David Blaikie | c38e524 | 2015-02-08 07:20:04 +0000 | [diff] [blame] | 1 | |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 2 | #include "llvm/Analysis/Passes.h" |
| 3 | #include "llvm/ExecutionEngine/Orc/CompileUtils.h" |
| 4 | #include "llvm/ExecutionEngine/Orc/IRCompileLayer.h" |
| 5 | #include "llvm/ExecutionEngine/Orc/LazyEmittingLayer.h" |
| 6 | #include "llvm/ExecutionEngine/Orc/ObjectLinkingLayer.h" |
| 7 | #include "llvm/IR/DataLayout.h" |
| 8 | #include "llvm/IR/DerivedTypes.h" |
| 9 | #include "llvm/IR/IRBuilder.h" |
| 10 | #include "llvm/IR/LLVMContext.h" |
| 11 | #include "llvm/IR/Module.h" |
| 12 | #include "llvm/IR/Verifier.h" |
| 13 | #include "llvm/PassManager.h" |
| 14 | #include "llvm/Support/TargetSelect.h" |
| 15 | #include "llvm/Transforms/Scalar.h" |
| 16 | #include <cctype> |
Lang Hames | be9df34 | 2015-02-08 19:14:56 +0000 | [diff] [blame] | 17 | #include <iomanip> |
| 18 | #include <iostream> |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 19 | #include <map> |
Lang Hames | be9df34 | 2015-02-08 19:14:56 +0000 | [diff] [blame] | 20 | #include <sstream> |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 21 | #include <string> |
| 22 | #include <vector> |
| 23 | using namespace llvm; |
| 24 | |
| 25 | //===----------------------------------------------------------------------===// |
| 26 | // Lexer |
| 27 | //===----------------------------------------------------------------------===// |
| 28 | |
| 29 | // The lexer returns tokens [0-255] if it is an unknown character, otherwise one |
| 30 | // of these for known things. |
| 31 | enum Token { |
| 32 | tok_eof = -1, |
| 33 | |
| 34 | // commands |
| 35 | tok_def = -2, tok_extern = -3, |
| 36 | |
| 37 | // primary |
| 38 | tok_identifier = -4, tok_number = -5, |
| 39 | |
| 40 | // control |
| 41 | tok_if = -6, tok_then = -7, tok_else = -8, |
| 42 | tok_for = -9, tok_in = -10, |
| 43 | |
| 44 | // operators |
| 45 | tok_binary = -11, tok_unary = -12, |
| 46 | |
| 47 | // var definition |
| 48 | tok_var = -13 |
| 49 | }; |
| 50 | |
| 51 | static std::string IdentifierStr; // Filled in if tok_identifier |
| 52 | static double NumVal; // Filled in if tok_number |
| 53 | |
| 54 | /// gettok - Return the next token from standard input. |
| 55 | static int gettok() { |
| 56 | static int LastChar = ' '; |
| 57 | |
| 58 | // Skip any whitespace. |
| 59 | while (isspace(LastChar)) |
| 60 | LastChar = getchar(); |
| 61 | |
| 62 | if (isalpha(LastChar)) { // identifier: [a-zA-Z][a-zA-Z0-9]* |
| 63 | IdentifierStr = LastChar; |
| 64 | while (isalnum((LastChar = getchar()))) |
| 65 | IdentifierStr += LastChar; |
| 66 | |
| 67 | if (IdentifierStr == "def") return tok_def; |
| 68 | if (IdentifierStr == "extern") return tok_extern; |
| 69 | if (IdentifierStr == "if") return tok_if; |
| 70 | if (IdentifierStr == "then") return tok_then; |
| 71 | if (IdentifierStr == "else") return tok_else; |
| 72 | if (IdentifierStr == "for") return tok_for; |
| 73 | if (IdentifierStr == "in") return tok_in; |
| 74 | if (IdentifierStr == "binary") return tok_binary; |
| 75 | if (IdentifierStr == "unary") return tok_unary; |
| 76 | if (IdentifierStr == "var") return tok_var; |
| 77 | return tok_identifier; |
| 78 | } |
| 79 | |
| 80 | if (isdigit(LastChar) || LastChar == '.') { // Number: [0-9.]+ |
| 81 | std::string NumStr; |
| 82 | do { |
| 83 | NumStr += LastChar; |
| 84 | LastChar = getchar(); |
| 85 | } while (isdigit(LastChar) || LastChar == '.'); |
| 86 | |
| 87 | NumVal = strtod(NumStr.c_str(), 0); |
| 88 | return tok_number; |
| 89 | } |
| 90 | |
| 91 | if (LastChar == '#') { |
| 92 | // Comment until end of line. |
| 93 | do LastChar = getchar(); |
| 94 | while (LastChar != EOF && LastChar != '\n' && LastChar != '\r'); |
| 95 | |
| 96 | if (LastChar != EOF) |
| 97 | return gettok(); |
| 98 | } |
| 99 | |
| 100 | // Check for end of file. Don't eat the EOF. |
| 101 | if (LastChar == EOF) |
| 102 | return tok_eof; |
| 103 | |
| 104 | // Otherwise, just return the character as its ascii value. |
| 105 | int ThisChar = LastChar; |
| 106 | LastChar = getchar(); |
| 107 | return ThisChar; |
| 108 | } |
| 109 | |
| 110 | //===----------------------------------------------------------------------===// |
| 111 | // Abstract Syntax Tree (aka Parse Tree) |
| 112 | //===----------------------------------------------------------------------===// |
| 113 | |
| 114 | class IRGenContext; |
| 115 | |
| 116 | /// ExprAST - Base class for all expression nodes. |
| 117 | struct ExprAST { |
| 118 | virtual ~ExprAST() {} |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 119 | virtual Value *IRGen(IRGenContext &C) const = 0; |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 120 | }; |
| 121 | |
| 122 | /// NumberExprAST - Expression class for numeric literals like "1.0". |
| 123 | struct NumberExprAST : public ExprAST { |
| 124 | NumberExprAST(double Val) : Val(Val) {} |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 125 | Value *IRGen(IRGenContext &C) const override; |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 126 | |
| 127 | double Val; |
| 128 | }; |
| 129 | |
| 130 | /// VariableExprAST - Expression class for referencing a variable, like "a". |
| 131 | struct VariableExprAST : public ExprAST { |
| 132 | VariableExprAST(std::string Name) : Name(std::move(Name)) {} |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 133 | Value *IRGen(IRGenContext &C) const override; |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 134 | |
| 135 | std::string Name; |
| 136 | }; |
| 137 | |
| 138 | /// UnaryExprAST - Expression class for a unary operator. |
| 139 | struct UnaryExprAST : public ExprAST { |
