xref: /llvm-project/llvm/examples/Kaleidoscope/Chapter8/toy.cpp (revision 114325b86a3d0360afc772c44335fc9092686f58)
1 #include "llvm/ADT/APFloat.h"
2 #include "llvm/ADT/STLExtras.h"
3 #include "llvm/IR/BasicBlock.h"
4 #include "llvm/IR/Constants.h"
5 #include "llvm/IR/DerivedTypes.h"
6 #include "llvm/IR/Function.h"
7 #include "llvm/IR/IRBuilder.h"
8 #include "llvm/IR/Instructions.h"
9 #include "llvm/IR/LLVMContext.h"
10 #include "llvm/IR/LegacyPassManager.h"
11 #include "llvm/IR/Module.h"
12 #include "llvm/IR/Type.h"
13 #include "llvm/IR/Verifier.h"
14 #include "llvm/MC/TargetRegistry.h"
15 #include "llvm/Support/FileSystem.h"
16 #include "llvm/Support/TargetSelect.h"
17 #include "llvm/Support/raw_ostream.h"
18 #include "llvm/Target/TargetMachine.h"
19 #include "llvm/Target/TargetOptions.h"
20 #include "llvm/TargetParser/Host.h"
21 #include <algorithm>
22 #include <cassert>
23 #include <cctype>
24 #include <cstdio>
25 #include <cstdlib>
26 #include <map>
27 #include <memory>
28 #include <string>
29 #include <system_error>
30 #include <utility>
31 #include <vector>
32 
33 using namespace llvm;
34 using namespace llvm::sys;
35 
36 //===----------------------------------------------------------------------===//
37 // Lexer
38 //===----------------------------------------------------------------------===//
39 
40 // The lexer returns tokens [0-255] if it is an unknown character, otherwise one
41 // of these for known things.
42 enum Token {
43   tok_eof = -1,
44 
45   // commands
46   tok_def = -2,
47   tok_extern = -3,
48 
49   // primary
50   tok_identifier = -4,
51   tok_number = -5,
52 
53   // control
54   tok_if = -6,
55   tok_then = -7,
56   tok_else = -8,
57   tok_for = -9,
58   tok_in = -10,
59 
60   // operators
61   tok_binary = -11,
62   tok_unary = -12,
63 
64   // var definition
65   tok_var = -13
66 };
67 
68 static std::string IdentifierStr; // Filled in if tok_identifier
69 static double NumVal;             // Filled in if tok_number
70 
71 /// gettok - Return the next token from standard input.
gettok()72 static int gettok() {
73   static int LastChar = ' ';
74 
75   // Skip any whitespace.
76   while (isspace(LastChar))
77     LastChar = getchar();
78 
79   if (isalpha(LastChar)) { // identifier: [a-zA-Z][a-zA-Z0-9]*
80     IdentifierStr = LastChar;
81     while (isalnum((LastChar = getchar())))
82       IdentifierStr += LastChar;
83 
84     if (IdentifierStr == "def")
85       return tok_def;
86     if (IdentifierStr == "extern")
87       return tok_extern;
88     if (IdentifierStr == "if")
89       return tok_if;
90     if (IdentifierStr == "then")
91       return tok_then;
92     if (IdentifierStr == "else")
93       return tok_else;
94     if (IdentifierStr == "for")
95       return tok_for;
96     if (IdentifierStr == "in")
97       return tok_in;
98     if (IdentifierStr == "binary")
99       return tok_binary;
100     if (IdentifierStr == "unary")
101       return tok_unary;
102     if (IdentifierStr == "var")
103       return tok_var;
104     return tok_identifier;
105   }
106 
107   if (isdigit(LastChar) || LastChar == '.') { // Number: [0-9.]+
108     std::string NumStr;
109     do {
110       NumStr += LastChar;
111       LastChar = getchar();
112     } while (isdigit(LastChar) || LastChar == '.');
113 
114     NumVal = strtod(NumStr.c_str(), nullptr);
115     return tok_number;
116   }
117 
118   if (LastChar == '#') {
119     // Comment until end of line.
120     do
121       LastChar = getchar();
122     while (LastChar != EOF && LastChar != '\n' && LastChar != '\r');
123 
124     if (LastChar != EOF)
125       return gettok();
126   }
127 
128   // Check for end of file.  Don't eat the EOF.
129   if (LastChar == EOF)
130     return tok_eof;
131 
132   // Otherwise, just return the character as its ascii value.
133   int ThisChar = LastChar;
134   LastChar = getchar();
135   return ThisChar;
136 }
137 
138 //===----------------------------------------------------------------------===//
139 // Abstract Syntax Tree (aka Parse Tree)
140 //===----------------------------------------------------------------------===//
141 
142 namespace {
143 
144 /// ExprAST - Base class for all expression nodes.
145 class ExprAST {
146 public:
147   virtual ~ExprAST() = default;
148 
149   virtual Value *codegen() = 0;
150 };
151 
152 /// NumberExprAST - Expression class for numeric literals like "1.0".
153 class NumberExprAST : public ExprAST {
154   double Val;
155 
156 public:
NumberExprAST(double Val)157   NumberExprAST(double Val) : Val(Val) {}
158 
159   Value *codegen() override;
160 };
161 
162 /// VariableExprAST - Expression class for referencing a variable, like "a".
163 class VariableExprAST : public ExprAST {
164   std::string Name;
165 
166 public:
VariableExprAST(const std::string & Name)167   VariableExprAST(const std::string &Name) : Name(Name) {}
168 
169   Value *codegen() override;
getName() const170   const std::string &getName() const { return Name; }
171 };
172 
173 /// UnaryExprAST - Expression class for a unary operator.