| 140 | UnaryExprAST(char Opcode, std::unique_ptr<ExprAST> Operand) |
| 141 | : Opcode(std::move(Opcode)), Operand(std::move(Operand)) {} |
| 142 | |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 143 | Value *IRGen(IRGenContext &C) const override; |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 144 | |
| 145 | char Opcode; |
| 146 | std::unique_ptr<ExprAST> Operand; |
| 147 | }; |
| 148 | |
| 149 | /// BinaryExprAST - Expression class for a binary operator. |
| 150 | struct BinaryExprAST : public ExprAST { |
| 151 | BinaryExprAST(char Op, std::unique_ptr<ExprAST> LHS, |
| 152 | std::unique_ptr<ExprAST> RHS) |
| 153 | : Op(Op), LHS(std::move(LHS)), RHS(std::move(RHS)) {} |
| 154 | |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 155 | Value *IRGen(IRGenContext &C) const override; |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 156 | |
| 157 | char Op; |
| 158 | std::unique_ptr<ExprAST> LHS, RHS; |
| 159 | }; |
| 160 | |
| 161 | /// CallExprAST - Expression class for function calls. |
| 162 | struct CallExprAST : public ExprAST { |
| 163 | CallExprAST(std::string CalleeName, |
| 164 | std::vector<std::unique_ptr<ExprAST>> Args) |
| 165 | : CalleeName(std::move(CalleeName)), Args(std::move(Args)) {} |
| 166 | |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 167 | Value *IRGen(IRGenContext &C) const override; |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 168 | |
| 169 | std::string CalleeName; |
| 170 | std::vector<std::unique_ptr<ExprAST>> Args; |
| 171 | }; |
| 172 | |
| 173 | /// IfExprAST - Expression class for if/then/else. |
| 174 | struct IfExprAST : public ExprAST { |
| 175 | IfExprAST(std::unique_ptr<ExprAST> Cond, std::unique_ptr<ExprAST> Then, |
| 176 | std::unique_ptr<ExprAST> Else) |
| 177 | : Cond(std::move(Cond)), Then(std::move(Then)), Else(std::move(Else)) {} |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 178 | Value *IRGen(IRGenContext &C) const override; |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 179 | |
| 180 | std::unique_ptr<ExprAST> Cond, Then, Else; |
| 181 | }; |
| 182 | |
| 183 | /// ForExprAST - Expression class for for/in. |
| 184 | struct ForExprAST : public ExprAST { |
| 185 | ForExprAST(std::string VarName, std::unique_ptr<ExprAST> Start, |
| 186 | std::unique_ptr<ExprAST> End, std::unique_ptr<ExprAST> Step, |
| 187 | std::unique_ptr<ExprAST> Body) |
| 188 | : VarName(std::move(VarName)), Start(std::move(Start)), End(std::move(End)), |
| 189 | Step(std::move(Step)), Body(std::move(Body)) {} |
| 190 | |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 191 | Value *IRGen(IRGenContext &C) const override; |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 192 | |
| 193 | std::string VarName; |
| 194 | std::unique_ptr<ExprAST> Start, End, Step, Body; |
| 195 | }; |
| 196 | |
| 197 | /// VarExprAST - Expression class for var/in |
| 198 | struct VarExprAST : public ExprAST { |
| 199 | typedef std::pair<std::string, std::unique_ptr<ExprAST>> Binding; |
| 200 | typedef std::vector<Binding> BindingList; |
| 201 | |
| 202 | VarExprAST(BindingList VarBindings, std::unique_ptr<ExprAST> Body) |
| 203 | : VarBindings(std::move(VarBindings)), Body(std::move(Body)) {} |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 204 | |
| 205 | Value *IRGen(IRGenContext &C) const override; |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 206 | |
| 207 | BindingList VarBindings; |
| 208 | std::unique_ptr<ExprAST> Body; |
| 209 | }; |
| 210 | |
| 211 | /// PrototypeAST - This class represents the "prototype" for a function, |
| 212 | /// which captures its argument names as well as if it is an operator. |
| 213 | struct PrototypeAST { |
| 214 | PrototypeAST(std::string Name, std::vector<std::string> Args, |
| 215 | bool IsOperator = false, unsigned Precedence = 0) |
| 216 | : Name(std::move(Name)), Args(std::move(Args)), IsOperator(IsOperator), |
| 217 | Precedence(Precedence) {} |
| 218 | |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 219 | Function *IRGen(IRGenContext &C) const; |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 220 | void CreateArgumentAllocas(Function *F, IRGenContext &C); |
| 221 | |
| 222 | bool isUnaryOp() const { return IsOperator && Args.size() == 1; } |
| 223 | bool isBinaryOp() const { return IsOperator && Args.size() == 2; } |
| 224 | |
| 225 | char getOperatorName() const { |
| 226 | assert(isUnaryOp() || isBinaryOp()); |
| 227 | return Name[Name.size()-1]; |
| 228 | } |
| 229 | |
| 230 | std::string Name; |
| 231 | std::vector<std::string> Args; |
| 232 | bool IsOperator; |
| 233 | unsigned Precedence; // Precedence if a binary op. |
| 234 | }; |
| 235 | |
| 236 | /// FunctionAST - This class represents a function definition itself. |
| 237 | struct FunctionAST { |
| 238 | FunctionAST(std::unique_ptr<PrototypeAST> Proto, |
| 239 | std::unique_ptr<ExprAST> Body) |
| 240 | : Proto(std::move(Proto)), Body(std::move(Body)) {} |
| 241 | |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 242 | Function *IRGen(IRGenContext &C) const; |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 243 | |
| 244 | std::unique_ptr<PrototypeAST> Proto; |
| 245 | std::unique_ptr<ExprAST> Body; |
| 246 | }; |
| 247 | |
| 248 | //===----------------------------------------------------------------------===// |
| 249 | // Parser |
| 250 | //===----------------------------------------------------------------------===// |
| 251 | |
| 252 | /// CurTok/getNextToken - Provide a simple token buffer. CurTok is the current |
| 253 | /// token the parser is looking at. getNextToken reads another token from the |
| 254 | /// lexer and updates CurTok with its results. |
| 255 | static int CurTok; |
| 256 | static int getNextToken() { |
| 257 | return CurTok = gettok(); |
| 258 | } |
| 259 | |
| 260 | /// BinopPrecedence - This holds the precedence for each binary operator that is |
| 261 | /// defined. |
| 262 | static std::map<char, int> BinopPrecedence; |
| 263 | |
| 264 | /// GetTokPrecedence - Get the precedence of the pending binary operator token. |
| 265 | static int GetTokPrecedence() { |
| 266 | if (!isascii(CurTok)) |
| 267 | return -1; |
| 268 | |
| 269 | // Make sure it's a declared binop. |
| 270 | int TokPrec = BinopPrecedence[CurTok]; |
| 271 | if (TokPrec <= 0) return -1; |
| 272 | return TokPrec; |
| 273 | } |
| 274 | |
| 275 | template <typename T> |
Lang Hames | be9df34 | 2015-02-08 19:14:56 +0000 | [diff] [blame] | 276 | std::unique_ptr<T> ErrorU(const std::string &Str) { |
| 277 | std::cerr << "Error: " << Str << "\n"; |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 278 | return nullptr; |
| 279 | } |
| 280 | |
| 281 | template <typename T> |
Lang Hames | be9df34 | 2015-02-08 19:14:56 +0000 | [diff] [blame] | 282 | T* ErrorP(const std::string &Str) { |
| 283 | std::cerr << "Error: " << Str << "\n"; |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 284 | return nullptr; |
| 285 | } |
| 286 | |
| 287 | static std::unique_ptr<ExprAST> ParseExpression(); |
| 288 | |
| 289 | /// identifierexpr |
| 290 | /// ::= identifier |
| 291 | /// ::= identifier '(' expression* ')' |
| 292 | static std::unique_ptr<ExprAST> ParseIdentifierExpr() { |
| 293 | std::string IdName = IdentifierStr; |
| 294 | |
| 295 | getNextToken(); // eat identifier. |
| 296 | |
| 297 | if (CurTok != '(') // Simple variable ref. |
| 298 | return llvm::make_unique<VariableExprAST>(IdName); |
| 299 | |
| 300 | // Call. |
| 301 | getNextToken(); // eat ( |
| 302 | std::vector<std::unique_ptr<ExprAST>> Args; |
| 303 | if (CurTok != ')') { |
| 304 | while (1) { |
| 305 | auto Arg = ParseExpression(); |
| 306 | if (!Arg) return nullptr; |
| 307 | Args.push_back(std::move(Arg)); |
| 308 | |
| 309 | if (CurTok == ')') break; |
| 310 | |
| 311 | if (CurTok != ',') |
| 312 | return ErrorU<CallExprAST>("Expected ')' or ',' in argument list"); |
| 313 | getNextToken(); |
| 314 | } |
| 315 | } |
| 316 | |
| 317 | // Eat the ')'. |
| 318 | getNextToken(); |
| 319 | |
| 320 | return llvm::make_unique<CallExprAST>(IdName, std::move(Args)); |
| 321 | } |
| 322 | |
| 323 | /// numberexpr ::= number |
| 324 | static std::unique_ptr<NumberExprAST> ParseNumberExpr() { |
| 325 | auto Result = llvm::make_unique<NumberExprAST>(NumVal); |
| 326 | getNextToken(); // consume the number |
| 327 | return Result; |
| 328 | } |
| 329 | |
| 330 | /// parenexpr ::= '(' expression ')' |
| 331 | static std::unique_ptr<ExprAST> ParseParenExpr() { |
| 332 | getNextToken(); // eat (. |
| 333 | auto V = ParseExpression(); |
| 334 | if (!V) |
| 335 | return nullptr; |
| 336 | |
| 337 | if (CurTok != ')') |
| 338 | return ErrorU<ExprAST>("expected ')'"); |
| 339 | getNextToken(); // eat ). |
| 340 | return V; |
| 341 | } |
| 342 | |
| 343 | /// ifexpr ::= 'if' expression 'then' expression 'else' expression |
| 344 | static std::unique_ptr<ExprAST> ParseIfExpr() { |
| 345 | getNextToken(); // eat the if. |
| 346 | |
| 347 | // condition. |
| 348 | auto Cond = ParseExpression(); |
| 349 | if (!Cond) |
| 350 | return nullptr; |
| 351 | |
| 352 | if (CurTok != tok_then) |
| 353 | return ErrorU<ExprAST>("expected then"); |
| 354 | getNextToken(); // eat the then |
| 355 | |
| 356 | auto Then = ParseExpression(); |
| 357 | if (!Then) |
| 358 | return nullptr; |
| 359 | |
| 360 | if (CurTok != tok_else) |
| 361 | return ErrorU<ExprAST>("expected else"); |
| 362 | |
| 363 | getNextToken(); |
| 364 | |
| 365 | auto Else = ParseExpression(); |
| 366 | if (!Else) |
| 367 | return nullptr; |
| 368 | |
| 369 | return llvm::make_unique<IfExprAST>(std::move(Cond), std::move(Then), |
| 370 | std::move(Else)); |
| 371 | } |
| 372 | |
| 373 | /// forexpr ::= 'for' identifier '=' expr ',' expr (',' expr)? 'in' expression |
| 374 | static std::unique_ptr<ForExprAST> ParseForExpr() { |
| 375 | getNextToken(); // eat the for. |
| 376 | |
| 377 | if (CurTok != tok_identifier) |
| 378 | return ErrorU<ForExprAST>("expected identifier after for"); |
| 379 | |
| 380 | std::string IdName = IdentifierStr; |
| 381 | getNextToken(); // eat identifier. |
| 382 | |
| 383 | if (CurTok != '=') |
| 384 | return ErrorU<ForExprAST>("expected '=' after for"); |
| 385 | getNextToken(); // eat '='. |
| 386 | |
| 387 | |
| 388 | auto Start = ParseExpression(); |
| 389 | if (!Start) |
| 390 | return nullptr; |
| 391 | if (CurTok != ',') |
| 392 | return ErrorU<ForExprAST>("expected ',' after for start value"); |
| 393 | getNextToken(); |
| 394 | |
| 395 | auto End = ParseExpression(); |
| 396 | if (!End) |
| 397 | return nullptr; |
| 398 | |
| 399 | // The step value is optional. |
| 400 | std::unique_ptr<ExprAST> Step; |
| 401 | if (CurTok == ',') { |
| 402 | getNextToken(); |
| 403 | Step = ParseExpression(); |
| 404 | if (!Step) |
| 405 | return nullptr; |
| 406 | } |
| 407 | |
| 408 | if (CurTok != tok_in) |
| 409 | return ErrorU<ForExprAST>("expected 'in' after for"); |
| 410 | getNextToken(); // eat 'in'. |
| 411 | |
| 412 | auto Body = ParseExpression(); |
| 413 | if (Body) |
| 414 | return nullptr; |
| 415 | |
| 416 | return llvm::make_unique<ForExprAST>(IdName, std::move(Start), std::move(End), |
| 417 | std::move(Step), std::move(Body)); |
| 418 | } |
| 419 | |
| 420 | /// varexpr ::= 'var' identifier ('=' expression)? |
| 421 | // (',' identifier ('=' expression)?)* 'in' expression |