174 class UnaryExprAST : public ExprAST {
175   char Opcode;
176   std::unique_ptr<ExprAST> Operand;
177 
178 public:
UnaryExprAST(char Opcode,std::unique_ptr<ExprAST> Operand)179   UnaryExprAST(char Opcode, std::unique_ptr<ExprAST> Operand)
180       : Opcode(Opcode), Operand(std::move(Operand)) {}
181 
182   Value *codegen() override;
183 };
184 
185 /// BinaryExprAST - Expression class for a binary operator.
186 class BinaryExprAST : public ExprAST {
187   char Op;
188   std::unique_ptr<ExprAST> LHS, RHS;
189 
190 public:
BinaryExprAST(char Op,std::unique_ptr<ExprAST> LHS,std::unique_ptr<ExprAST> RHS)191   BinaryExprAST(char Op, std::unique_ptr<ExprAST> LHS,
192                 std::unique_ptr<ExprAST> RHS)
193       : Op(Op), LHS(std::move(LHS)), RHS(std::move(RHS)) {}
194 
195   Value *codegen() override;
196 };
197 
198 /// CallExprAST - Expression class for function calls.
199 class CallExprAST : public ExprAST {
200   std::string Callee;
201   std::vector<std::unique_ptr<ExprAST>> Args;
202 
203 public:
CallExprAST(const std::string & Callee,std::vector<std::unique_ptr<ExprAST>> Args)204   CallExprAST(const std::string &Callee,
205               std::vector<std::unique_ptr<ExprAST>> Args)
206       : Callee(Callee), Args(std::move(Args)) {}
207 
208   Value *codegen() override;
209 };
210 
211 /// IfExprAST - Expression class for if/then/else.
212 class IfExprAST : public ExprAST {
213   std::unique_ptr<ExprAST> Cond, Then, Else;
214 
215 public:
IfExprAST(std::unique_ptr<ExprAST> Cond,std::unique_ptr<ExprAST> Then,std::unique_ptr<ExprAST> Else)216   IfExprAST(std::unique_ptr<ExprAST> Cond, std::unique_ptr<ExprAST> Then,
217             std::unique_ptr<ExprAST> Else)
218       : Cond(std::move(Cond)), Then(std::move(Then)), Else(std::move(Else)) {}
219 
220   Value *codegen() override;
221 };
222 
223 /// ForExprAST - Expression class for for/in.
224 class ForExprAST : public ExprAST {
225   std::string VarName;
226   std::unique_ptr<ExprAST> Start, End, Step, Body;
227 
228 public:
ForExprAST(const std::string & VarName,std::unique_ptr<ExprAST> Start,std::unique_ptr<ExprAST> End,std::unique_ptr<ExprAST> Step,std::unique_ptr<ExprAST> Body)229   ForExprAST(const std::string &VarName, std::unique_ptr<ExprAST> Start,
230              std::unique_ptr<ExprAST> End, std::unique_ptr<ExprAST> Step,
231              std::unique_ptr<ExprAST> Body)
232       : VarName(VarName), Start(std::move(Start)), End(std::move(End)),
233         Step(std::move(Step)), Body(std::move(Body)) {}
234 
235   Value *codegen() override;
236 };
237 
238 /// VarExprAST - Expression class for var/in
239 class VarExprAST : public ExprAST {
240   std::vector<std::pair<std::string, std::unique_ptr<ExprAST>>> VarNames;
241   std::unique_ptr<ExprAST> Body;
242 
243 public:
VarExprAST(std::vector<std::pair<std::string,std::unique_ptr<ExprAST>>> VarNames,std::unique_ptr<ExprAST> Body)244   VarExprAST(
245       std::vector<std::pair<std::string, std::unique_ptr<ExprAST>>> VarNames,
246       std::unique_ptr<ExprAST> Body)
247       : VarNames(std::move(VarNames)), Body(std::move(Body)) {}
248 
249   Value *codegen() override;
250 };
251 
252 /// PrototypeAST - This class represents the "prototype" for a function,
253 /// which captures its name, and its argument names (thus implicitly the number
254 /// of arguments the function takes), as well as if it is an operator.
255 class PrototypeAST {
256   std::string Name;
257   std::vector<std::string> Args;
258   bool IsOperator;
259   unsigned Precedence; // Precedence if a binary op.
260 
261 public:
PrototypeAST(const std::string & Name,std::vector<std::string> Args,bool IsOperator=false,unsigned Prec=0)262   PrototypeAST(const std::string &Name, std::vector<std::string> Args,
263                bool IsOperator = false, unsigned Prec = 0)
264       : Name(Name), Args(std::move(Args)), IsOperator(IsOperator),
265         Precedence(Prec) {}
266 
267   Function *codegen();
getName() const268   const std::string &getName() const { return Name; }
269 
isUnaryOp() const270   bool isUnaryOp() const { return IsOperator && Args.size() == 1; }
isBinaryOp() const271   bool isBinaryOp() const { return IsOperator && Args.size() == 2; }
272 
getOperatorName() const273   char getOperatorName() const {
274     assert(isUnaryOp() || isBinaryOp());
275     return Name[Name.size() - 1];
276   }
277 
getBinaryPrecedence() const278   unsigned getBinaryPrecedence() const { return Precedence; }
279 };
280 
281 /// FunctionAST - This class represents a function definition itself.
282 class FunctionAST {
283   std::unique_ptr<PrototypeAST> Proto;
284   std::unique_ptr<ExprAST> Body;
285 
286 public:
FunctionAST(std::unique_ptr<PrototypeAST> Proto,std::unique_ptr<ExprAST> Body)287   FunctionAST(std::unique_ptr<PrototypeAST> Proto,
288               std::unique_ptr<ExprAST> Body)
289       : Proto(std::move(Proto)), Body(std::move(Body)) {}
290 
291   Function *codegen();
292 };
293 
294 } // end anonymous namespace
295 
296 //===----------------------------------------------------------------------===//
297 // Parser
298 //===----------------------------------------------------------------------===//
299 
300 /// CurTok/getNextToken - Provide a simple token buffer.  CurTok is the current
301 /// token the parser is looking at.  getNextToken reads another token from the
302 /// lexer and updates CurTok with its results.
303 static int CurTok;
getNextToken()304 static int getNextToken() { return CurTok = gettok(); }
305 
306 /// BinopPrecedence - This holds the precedence for each binary operator that is
307 /// defined.
308 static std::map<char, int> BinopPrecedence;
309 
310 /// GetTokPrecedence - Get the precedence of the pending binary operator token.
GetTokPrecedence()311 static int GetTokPrecedence() {
312   if (!isascii(CurTok))
313     return -1;
314 
315   // Make sure it's a declared binop.
316   int TokPrec = BinopPrecedence[CurTok];
317   if (TokPrec <= 0)
318     return -1;
319   return TokPrec;
320 }
321 
322 /// LogError* - These are little helper functions for error handling.
LogError(const char * Str)323 std::unique_ptr<ExprAST> LogError(const char *Str) {
324   fprintf(stderr, "Error: %s\n", Str);
325   return nullptr;
326 }
327 
LogErrorP(const char * Str)328 std::unique_ptr<PrototypeAST> LogErrorP(const char *Str) {
329   LogError(Str);
330   return nullptr;
331 }
332 
333 static std::unique_ptr<ExprAST> ParseExpression();
334 
335 /// numberexpr ::= number
ParseNumberExpr()336 static std::unique_ptr<ExprAST> ParseNumberExpr() {
337   auto Result = std::make_unique<NumberExprAST>(NumVal);
338   getNextToken(); // consume the number
339   return std::move(Result);
340 }
341 
342 /// parenexpr ::= '(' expression ')'
ParseParenExpr()343 static std::unique_ptr<ExprAST> ParseParenExpr() {
344   getNextToken(); // eat (.
345   auto V = ParseExpression();
346   if (!V)
347     return nullptr;
348 
349   if (CurTok != ')')
350     return LogError("expected ')'");
351   getNextToken(); // eat ).
352   return V;
353 }
354 
355 /// identifierexpr
356 ///   ::= identifier
357 ///   ::= identifier '(' expression* ')'
ParseIdentifierExpr()358 static std::unique_ptr<ExprAST> ParseIdentifierExpr() {
359   std::string IdName = IdentifierStr;
360 
361   getNextToken(); // eat identifier.
362 
363   if (CurTok != '(') // Simple variable ref.
364     return std::make_unique<VariableExprAST>(IdName);
365 
366   // Call.
367   getNextToken(); // eat (
368   std::vector<std::unique_ptr<ExprAST>> Args;
369   if (CurTok != ')') {
370     while (true) {
371       if (auto Arg = ParseExpression())
372         Args.push_back(std::move(Arg));
373       else
374         return nullptr;
375 
376       if (CurTok == ')')
377         break;
378 
379       if (CurTok != ',')
380         return LogError("Expected ')' or ',' in argument list");
381       getNextToken();
382     }
383   }
384 
385   // Eat the ')'.
386   getNextToken();
387 
388   return std::make_unique<CallExprAST>(IdName, std::move(Args));
389 }
390 
391 /// ifexpr ::= 'if' expression 'then' expression 'else' expression
ParseIfExpr()392 static std::unique_ptr<ExprAST> ParseIfExpr() {
393   getNextToken(); // eat the if.
394 
395   // condition.
396   auto Cond = ParseExpression();
397   if (!Cond)
398     return nullptr;
399 
400   if (CurTok != tok_then)
401     return LogError("expected then");
402   getNextToken(); // eat the then
403 
404   auto Then = ParseExpression();
405   if (!Then)
406     return nullptr;
407 
408   if (CurTok != tok_else)
409     return LogError("expected else");
410 
411   getNextToken();
412 
413   auto Else = ParseExpression();
414   if (!Else)
415     return nullptr;
416 
417   return std::make_unique<IfExprAST>(std::move(Cond), std::move(Then),
418                                       std::move(Else));
419 }
420 
421 /// forexpr ::= 'for' identifier '=' expr ',' expr (',' expr)? 'in' expression
ParseForExpr()422 static std::unique_ptr<ExprAST> ParseForExpr() {
423   getNextToken(); // eat the for.
424 
425   if (CurTok != tok_identifier)
426     return LogError("expected identifier after for");
427 
428   std::string IdName = IdentifierStr;
429   getNextToken(); // eat identifier.
430 
431   if (CurTok != '=')
432     return LogError("expected '=' after for");
433   getNextToken(); // eat '='.
434 
435   auto Start = ParseExpression();
436   if (!Start)
437     return nullptr;
438   if (CurTok != ',')
439     return LogError("expected ',' after for start value");
440   getNextToken();
441 
442   auto End = ParseExpression();
443   if (!End)
444     return nullptr;
445 
446   // The step value is optional.
447   std::unique_ptr<ExprAST> Step;
448   if (CurTok == ',') {
449     getNextToken();
450     Step = ParseExpression();
451     if (!Step)
452       return nullptr;
453   }
454 
455   if (CurTok != tok_in)
456     return LogError("expected 'in' after for");
457   getNextToken(); // eat 'in'.
458 
459   auto Body = ParseExpression();
460   if (!Body)
461     return nullptr;
462 
463   return std::make_unique<ForExprAST>(IdName, std::move(Start), std::move(End),
464                                        std::move(Step), std::move(Body));
465 }
466 
467 /// varexpr ::= 'var' identifier ('=' expression)?
468 //                    (',' identifier ('=' expression)?)* 'in' expression
ParseVarExpr()469 static std::unique_ptr<ExprAST> ParseVarExpr() {
470   getNextToken(); // eat the var.
471 
472   std::vector<std::pair<std::string, std::unique_ptr<ExprAST>>> VarNames;
473 
474   // At least one variable name is required.