| 422 | static std::unique_ptr<VarExprAST> ParseVarExpr() { |
| 423 | getNextToken(); // eat the var. |
| 424 | |
| 425 | VarExprAST::BindingList VarBindings; |
| 426 | |
| 427 | // At least one variable name is required. |
| 428 | if (CurTok != tok_identifier) |
| 429 | return ErrorU<VarExprAST>("expected identifier after var"); |
| 430 | |
| 431 | while (1) { |
| 432 | std::string Name = IdentifierStr; |
| 433 | getNextToken(); // eat identifier. |
| 434 | |
| 435 | // Read the optional initializer. |
| 436 | std::unique_ptr<ExprAST> Init; |
| 437 | if (CurTok == '=') { |
| 438 | getNextToken(); // eat the '='. |
| 439 | |
| 440 | Init = ParseExpression(); |
| 441 | if (!Init) |
| 442 | return nullptr; |
| 443 | } |
| 444 | |
| 445 | VarBindings.push_back(VarExprAST::Binding(Name, std::move(Init))); |
| 446 | |
| 447 | // End of var list, exit loop. |
| 448 | if (CurTok != ',') break; |
| 449 | getNextToken(); // eat the ','. |
| 450 | |
| 451 | if (CurTok != tok_identifier) |
| 452 | return ErrorU<VarExprAST>("expected identifier list after var"); |
| 453 | } |
| 454 | |
| 455 | // At this point, we have to have 'in'. |
| 456 | if (CurTok != tok_in) |
| 457 | return ErrorU<VarExprAST>("expected 'in' keyword after 'var'"); |
| 458 | getNextToken(); // eat 'in'. |
| 459 | |
| 460 | auto Body = ParseExpression(); |
| 461 | if (!Body) |
| 462 | return nullptr; |
| 463 | |
| 464 | return llvm::make_unique<VarExprAST>(std::move(VarBindings), std::move(Body)); |
| 465 | } |
| 466 | |
| 467 | /// primary |
| 468 | /// ::= identifierexpr |
| 469 | /// ::= numberexpr |
| 470 | /// ::= parenexpr |
| 471 | /// ::= ifexpr |
| 472 | /// ::= forexpr |
| 473 | /// ::= varexpr |
| 474 | static std::unique_ptr<ExprAST> ParsePrimary() { |
| 475 | switch (CurTok) { |
| 476 | default: return ErrorU<ExprAST>("unknown token when expecting an expression"); |
| 477 | case tok_identifier: return ParseIdentifierExpr(); |
| 478 | case tok_number: return ParseNumberExpr(); |
| 479 | case '(': return ParseParenExpr(); |
| 480 | case tok_if: return ParseIfExpr(); |
| 481 | case tok_for: return ParseForExpr(); |
| 482 | case tok_var: return ParseVarExpr(); |
| 483 | } |
| 484 | } |
| 485 | |
| 486 | /// unary |
| 487 | /// ::= primary |
| 488 | /// ::= '!' unary |
| 489 | static std::unique_ptr<ExprAST> ParseUnary() { |
| 490 | // If the current token is not an operator, it must be a primary expr. |
| 491 | if (!isascii(CurTok) || CurTok == '(' || CurTok == ',') |
| 492 | return ParsePrimary(); |
| 493 | |
| 494 | // If this is a unary operator, read it. |
| 495 | int Opc = CurTok; |
| 496 | getNextToken(); |
| 497 | if (auto Operand = ParseUnary()) |
| 498 | return llvm::make_unique<UnaryExprAST>(Opc, std::move(Operand)); |
| 499 | return nullptr; |
| 500 | } |
| 501 | |
| 502 | /// binoprhs |
| 503 | /// ::= ('+' unary)* |
| 504 | static std::unique_ptr<ExprAST> ParseBinOpRHS(int ExprPrec, |
| 505 | std::unique_ptr<ExprAST> LHS) { |
| 506 | // If this is a binop, find its precedence. |
| 507 | while (1) { |
| 508 | int TokPrec = GetTokPrecedence(); |
| 509 | |
| 510 | // If this is a binop that binds at least as tightly as the current binop, |
| 511 | // consume it, otherwise we are done. |
| 512 | if (TokPrec < ExprPrec) |
| 513 | return LHS; |
| 514 | |
| 515 | // Okay, we know this is a binop. |
| 516 | int BinOp = CurTok; |
| 517 | getNextToken(); // eat binop |
| 518 | |
| 519 | // Parse the unary expression after the binary operator. |
| 520 | auto RHS = ParseUnary(); |
| 521 | if (!RHS) |
| 522 | return nullptr; |
| 523 | |
| 524 | // If BinOp binds less tightly with RHS than the operator after RHS, let |
| 525 | // the pending operator take RHS as its LHS. |
| 526 | int NextPrec = GetTokPrecedence(); |
| 527 | if (TokPrec < NextPrec) { |
| 528 | RHS = ParseBinOpRHS(TokPrec+1, std::move(RHS)); |
| 529 | if (!RHS) |
| 530 | return nullptr; |
| 531 | } |
| 532 | |
| 533 | // Merge LHS/RHS. |
| 534 | LHS = llvm::make_unique<BinaryExprAST>(BinOp, std::move(LHS), std::move(RHS)); |
| 535 | } |
| 536 | } |
| 537 | |
| 538 | /// expression |
| 539 | /// ::= unary binoprhs |
| 540 | /// |
| 541 | static std::unique_ptr<ExprAST> ParseExpression() { |
| 542 | auto LHS = ParseUnary(); |
| 543 | if (!LHS) |
| 544 | return nullptr; |
| 545 | |
| 546 | return ParseBinOpRHS(0, std::move(LHS)); |
| 547 | } |
| 548 | |
| 549 | /// prototype |
| 550 | /// ::= id '(' id* ')' |
| 551 | /// ::= binary LETTER number? (id, id) |
| 552 | /// ::= unary LETTER (id) |
| 553 | static std::unique_ptr<PrototypeAST> ParsePrototype() { |
| 554 | std::string FnName; |
| 555 | |
| 556 | unsigned Kind = 0; // 0 = identifier, 1 = unary, 2 = binary. |
| 557 | unsigned BinaryPrecedence = 30; |
| 558 | |
| 559 | switch (CurTok) { |
| 560 | default: |
| 561 | return ErrorU<PrototypeAST>("Expected function name in prototype"); |
| 562 | case tok_identifier: |
| 563 | FnName = IdentifierStr; |
| 564 | Kind = 0; |
| 565 | getNextToken(); |
| 566 | break; |
| 567 | case tok_unary: |
| 568 | getNextToken(); |
| 569 | if (!isascii(CurTok)) |
| 570 | return ErrorU<PrototypeAST>("Expected unary operator"); |
| 571 | FnName = "unary"; |
| 572 | FnName += (char)CurTok; |
| 573 | Kind = 1; |
| 574 | getNextToken(); |
| 575 | break; |
| 576 | case tok_binary: |
| 577 | getNextToken(); |
| 578 | if (!isascii(CurTok)) |
| 579 | return ErrorU<PrototypeAST>("Expected binary operator"); |
| 580 | FnName = "binary"; |
| 581 | FnName += (char)CurTok; |
| 582 | Kind = 2; |
| 583 | getNextToken(); |
| 584 | |
| 585 | // Read the precedence if present. |
| 586 | if (CurTok == tok_number) { |
| 587 | if (NumVal < 1 || NumVal > 100) |