475   if (CurTok != tok_identifier)
476     return LogError("expected identifier after var");
477 
478   while (true) {
479     std::string Name = IdentifierStr;
480     getNextToken(); // eat identifier.
481 
482     // Read the optional initializer.
483     std::unique_ptr<ExprAST> Init = nullptr;
484     if (CurTok == '=') {
485       getNextToken(); // eat the '='.
486 
487       Init = ParseExpression();
488       if (!Init)
489         return nullptr;
490     }
491 
492     VarNames.push_back(std::make_pair(Name, std::move(Init)));
493 
494     // End of var list, exit loop.
495     if (CurTok != ',')
496       break;
497     getNextToken(); // eat the ','.
498 
499     if (CurTok != tok_identifier)
500       return LogError("expected identifier list after var");
501   }
502 
503   // At this point, we have to have 'in'.
504   if (CurTok != tok_in)
505     return LogError("expected 'in' keyword after 'var'");
506   getNextToken(); // eat 'in'.
507 
508   auto Body = ParseExpression();
509   if (!Body)
510     return nullptr;
511 
512   return std::make_unique<VarExprAST>(std::move(VarNames), std::move(Body));
513 }
514 
515 /// primary
516 ///   ::= identifierexpr
517 ///   ::= numberexpr
518 ///   ::= parenexpr
519 ///   ::= ifexpr
520 ///   ::= forexpr
521 ///   ::= varexpr
ParsePrimary()522 static std::unique_ptr<ExprAST> ParsePrimary() {
523   switch (CurTok) {
524   default:
525     return LogError("unknown token when expecting an expression");
526   case tok_identifier:
527     return ParseIdentifierExpr();
528   case tok_number:
529     return ParseNumberExpr();
530   case '(':
531     return ParseParenExpr();
532   case tok_if:
533     return ParseIfExpr();
534   case tok_for:
535     return ParseForExpr();
536   case tok_var:
537     return ParseVarExpr();
538   }
539 }
540 
541 /// unary
542 ///   ::= primary
543 ///   ::= '!' unary
ParseUnary()544 static std::unique_ptr<ExprAST> ParseUnary() {
545   // If the current token is not an operator, it must be a primary expr.
546   if (!isascii(CurTok) || CurTok == '(' || CurTok == ',')
547     return ParsePrimary();
548 
549   // If this is a unary operator, read it.
550   int Opc = CurTok;
551   getNextToken();
552   if (auto Operand = ParseUnary())
553     return std::make_unique<UnaryExprAST>(Opc, std::move(Operand));
554   return nullptr;
555 }
556 
557 /// binoprhs
558 ///   ::= ('+' unary)*
ParseBinOpRHS(int ExprPrec,std::unique_ptr<ExprAST> LHS)559 static std::unique_ptr<ExprAST> ParseBinOpRHS(int ExprPrec,
560                                               std::unique_ptr<ExprAST> LHS) {
561   // If this is a binop, find its precedence.
562   while (true) {
563     int TokPrec = GetTokPrecedence();
564 
565     // If this is a binop that binds at least as tightly as the current binop,
566     // consume it, otherwise we are done.
567     if (TokPrec < ExprPrec)
568       return LHS;
569 
570     // Okay, we know this is a binop.
571     int BinOp = CurTok;
572     getNextToken(); // eat binop
573 
574     // Parse the unary expression after the binary operator.
575     auto RHS = ParseUnary();
576     if (!RHS)
577       return nullptr;
578 
579     // If BinOp binds less tightly with RHS than the operator after RHS, let
580     // the pending operator take RHS as its LHS.
581     int NextPrec = GetTokPrecedence();
582     if (TokPrec < NextPrec) {
583       RHS = ParseBinOpRHS(TokPrec + 1, std::move(RHS));
584       if (!RHS)
585         return nullptr;
586     }
587 
588     // Merge LHS/RHS.
589     LHS =
590         std::make_unique<BinaryExprAST>(BinOp, std::move(LHS), std::move(RHS));
591   }
592 }
593 
594 /// expression
595 ///   ::= unary binoprhs
596 ///
ParseExpression()597 static std::unique_ptr<ExprAST> ParseExpression() {
598   auto LHS = ParseUnary();
599   if (!LHS)
600     return nullptr;
601 
602   return ParseBinOpRHS(0, std::move(LHS));
603 }
604 
605 /// prototype
606 ///   ::= id '(' id* ')'
607 ///   ::= binary LETTER number? (id, id)
608 ///   ::= unary LETTER (id)
ParsePrototype()609 static std::unique_ptr<PrototypeAST> ParsePrototype() {
610   std::string FnName;
611 
612   unsigned Kind = 0; // 0 = identifier, 1 = unary, 2 = binary.
613   unsigned BinaryPrecedence = 30;
614 
615   switch (CurTok) {
616   default:
617     return LogErrorP("Expected function name in prototype");
618   case tok_identifier:
619     FnName = IdentifierStr;
620     Kind = 0;
621     getNextToken();
622     break;
623   case tok_unary:
624     getNextToken();
625     if (!isascii(CurTok))
626       return LogErrorP("Expected unary operator");
627     FnName = "unary";
628     FnName += (char)CurTok;
629     Kind = 1;
630     getNextToken();
631     break;
632   case tok_binary:
633     getNextToken();
634     if (!isascii(CurTok))
635       return LogErrorP("Expected binary operator");
636     FnName = "binary";
637     FnName += (char)CurTok;
638     Kind = 2;
639     getNextToken();
640 
641     // Read the precedence if present.
642     if (CurTok == tok_number) {
643       if (NumVal < 1 || NumVal > 100)
644         return LogErrorP("Invalid precedence: must be 1..100");
645       BinaryPrecedence = (unsigned)NumVal;
646       getNextToken();
647     }
648     break;
649   }
650 
651   if (CurTok != '(')
652     return LogErrorP("Expected '(' in prototype");
653 
654   std::vector<std::string> ArgNames;
655   while (getNextToken() == tok_identifier)
656     ArgNames.push_back(IdentifierStr);
657   if (CurTok != ')')
658     return LogErrorP("Expected ')' in prototype");
659 
660   // success.
661   getNextToken(); // eat ')'.
662 
663   // Verify right number of names for operator.
664   if (Kind && ArgNames.size() != Kind)
665     return LogErrorP("Invalid number of operands for operator");
666 
667   return std::make_unique<PrototypeAST>(FnName, ArgNames, Kind != 0,
668                                          BinaryPrecedence);
669 }
670 
671 /// definition ::= 'def' prototype expression
ParseDefinition()672 static std::unique_ptr<FunctionAST> ParseDefinition() {
673   getNextToken(); // eat def.
674   auto Proto = ParsePrototype();
675   if (!Proto)
676     return nullptr;
677 
678   if (auto E = ParseExpression())
679     return std::make_unique<FunctionAST>(std::move(Proto), std::move(E));
680   return nullptr;
681 }
682 
683 /// toplevelexpr ::= expression
ParseTopLevelExpr()684 static std::unique_ptr<FunctionAST> ParseTopLevelExpr() {
685   if (auto E = ParseExpression()) {
686     // Make an anonymous proto.
687     auto Proto = std::make_unique<PrototypeAST>("__anon_expr",
688                                                  std::vector<std::string>());
689     return std::make_unique<FunctionAST>(std::move(Proto), std::move(E));
690   }
691   return nullptr;
692 }
693 
694 /// external ::= 'extern' prototype
ParseExtern()695 static std::unique_ptr<PrototypeAST> ParseExtern() {
696   getNextToken(); // eat extern.
697   return ParsePrototype();
698 }
699 
700 //===----------------------------------------------------------------------===//
701 // Code Generation
702 //===----------------------------------------------------------------------===//
703 
704 static std::unique_ptr<LLVMContext> TheContext;
705 static std::unique_ptr<Module> TheModule;
706 static std::unique_ptr<IRBuilder<>> Builder;
707 static std::map<std::string, AllocaInst *> NamedValues;
708 static std::map<std::string, std::unique_ptr<PrototypeAST>> FunctionProtos;
709 static ExitOnError ExitOnErr;
710 
LogErrorV(const char * Str)711 Value *LogErrorV(const char *Str) {
712   LogError(Str);
713   return nullptr;
714 }
715 
getFunction(std::string Name)716 Function *getFunction(std::string Name) {
717   // First, see if the function has already been added to the current module.
718   if (auto *F = TheModule->getFunction(Name))
719     return F;
720 
721   // If not, check whether we can codegen the declaration from some existing
722   // prototype.
723   auto FI = FunctionProtos.find(Name);
724   if (FI != FunctionProtos.end())
725     return FI->second->codegen();
726 
727   // If no existing prototype exists, return null.
728   return nullptr;
729 }
730 
731 /// CreateEntryBlockAlloca - Create an alloca instruction in the entry block of
732 /// the function.  This is used for mutable variables etc.
CreateEntryBlockAlloca(Function * TheFunction,StringRef VarName)733 static AllocaInst *CreateEntryBlockAlloca(Function *TheFunction,
734                                           StringRef VarName) {
735   IRBuilder<> TmpB(&TheFunction->getEntryBlock(),
736                    TheFunction->getEntryBlock().begin());
737   return TmpB.CreateAlloca(Type::getDoubleTy(*TheContext), nullptr, VarName);
738 }
739 
codegen()740 Value *NumberExprAST::codegen() {
741   return ConstantFP::get(*TheContext, APFloat(Val));
742 }
743 
codegen()744 Value *VariableExprAST::codegen() {
745   // Look this variable up in the function.
746   Value *V = NamedValues[Name];
747   if (!V)
748     return LogErrorV("Unknown variable name");
749 
750   // Load the value.
751   return Builder->CreateLoad(Type::getDoubleTy(*TheContext), V, Name.c_str());
752 }
753 
codegen()754 Value *UnaryExprAST::codegen() {
755   Value *OperandV = Operand->codegen();
756   if (!OperandV)
757     return nullptr;
758 
759   Function *F = getFunction(std::string("unary") + Opcode);
760   if (!F)
761     return LogErrorV("Unknown unary operator");
762 
763   return Builder->CreateCall(F, OperandV, "unop");
764 }
765 
codegen()766 Value *BinaryExprAST::codegen() {
767   // Special case '=' because we don't want to emit the LHS as an expression.
768   if (Op == '=') {
769     // Assignment requires the LHS to be an identifier.
770     // This assume we're building without RTTI because LLVM builds that way by
771     // default.  If you build LLVM with RTTI this can be changed to a
772     // dynamic_cast for automatic error checking.
773     VariableExprAST *LHSE = static_cast<VariableExprAST *>(LHS.get());
774     if (!LHSE)
775       return LogErrorV("destination of '=' must be a variable");
776     // Codegen the RHS.
777     Value *Val = RHS->codegen();
778     if (!Val)
779       return nullptr;
780 
781     // Look up the name.
782     Value *Variable = NamedValues[LHSE->getName()];
783     if (!Variable)
784       return LogErrorV("Unknown variable name");
785 
786     Builder->CreateStore(Val, Variable);
787     return Val;
788   }
789 
790   Value *L = LHS->codegen();
791   Value *R = RHS->codegen();
792   if (!L || !R)
793     return nullptr;
794 
795   switch (Op) {
796   case '+':
797     return Builder->CreateFAdd(L, R, "addtmp");
798   case '-':
799     return Builder->CreateFSub(L, R, "subtmp");
800   case '*':
801     return Builder->CreateFMul(L, R, "multmp");
802   case '<':
803     L = Builder->CreateFCmpULT(L, R, "cmptmp");
804     // Convert bool 0/1 to double 0.0 or 1.0
805     return Builder->CreateUIToFP(L, Type::getDoubleTy(*TheContext), "booltmp");
806   default:
807     break;
808   }
809 
810   // If it wasn't a builtin binary operator, it must be a user defined one. Emit
811   // a call to it.