| 588 | return ErrorU<PrototypeAST>("Invalid precedecnce: must be 1..100"); |
| 589 | BinaryPrecedence = (unsigned)NumVal; |
| 590 | getNextToken(); |
| 591 | } |
| 592 | break; |
| 593 | } |
| 594 | |
| 595 | if (CurTok != '(') |
| 596 | return ErrorU<PrototypeAST>("Expected '(' in prototype"); |
| 597 | |
| 598 | std::vector<std::string> ArgNames; |
| 599 | while (getNextToken() == tok_identifier) |
| 600 | ArgNames.push_back(IdentifierStr); |
| 601 | if (CurTok != ')') |
| 602 | return ErrorU<PrototypeAST>("Expected ')' in prototype"); |
| 603 | |
| 604 | // success. |
| 605 | getNextToken(); // eat ')'. |
| 606 | |
| 607 | // Verify right number of names for operator. |
| 608 | if (Kind && ArgNames.size() != Kind) |
| 609 | return ErrorU<PrototypeAST>("Invalid number of operands for operator"); |
| 610 | |
| 611 | return llvm::make_unique<PrototypeAST>(FnName, std::move(ArgNames), Kind != 0, |
| 612 | BinaryPrecedence); |
| 613 | } |
| 614 | |
| 615 | /// definition ::= 'def' prototype expression |
| 616 | static std::unique_ptr<FunctionAST> ParseDefinition() { |
| 617 | getNextToken(); // eat def. |
| 618 | auto Proto = ParsePrototype(); |
| 619 | if (!Proto) |
| 620 | return nullptr; |
| 621 | |
| 622 | if (auto Body = ParseExpression()) |
| 623 | return llvm::make_unique<FunctionAST>(std::move(Proto), std::move(Body)); |
| 624 | return nullptr; |
| 625 | } |
| 626 | |
| 627 | /// toplevelexpr ::= expression |
| 628 | static std::unique_ptr<FunctionAST> ParseTopLevelExpr() { |
| 629 | if (auto E = ParseExpression()) { |
| 630 | // Make an anonymous proto. |
| 631 | auto Proto = |
| 632 | llvm::make_unique<PrototypeAST>("__anon_expr", std::vector<std::string>()); |
| 633 | return llvm::make_unique<FunctionAST>(std::move(Proto), std::move(E)); |
| 634 | } |
| 635 | return nullptr; |
| 636 | } |
| 637 | |
| 638 | /// external ::= 'extern' prototype |
| 639 | static std::unique_ptr<PrototypeAST> ParseExtern() { |
| 640 | getNextToken(); // eat extern. |
| 641 | return ParsePrototype(); |
| 642 | } |
| 643 | |
| 644 | //===----------------------------------------------------------------------===// |
| 645 | // Code Generation |
| 646 | //===----------------------------------------------------------------------===// |
| 647 | |
| 648 | // FIXME: Obviously we can do better than this |
Lang Hames | be9df34 | 2015-02-08 19:14:56 +0000 | [diff] [blame] | 649 | std::string GenerateUniqueName(const std::string &Root) { |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 650 | static int i = 0; |
Lang Hames | be9df34 | 2015-02-08 19:14:56 +0000 | [diff] [blame] | 651 | std::ostringstream NameStream; |
| 652 | NameStream << Root << ++i; |
| 653 | return NameStream.str(); |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 654 | } |
| 655 | |
| 656 | std::string MakeLegalFunctionName(std::string Name) |
| 657 | { |
| 658 | std::string NewName; |
| 659 | assert(!Name.empty() && "Base name must not be empty"); |
| 660 | |
| 661 | // Start with what we have |
| 662 | NewName = Name; |
| 663 | |
| 664 | // Look for a numberic first character |
| 665 | if (NewName.find_first_of("0123456789") == 0) { |
| 666 | NewName.insert(0, 1, 'n'); |
| 667 | } |
| 668 | |
| 669 | // Replace illegal characters with their ASCII equivalent |
| 670 | std::string legal_elements = "_abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789"; |
| 671 | size_t pos; |
| 672 | while ((pos = NewName.find_first_not_of(legal_elements)) != std::string::npos) { |
Lang Hames | be9df34 | 2015-02-08 19:14:56 +0000 | [diff] [blame] | 673 | std::ostringstream NumStream; |
| 674 | NumStream << (int)NewName.at(pos); |
| 675 | NewName = NewName.replace(pos, 1, NumStream.str()); |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 676 | } |
| 677 | |
| 678 | return NewName; |
| 679 | } |
| 680 | |
| 681 | class SessionContext { |
| 682 | public: |
| 683 | SessionContext(LLVMContext &C) : Context(C) {} |
| 684 | LLVMContext& getLLVMContext() const { return Context; } |
| 685 | void addPrototypeAST(std::unique_ptr<PrototypeAST> P); |
| 686 | PrototypeAST* getPrototypeAST(const std::string &Name); |
| 687 | private: |
| 688 | typedef std::map<std::string, std::unique_ptr<PrototypeAST>> PrototypeMap; |
| 689 | LLVMContext &Context; |
| 690 | PrototypeMap Prototypes; |
| 691 | }; |
| 692 | |
| 693 | void SessionContext::addPrototypeAST(std::unique_ptr<PrototypeAST> P) { |
| 694 | Prototypes[P->Name] = std::move(P); |
| 695 | } |
| 696 | |
| 697 | PrototypeAST* SessionContext::getPrototypeAST(const std::string &Name) { |
| 698 | PrototypeMap::iterator I = Prototypes.find(Name); |
| 699 | if (I != Prototypes.end()) |
| 700 | return I->second.get(); |
| 701 | return nullptr; |
Lang Hames | be9df34 | 2015-02-08 19:14:56 +0000 | [diff] [blame] | 702 | } |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 703 | |
| 704 | class IRGenContext { |
| 705 | public: |
| 706 | |
| 707 | IRGenContext(SessionContext &S) |
| 708 | : Session(S), |
| 709 | M(new Module(GenerateUniqueName("jit_module_"), |
| 710 | Session.getLLVMContext())), |
| 711 | Builder(Session.getLLVMContext()) {} |
| 712 | |
| 713 | SessionContext& getSession() { return Session; } |
| 714 | Module& getM() const { return *M; } |
| 715 | std::unique_ptr<Module> takeM() { return std::move(M); } |
| 716 | IRBuilder<>& getBuilder() { return Builder; } |
| 717 | LLVMContext& getLLVMContext() { return Session.getLLVMContext(); } |
| 718 | Function* getPrototype(const std::string &Name); |
| 719 | |
| 720 | std::map<std::string, AllocaInst*> NamedValues; |
| 721 | private: |
| 722 | SessionContext &Session; |
| 723 | std::unique_ptr<Module> M; |
| 724 | IRBuilder<> Builder; |