812   Function *F = getFunction(std::string("binary") + Op);
813   assert(F && "binary operator not found!");
814 
815   Value *Ops[] = {L, R};
816   return Builder->CreateCall(F, Ops, "binop");
817 }
818 
codegen()819 Value *CallExprAST::codegen() {
820   // Look up the name in the global module table.
821   Function *CalleeF = getFunction(Callee);
822   if (!CalleeF)
823     return LogErrorV("Unknown function referenced");
824 
825   // If argument mismatch error.
826   if (CalleeF->arg_size() != Args.size())
827     return LogErrorV("Incorrect # arguments passed");
828 
829   std::vector<Value *> ArgsV;
830   for (unsigned i = 0, e = Args.size(); i != e; ++i) {
831     ArgsV.push_back(Args[i]->codegen());
832     if (!ArgsV.back())
833       return nullptr;
834   }
835 
836   return Builder->CreateCall(CalleeF, ArgsV, "calltmp");
837 }
838 
codegen()839 Value *IfExprAST::codegen() {
840   Value *CondV = Cond->codegen();
841   if (!CondV)
842     return nullptr;
843 
844   // Convert condition to a bool by comparing non-equal to 0.0.
845   CondV = Builder->CreateFCmpONE(
846       CondV, ConstantFP::get(*TheContext, APFloat(0.0)), "ifcond");
847 
848   Function *TheFunction = Builder->GetInsertBlock()->getParent();
849 
850   // Create blocks for the then and else cases.  Insert the 'then' block at the
851   // end of the function.
852   BasicBlock *ThenBB = BasicBlock::Create(*TheContext, "then", TheFunction);
853   BasicBlock *ElseBB = BasicBlock::Create(*TheContext, "else");
854   BasicBlock *MergeBB = BasicBlock::Create(*TheContext, "ifcont");
855 
856   Builder->CreateCondBr(CondV, ThenBB, ElseBB);
857 
858   // Emit then value.
859   Builder->SetInsertPoint(ThenBB);
860 
861   Value *ThenV = Then->codegen();
862   if (!ThenV)
863     return nullptr;
864 
865   Builder->CreateBr(MergeBB);
866   // Codegen of 'Then' can change the current block, update ThenBB for the PHI.
867   ThenBB = Builder->GetInsertBlock();
868 
869   // Emit else block.
870   TheFunction->insert(TheFunction->end(), ElseBB);
871   Builder->SetInsertPoint(ElseBB);
872 
873   Value *ElseV = Else->codegen();
874   if (!ElseV)
875     return nullptr;
876 
877   Builder->CreateBr(MergeBB);
878   // Codegen of 'Else' can change the current block, update ElseBB for the PHI.
879   ElseBB = Builder->GetInsertBlock();
880 
881   // Emit merge block.
882   TheFunction->insert(TheFunction->end(), MergeBB);
883   Builder->SetInsertPoint(MergeBB);
884   PHINode *PN = Builder->CreatePHI(Type::getDoubleTy(*TheContext), 2, "iftmp");
885 
886   PN->addIncoming(ThenV, ThenBB);
887   PN->addIncoming(ElseV, ElseBB);
888   return PN;
889 }
890 
891 // Output for-loop as:
892 //   var = alloca double
893 //   ...
894 //   start = startexpr
895 //   store start -> var
896 //   goto loop
897 // loop:
898 //   ...
899 //   bodyexpr
900 //   ...
901 // loopend:
902 //   step = stepexpr
903 //   endcond = endexpr
904 //
905 //   curvar = load var
906 //   nextvar = curvar + step
907 //   store nextvar -> var
908 //   br endcond, loop, endloop
909 // outloop:
codegen()910 Value *ForExprAST::codegen() {
911   Function *TheFunction = Builder->GetInsertBlock()->getParent();
912 
913   // Create an alloca for the variable in the entry block.
914   AllocaInst *Alloca = CreateEntryBlockAlloca(TheFunction, VarName);
915 
916   // Emit the start code first, without 'variable' in scope.
917   Value *StartVal = Start->codegen();
918   if (!StartVal)
919     return nullptr;
920 
921   // Store the value into the alloca.
922   Builder->CreateStore(StartVal, Alloca);
923 
924   // Make the new basic block for the loop header, inserting after current
925   // block.
926   BasicBlock *LoopBB = BasicBlock::Create(*TheContext, "loop", TheFunction);
927 
928   // Insert an explicit fall through from the current block to the LoopBB.
929   Builder->CreateBr(LoopBB);
930 
931   // Start insertion in LoopBB.
932   Builder->SetInsertPoint(LoopBB);
933 
934   // Within the loop, the variable is defined equal to the PHI node.  If it
935   // shadows an existing variable, we have to restore it, so save it now.
936   AllocaInst *OldVal = NamedValues[VarName];
937   NamedValues[VarName] = Alloca;
938 
939   // Emit the body of the loop.  This, like any other expr, can change the
940   // current BB.  Note that we ignore the value computed by the body, but don't
941   // allow an error.
942   if (!Body->codegen())
943     return nullptr;
944 
945   // Emit the step value.
946   Value *StepVal = nullptr;
947   if (Step) {
948     StepVal = Step->codegen();
949     if (!StepVal)
950       return nullptr;
951   } else {
952     // If not specified, use 1.0.
953     StepVal = ConstantFP::get(*TheContext, APFloat(1.0));
954   }
955 
956   // Compute the end condition.