| 725 | }; |
| 726 | |
| 727 | Function* IRGenContext::getPrototype(const std::string &Name) { |
| 728 | if (Function *ExistingProto = M->getFunction(Name)) |
| 729 | return ExistingProto; |
| 730 | if (PrototypeAST *ProtoAST = Session.getPrototypeAST(Name)) |
| 731 | return ProtoAST->IRGen(*this); |
| 732 | return nullptr; |
| 733 | } |
| 734 | |
| 735 | /// CreateEntryBlockAlloca - Create an alloca instruction in the entry block of |
| 736 | /// the function. This is used for mutable variables etc. |
| 737 | static AllocaInst *CreateEntryBlockAlloca(Function *TheFunction, |
| 738 | const std::string &VarName) { |
| 739 | IRBuilder<> TmpB(&TheFunction->getEntryBlock(), |
| 740 | TheFunction->getEntryBlock().begin()); |
| 741 | return TmpB.CreateAlloca(Type::getDoubleTy(getGlobalContext()), 0, |
| 742 | VarName.c_str()); |
| 743 | } |
| 744 | |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 745 | Value *NumberExprAST::IRGen(IRGenContext &C) const { |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 746 | return ConstantFP::get(C.getLLVMContext(), APFloat(Val)); |
| 747 | } |
| 748 | |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 749 | Value *VariableExprAST::IRGen(IRGenContext &C) const { |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 750 | // Look this variable up in the function. |
| 751 | Value *V = C.NamedValues[Name]; |
| 752 | |
Lang Hames | be9df34 | 2015-02-08 19:14:56 +0000 | [diff] [blame] | 753 | if (V == 0) |
| 754 | return ErrorP<Value>("Unknown variable name '" + Name + "'"); |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 755 | |
| 756 | // Load the value. |
| 757 | return C.getBuilder().CreateLoad(V, Name.c_str()); |
| 758 | } |
| 759 | |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 760 | Value *UnaryExprAST::IRGen(IRGenContext &C) const { |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 761 | if (Value *OperandV = Operand->IRGen(C)) { |
| 762 | std::string FnName = MakeLegalFunctionName(std::string("unary")+Opcode); |
| 763 | if (Function *F = C.getPrototype(FnName)) |
| 764 | return C.getBuilder().CreateCall(F, OperandV, "unop"); |
| 765 | return ErrorP<Value>("Unknown unary operator"); |
| 766 | } |
| 767 | |
| 768 | // Could not codegen operand - return null. |
| 769 | return nullptr; |
| 770 | } |
| 771 | |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 772 | Value *BinaryExprAST::IRGen(IRGenContext &C) const { |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 773 | // Special case '=' because we don't want to emit the LHS as an expression. |
| 774 | if (Op == '=') { |
| 775 | // Assignment requires the LHS to be an identifier. |
| 776 | auto LHSVar = static_cast<VariableExprAST&>(*LHS); |
| 777 | // Codegen the RHS. |
| 778 | Value *Val = RHS->IRGen(C); |
| 779 | if (!Val) return nullptr; |
| 780 | |
| 781 | // Look up the name. |
| 782 | if (auto Variable = C.NamedValues[LHSVar.Name]) { |
| 783 | C.getBuilder().CreateStore(Val, Variable); |
| 784 | return Val; |
| 785 | } |
| 786 | return ErrorP<Value>("Unknown variable name"); |
| 787 | } |
| 788 | |
| 789 | Value *L = LHS->IRGen(C); |
| 790 | Value *R = RHS->IRGen(C); |
| 791 | if (!L || !R) return nullptr; |
| 792 | |
| 793 | switch (Op) { |
| 794 | case '+': return C.getBuilder().CreateFAdd(L, R, "addtmp"); |
| 795 | case '-': return C.getBuilder().CreateFSub(L, R, "subtmp"); |
| 796 | case '*': return C.getBuilder().CreateFMul(L, R, "multmp"); |
| 797 | case '/': return C.getBuilder().CreateFDiv(L, R, "divtmp"); |
| 798 | case '<': |
| 799 | L = C.getBuilder().CreateFCmpULT(L, R, "cmptmp"); |
| 800 | // Convert bool 0/1 to double 0.0 or 1.0 |
| 801 | return C.getBuilder().CreateUIToFP(L, Type::getDoubleTy(getGlobalContext()), |
| 802 | "booltmp"); |
| 803 | default: break; |
| 804 | } |
| 805 | |
| 806 | // If it wasn't a builtin binary operator, it must be a user defined one. Emit |
| 807 | // a call to it. |
| 808 | std::string FnName = MakeLegalFunctionName(std::string("binary")+Op); |
| 809 | if (Function *F = C.getPrototype(FnName)) { |
| 810 | Value *Ops[] = { L, R }; |
| 811 | return C.getBuilder().CreateCall(F, Ops, "binop"); |
| 812 | } |
| 813 | |
| 814 | return ErrorP<Value>("Unknown binary operator"); |
| 815 | } |
| 816 | |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 817 | Value *CallExprAST::IRGen(IRGenContext &C) const { |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 818 | // Look up the name in the global module table. |
| 819 | if (auto CalleeF = C.getPrototype(CalleeName)) { |
| 820 | // If argument mismatch error. |
| 821 | if (CalleeF->arg_size() != Args.size()) |
| 822 | return ErrorP<Value>("Incorrect # arguments passed"); |
| 823 | |
| 824 | std::vector<Value*> ArgsV; |
| 825 | for (unsigned i = 0, e = Args.size(); i != e; ++i) { |
| 826 | ArgsV.push_back(Args[i]->IRGen(C)); |
| 827 | if (!ArgsV.back()) return nullptr; |
| 828 | } |
| 829 | |
| 830 | return C.getBuilder().CreateCall(CalleeF, ArgsV, "calltmp"); |
| 831 | } |
| 832 | |
| 833 | return ErrorP<Value>("Unknown function referenced"); |
| 834 | } |
| 835 | |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 836 | Value *IfExprAST::IRGen(IRGenContext &C) const { |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 837 | Value *CondV = Cond->IRGen(C); |
| 838 | if (!CondV) return nullptr; |
| 839 | |
| 840 | // Convert condition to a bool by comparing equal to 0.0. |
| 841 | ConstantFP *FPZero = |
| 842 | ConstantFP::get(C.getLLVMContext(), APFloat(0.0)); |
| 843 | CondV = C.getBuilder().CreateFCmpONE(CondV, FPZero, "ifcond"); |
| 844 | |
| 845 | Function *TheFunction = C.getBuilder().GetInsertBlock()->getParent(); |
| 846 | |
| 847 | // Create blocks for the then and else cases. Insert the 'then' block at the |
| 848 | // end of the function. |
| 849 | BasicBlock *ThenBB = BasicBlock::Create(C.getLLVMContext(), "then", TheFunction); |
| 850 | BasicBlock *ElseBB = BasicBlock::Create(C.getLLVMContext(), "else"); |
| 851 | BasicBlock *MergeBB = BasicBlock::Create(C.getLLVMContext(), "ifcont"); |
| 852 | |
| 853 | C.getBuilder().CreateCondBr(CondV, ThenBB, ElseBB); |
| 854 | |
| 855 | // Emit then value. |
| 856 | C.getBuilder().SetInsertPoint(ThenBB); |
| 857 | |
| 858 | Value *ThenV = Then->IRGen(C); |
| 859 | if (!ThenV) return nullptr; |
| 860 | |
| 861 | C.getBuilder().CreateBr(MergeBB); |
| 862 | // Codegen of 'Then' can change the current block, update ThenBB for the PHI. |
| 863 | ThenBB = C.getBuilder().GetInsertBlock(); |
| 864 | |
| 865 | // Emit else block. |
| 866 | TheFunction->getBasicBlockList().push_back(ElseBB); |
| 867 | C.getBuilder().SetInsertPoint(ElseBB); |
| 868 | |
| 869 | Value *ElseV = Else->IRGen(C); |
| 870 | if (!ElseV) return nullptr; |
| 871 | |
| 872 | C.getBuilder().CreateBr(MergeBB); |
| 873 | // Codegen of 'Else' can change the current block, update ElseBB for the PHI. |
| 874 | ElseBB = C.getBuilder().GetInsertBlock(); |
| 875 | |
| 876 | // Emit merge block. |
| 877 | TheFunction->getBasicBlockList().push_back(MergeBB); |
| 878 | C.getBuilder().SetInsertPoint(MergeBB); |
| 879 | PHINode *PN = C.getBuilder().CreatePHI(Type::getDoubleTy(getGlobalContext()), 2, |
| 880 | "iftmp"); |
| 881 | |
| 882 | PN->addIncoming(ThenV, ThenBB); |
| 883 | PN->addIncoming(ElseV, ElseBB); |
| 884 | return PN; |
| 885 | } |
| 886 | |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 887 | Value *ForExprAST::IRGen(IRGenContext &C) const { |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 888 | // Output this as: |
| 889 | // var = alloca double |
| 890 | // ... |
| 891 | // start = startexpr |
| 892 | // store start -> var |
| 893 | // goto loop |
| 894 | // loop: |
| 895 | // ... |
| 896 | // bodyexpr |
| 897 | // ... |
| 898 | // loopend: |
| 899 | // step = stepexpr |
| 900 | // endcond = endexpr |
| 901 | // |
| 902 | // curvar = load var |
| 903 | // nextvar = curvar + step |
| 904 | // store nextvar -> var |
| 905 | // br endcond, loop, endloop |
| 906 | // outloop: |
| 907 | |
| 908 | Function *TheFunction = C.getBuilder().GetInsertBlock()->getParent(); |
| 909 | |
| 910 | // Create an alloca for the variable in the entry block. |
| 911 | AllocaInst *Alloca = CreateEntryBlockAlloca(TheFunction, VarName); |
| 912 | |
| 913 | // Emit the start code first, without 'variable' in scope. |
| 914 | Value *StartVal = Start->IRGen(C); |
| 915 | if (!StartVal) return nullptr; |
| 916 | |
| 917 | // Store the value into the alloca. |
| 918 | C.getBuilder().CreateStore(StartVal, Alloca); |
| 919 | |
| 920 | // Make the new basic block for the loop header, inserting after current |
| 921 | // block. |
| 922 | BasicBlock *LoopBB = BasicBlock::Create(getGlobalContext(), "loop", TheFunction); |
| 923 | |
| 924 | // Insert an explicit fall through from the current block to the LoopBB. |
| 925 | C.getBuilder().CreateBr(LoopBB); |
| 926 | |
| 927 | // Start insertion in LoopBB. |
| 928 | C.getBuilder().SetInsertPoint(LoopBB); |
| 929 | |
| 930 | // Within the loop, the variable is defined equal to the PHI node. If it |
| 931 | // shadows an existing variable, we have to restore it, so save it now. |
| 932 | AllocaInst *OldVal = C.NamedValues[VarName]; |
| 933 | C.NamedValues[VarName] = Alloca; |
| 934 | |
| 935 | // Emit the body of the loop. This, like any other expr, can change the |
| 936 | // current BB. Note that we ignore the value computed by the body, but don't |
| 937 | // allow an error. |
| 938 | if (!Body->IRGen(C)) |
| 939 | return nullptr; |
| 940 | |
| 941 | // Emit the step value. |
| 942 | Value *StepVal; |
| 943 | if (Step) { |
| 944 | StepVal = Step->IRGen(C); |
| 945 | if (!StepVal) return nullptr; |
| 946 | } else { |
| 947 | // If not specified, use 1.0. |
| 948 | StepVal = ConstantFP::get(getGlobalContext(), APFloat(1.0)); |
| 949 | } |
| 950 | |
| 951 | // Compute the end condition. |
| 952 | Value *EndCond = End->IRGen(C); |
| 953 | if (EndCond == 0) return EndCond; |
| 954 | |
| 955 | // Reload, increment, and restore the alloca. This handles the case where |
| 956 | // the body of the loop mutates the variable. |
| 957 | Value *CurVar = C.getBuilder().CreateLoad(Alloca, VarName.c_str()); |
| 958 | Value *NextVar = C.getBuilder().CreateFAdd(CurVar, StepVal, "nextvar"); |
| 959 | C.getBuilder().CreateStore(NextVar, Alloca); |
| 960 | |
| 961 | // Convert condition to a bool by comparing equal to 0.0. |
| 962 | EndCond = C.getBuilder().CreateFCmpONE(EndCond, |
| 963 | ConstantFP::get(getGlobalContext(), APFloat(0.0)), |
| 964 | "loopcond"); |
| 965 | |
| 966 | // Create the "after loop" block and insert it. |
| 967 | BasicBlock *AfterBB = BasicBlock::Create(getGlobalContext(), "afterloop", TheFunction); |
| 968 | |
| 969 | // Insert the conditional branch into the end of LoopEndBB. |
| 970 | C.getBuilder().CreateCondBr(EndCond, LoopBB, AfterBB); |
| 971 | |
| 972 | // Any new code will be inserted in AfterBB. |
| 973 | C.getBuilder().SetInsertPoint(AfterBB); |
| 974 | |
| 975 | // Restore the unshadowed variable. |
| 976 | if (OldVal) |
| 977 | C.NamedValues[VarName] = OldVal; |
| 978 | else |
| 979 | C.NamedValues.erase(VarName); |
| 980 | |
| 981 | |
| 982 | // for expr always returns 0.0. |
| 983 | return Constant::getNullValue(Type::getDoubleTy(getGlobalContext())); |
| 984 | } |