957   Value *EndCond = End->codegen();
958   if (!EndCond)
959     return nullptr;
960 
961   // Reload, increment, and restore the alloca.  This handles the case where
962   // the body of the loop mutates the variable.
963   Value *CurVar = Builder->CreateLoad(Type::getDoubleTy(*TheContext), Alloca,
964                                       VarName.c_str());
965   Value *NextVar = Builder->CreateFAdd(CurVar, StepVal, "nextvar");
966   Builder->CreateStore(NextVar, Alloca);
967 
968   // Convert condition to a bool by comparing non-equal to 0.0.
969   EndCond = Builder->CreateFCmpONE(
970       EndCond, ConstantFP::get(*TheContext, APFloat(0.0)), "loopcond");
971 
972   // Create the "after loop" block and insert it.
973   BasicBlock *AfterBB =
974       BasicBlock::Create(*TheContext, "afterloop", TheFunction);
975 
976   // Insert the conditional branch into the end of LoopEndBB.
977   Builder->CreateCondBr(EndCond, LoopBB, AfterBB);
978 
979   // Any new code will be inserted in AfterBB.
980   Builder->SetInsertPoint(AfterBB);
981 
982   // Restore the unshadowed variable.
983   if (OldVal)
984     NamedValues[VarName] = OldVal;
985   else
986     NamedValues.erase(VarName);
987 
988   // for expr always returns 0.0.
989   return Constant::getNullValue(Type::getDoubleTy(*TheContext));
990 }
991 
codegen()992 Value *VarExprAST::codegen() {
993   std::vector<AllocaInst *> OldBindings;
994 
995   Function *TheFunction = Builder->GetInsertBlock()->getParent();
996 
997   // Register all variables and emit their initializer.
998   for (unsigned i = 0, e = VarNames.size(); i != e; ++i) {
999     const std::string &VarName = VarNames[i].first;
1000     ExprAST *Init = VarNames[i].second.get();
1001 
1002     // Emit the initializer before adding the variable to scope, this prevents
1003     // the initializer from referencing the variable itself, and permits stuff
1004     // like this:
1005     //  var a = 1 in
1006     //    var a = a in ...   # refers to outer 'a'.
1007     Value *InitVal;
1008     if (Init) {
1009       InitVal = Init->codegen();
1010       if (!InitVal)
1011         return nullptr;
1012     } else { // If not specified, use 0.0.
1013       InitVal = ConstantFP::get(*TheContext, APFloat(0.0));
1014     }
1015 
1016     AllocaInst *Alloca = CreateEntryBlockAlloca(TheFunction, VarName);
1017     Builder->CreateStore(InitVal, Alloca);
1018 
1019     // Remember the old variable binding so that we can restore the binding when
1020     // we unrecurse.
1021     OldBindings.push_back(NamedValues[VarName]);
1022 
1023     // Remember this binding.
1024     NamedValues[VarName] = Alloca;
1025   }
1026 
1027   // Codegen the body, now that all vars are in scope.
1028   Value *BodyVal = Body->codegen();
1029   if (!BodyVal)
1030     return nullptr;
1031 
1032   // Pop all our variables from scope.
1033   for (unsigned i = 0, e = VarNames.size(); i != e; ++i)
1034     NamedValues[VarNames[i].first] = OldBindings[i];
1035 
1036   // Return the body computation.
1037   return BodyVal;
1038 }
1039 
codegen()1040 Function *PrototypeAST::codegen() {
1041   // Make the function type:  double(double,double) etc.
1042   std::vector<Type *> Doubles(Args.size(), Type::getDoubleTy(*TheContext));
1043   FunctionType *FT =
1044       FunctionType::get(Type::getDoubleTy(*TheContext), Doubles, false);
1045 
1046   Function *F =
1047       Function::Create(FT, Function::ExternalLinkage, Name, TheModule.get());
1048 
1049   // Set names for all arguments.
1050   unsigned Idx = 0;
1051   for (auto &Arg : F->args())
1052     Arg.setName(Args[Idx++]);
1053 
1054   return F;
1055 }
1056 
codegen()1057 Function *FunctionAST::codegen() {
1058   // Transfer ownership of the prototype to the FunctionProtos map, but keep a
1059   // reference to it for use below.
1060   auto &P = *Proto;
1061   FunctionProtos[Proto->getName()] = std::move(Proto);
1062   Function *TheFunction = getFunction(P.getName());
1063   if (!TheFunction)
1064     return nullptr;
1065 
1066   // If this is an operator, install it.
1067   if (P.isBinaryOp())
1068     BinopPrecedence[P.getOperatorName()] = P.getBinaryPrecedence();
1069 
1070   // Create a new basic block to start insertion into.
1071   BasicBlock *BB = BasicBlock::Create(*TheContext, "entry", TheFunction);
1072   Builder->SetInsertPoint(BB);
1073 
1074   // Record the function arguments in the NamedValues map.
1075   NamedValues.clear();
1076   for (auto &Arg : TheFunction->args()) {
1077     // Create an alloca for this variable.
1078     AllocaInst *Alloca = CreateEntryBlockAlloca(TheFunction, Arg.getName());
1079 
1080     // Store the initial value into the alloca.
1081     Builder->CreateStore(&Arg, Alloca);
1082 
1083     // Add arguments to variable symbol table.
1084     NamedValues[std::string(Arg.getName())] = Alloca;
1085   }
1086 
1087   if (Value *RetVal = Body->codegen()) {
1088     // Finish off the function.
1089     Builder->CreateRet(RetVal);
1090 
1091     // Validate the generated code, checking for consistency.
1092     verifyFunction(*TheFunction);
1093 
1094     return TheFunction;
1095   }
1096 
1097   // Error reading body, remove function.
1098   TheFunction->eraseFromParent();
1099 
1100   if (P.isBinaryOp())
1101     BinopPrecedence.erase(P.getOperatorName());
1102   return nullptr;
1103 }
1104 
1105 //===----------------------------------------------------------------------===//
1106 // Top-Level parsing and JIT Driver
1107 //===----------------------------------------------------------------------===//
1108 
InitializeModuleAndPassManager()1109 static void InitializeModuleAndPassManager() {
1110   // Open a new module.
1111   TheContext = std::make_unique<LLVMContext>();
1112   TheModule = std::make_unique<Module>("my cool jit", *TheContext);
1113 
1114   // Create a new builder for the module.
1115   Builder = std::make_unique<IRBuilder<>>(*TheContext);
1116 }
1117 
HandleDefinition()1118 static void HandleDefinition() {
1119   if (auto FnAST = ParseDefinition()) {
1120     if (auto *FnIR = FnAST->codegen()) {
1121       fprintf(stderr, "Read function definition:");
1122       FnIR->print(errs());
1123       fprintf(stderr, "\n");
1124     }
1125   } else {
1126     // Skip token for error recovery.
1127     getNextToken();
1128   }
1129 }
1130 
HandleExtern()1131 static void HandleExtern() {
1132   if (auto ProtoAST = ParseExtern()) {
1133     if (auto *FnIR = ProtoAST->codegen()) {
1134       fprintf(stderr, "Read extern: ");
1135       FnIR->print(errs());
1136       fprintf(stderr, "\n");
1137       FunctionProtos[ProtoAST->getName()] = std::move(ProtoAST);
1138     }
1139   } else {
1140     // Skip token for error recovery.
1141     getNextToken();
1142   }
1143 }
1144 
HandleTopLevelExpression()1145 static void HandleTopLevelExpression() {
1146   // Evaluate a top-level expression into an anonymous function.
1147   if (auto FnAST = ParseTopLevelExpr()) {
1148     FnAST->codegen();
1149   } else {
1150     // Skip token for error recovery.
1151     getNextToken();
1152   }
1153 }
1154 
1155 /// top ::= definition | external | expression | ';'
MainLoop()1156 static void MainLoop() {
1157   while (true) {
1158     switch (CurTok) {
1159     case tok_eof:
1160       return;
1161     case ';': // ignore top-level semicolons.
1162       getNextToken();
1163       break;
1164     case tok_def:
1165       HandleDefinition();
1166       break;
1167     case tok_extern:
1168       HandleExtern();
1169       break;
1170     default:
1171       HandleTopLevelExpression();
1172       break;
1173     }
1174   }
1175 }
1176 
1177 //===----------------------------------------------------------------------===//
1178 // "Library" functions that can be "extern'd" from user code.
1179 //===----------------------------------------------------------------------===//
1180 
1181 #ifdef _WIN32
1182 #define DLLEXPORT __declspec(dllexport)
1183 #else
1184 #define DLLEXPORT
1185 #endif
1186 
1187 /// putchard - putchar that takes a double and returns 0.
putchard(double X)1188 extern "C" DLLEXPORT double putchard(double X) {
1189   fputc((char)X, stderr);
1190   return 0;
1191 }
1192 
1193 /// printd - printf that takes a double prints it as "%f\n", returning 0.
printd(double X)1194 extern "C" DLLEXPORT double printd(double X) {
1195   fprintf(stderr, "%f\n", X);
1196   return 0;
1197 }
1198 
1199 //===----------------------------------------------------------------------===//
1200 // Main driver code.
1201 //===----------------------------------------------------------------------===//
1202 
main()1203 int main() {
1204   // Install standard binary operators.
1205   // 1 is lowest precedence.
1206   BinopPrecedence['<'] = 10;
1207   BinopPrecedence['+'] = 20;
1208   BinopPrecedence['-'] = 20;
1209   BinopPrecedence['*'] = 40; // highest.
1210 
1211   // Prime the first token.
1212   fprintf(stderr, "ready> ");
1213   getNextToken();
1214 
1215   InitializeModuleAndPassManager();
1216 
1217   // Run the main "interpreter loop" now.
1218   MainLoop();
1219 
1220   // Initialize the target registry etc.
1221   InitializeAllTargetInfos();
1222   InitializeAllTargets();
1223   InitializeAllTargetMCs();
1224   InitializeAllAsmParsers();
1225   InitializeAllAsmPrinters();
1226 
1227   auto TargetTriple = sys::getDefaultTargetTriple();
1228   TheModule->setTargetTriple(TargetTriple);
1229 
1230   std::string Error;
1231   auto Target = TargetRegistry::lookupTarget(TargetTriple, Error);
1232 
1233   // Print an error and exit if we couldn't find the requested target.
1234   // This generally occurs if we've forgotten to initialise the
1235   // TargetRegistry or we have a bogus target triple.
1236   if (!Target) {
1237     errs() << Error;
1238     return 1;
1239   }
1240 
1241   auto CPU = "generic";
1242   auto Features = "";
1243 
1244   TargetOptions opt;
1245   auto TheTargetMachine = Target->createTargetMachine(
1246       TargetTriple, CPU, Features, opt, Reloc::PIC_);
1247 
1248   TheModule->setDataLayout(TheTargetMachine->createDataLayout());
1249 
1250   auto Filename = "output.o";
1251   std::error_code EC;
1252   raw_fd_ostream dest(Filename, EC, sys::fs::OF_None);
1253 
1254   if (EC) {
1255     errs() << "Could not open file: " << EC.message();
1256     return 1;
1257   }
1258 
1259   legacy::PassManager pass;
1260   auto FileType = CodeGenFileType::ObjectFile;
1261 
1262   if (TheTargetMachine->addPassesToEmitFile(pass, dest, nullptr, FileType)) {
1263     errs() << "TheTargetMachine can't emit a file of this type";
1264     return 1;
1265   }
1266 
1267   pass.run(*TheModule);
1268   dest.flush();
1269 
1270   outs() << "Wrote " << Filename << "\n";
1271 
1272   return 0;
1273 }
1274