| 985 | |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 986 | Value *VarExprAST::IRGen(IRGenContext &C) const { |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 987 | std::vector<AllocaInst *> OldBindings; |
| 988 | |
| 989 | Function *TheFunction = C.getBuilder().GetInsertBlock()->getParent(); |
| 990 | |
| 991 | // Register all variables and emit their initializer. |
| 992 | for (unsigned i = 0, e = VarBindings.size(); i != e; ++i) { |
| 993 | auto &VarName = VarBindings[i].first; |
| 994 | auto &Init = VarBindings[i].second; |
| 995 | |
| 996 | // Emit the initializer before adding the variable to scope, this prevents |
| 997 | // the initializer from referencing the variable itself, and permits stuff |
| 998 | // like this: |
| 999 | // var a = 1 in |
| 1000 | // var a = a in ... # refers to outer 'a'. |
| 1001 | Value *InitVal; |
| 1002 | if (Init) { |
| 1003 | InitVal = Init->IRGen(C); |
| 1004 | if (!InitVal) return nullptr; |
| 1005 | } else // If not specified, use 0.0. |
| 1006 | InitVal = ConstantFP::get(getGlobalContext(), APFloat(0.0)); |
| 1007 | |
| 1008 | AllocaInst *Alloca = CreateEntryBlockAlloca(TheFunction, VarName); |
| 1009 | C.getBuilder().CreateStore(InitVal, Alloca); |
| 1010 | |
| 1011 | // Remember the old variable binding so that we can restore the binding when |
| 1012 | // we unrecurse. |
| 1013 | OldBindings.push_back(C.NamedValues[VarName]); |
| 1014 | |
| 1015 | // Remember this binding. |
| 1016 | C.NamedValues[VarName] = Alloca; |
| 1017 | } |
| 1018 | |
| 1019 | // Codegen the body, now that all vars are in scope. |
| 1020 | Value *BodyVal = Body->IRGen(C); |
| 1021 | if (!BodyVal) return nullptr; |
| 1022 | |
| 1023 | // Pop all our variables from scope. |
| 1024 | for (unsigned i = 0, e = VarBindings.size(); i != e; ++i) |
| 1025 | C.NamedValues[VarBindings[i].first] = OldBindings[i]; |
| 1026 | |
| 1027 | // Return the body computation. |
| 1028 | return BodyVal; |
| 1029 | } |
| 1030 | |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 1031 | Function *PrototypeAST::IRGen(IRGenContext &C) const { |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 1032 | std::string FnName = MakeLegalFunctionName(Name); |
| 1033 | |
| 1034 | // Make the function type: double(double,double) etc. |
| 1035 | std::vector<Type*> Doubles(Args.size(), |
| 1036 | Type::getDoubleTy(getGlobalContext())); |
| 1037 | FunctionType *FT = FunctionType::get(Type::getDoubleTy(getGlobalContext()), |
| 1038 | Doubles, false); |
| 1039 | Function *F = Function::Create(FT, Function::ExternalLinkage, FnName, |
| 1040 | &C.getM()); |
| 1041 | |
| 1042 | // If F conflicted, there was already something named 'FnName'. If it has a |
| 1043 | // body, don't allow redefinition or reextern. |
| 1044 | if (F->getName() != FnName) { |
| 1045 | // Delete the one we just made and get the existing one. |
| 1046 | F->eraseFromParent(); |
| 1047 | F = C.getM().getFunction(Name); |
| 1048 | |
| 1049 | // If F already has a body, reject this. |
| 1050 | if (!F->empty()) { |
| 1051 | ErrorP<Function>("redefinition of function"); |
| 1052 | return nullptr; |
| 1053 | } |
| 1054 | |
| 1055 | // If F took a different number of args, reject. |
| 1056 | if (F->arg_size() != Args.size()) { |
| 1057 | ErrorP<Function>("redefinition of function with different # args"); |
| 1058 | return nullptr; |
| 1059 | } |
| 1060 | } |
| 1061 | |
| 1062 | // Set names for all arguments. |
| 1063 | unsigned Idx = 0; |
| 1064 | for (Function::arg_iterator AI = F->arg_begin(); Idx != Args.size(); |
| 1065 | ++AI, ++Idx) |
| 1066 | AI->setName(Args[Idx]); |
| 1067 | |
| 1068 | return F; |
| 1069 | } |
| 1070 | |
| 1071 | /// CreateArgumentAllocas - Create an alloca for each argument and register the |
| 1072 | /// argument in the symbol table so that references to it will succeed. |
| 1073 | void PrototypeAST::CreateArgumentAllocas(Function *F, IRGenContext &C) { |
| 1074 | Function::arg_iterator AI = F->arg_begin(); |
| 1075 | for (unsigned Idx = 0, e = Args.size(); Idx != e; ++Idx, ++AI) { |
| 1076 | // Create an alloca for this variable. |
| 1077 | AllocaInst *Alloca = CreateEntryBlockAlloca(F, Args[Idx]); |
| 1078 | |
| 1079 | // Store the initial value into the alloca. |
| 1080 | C.getBuilder().CreateStore(AI, Alloca); |
| 1081 | |
| 1082 | // Add arguments to variable symbol table. |
| 1083 | C.NamedValues[Args[Idx]] = Alloca; |
| 1084 | } |
| 1085 | } |
| 1086 | |
David Blaikie | 055811e | 2015-02-08 20:15:01 +0000 | [diff] [blame] | 1087 | Function *FunctionAST::IRGen(IRGenContext &C) const { |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 1088 | C.NamedValues.clear(); |
| 1089 | |
| 1090 | Function *TheFunction = Proto->IRGen(C); |
| 1091 | if (!TheFunction) |
| 1092 | return nullptr; |
| 1093 | |
| 1094 | // If this is an operator, install it. |
| 1095 | if (Proto->isBinaryOp()) |
| 1096 | BinopPrecedence[Proto->getOperatorName()] = Proto->Precedence; |
| 1097 | |
| 1098 | // Create a new basic block to start insertion into. |
| 1099 | BasicBlock *BB = BasicBlock::Create(getGlobalContext(), "entry", TheFunction); |
| 1100 | C.getBuilder().SetInsertPoint(BB); |
| 1101 | |
| 1102 | // Add all arguments to the symbol table and create their allocas. |
| 1103 | Proto->CreateArgumentAllocas(TheFunction, C); |
| 1104 | |
| 1105 | if (Value *RetVal = Body->IRGen(C)) { |
| 1106 | // Finish off the function. |
| 1107 | C.getBuilder().CreateRet(RetVal); |
| 1108 | |
| 1109 | // Validate the generated code, checking for consistency. |
| 1110 | verifyFunction(*TheFunction); |
| 1111 | |
| 1112 | return TheFunction; |
| 1113 | } |
| 1114 | |
| 1115 | // Error reading body, remove function. |
| 1116 | TheFunction->eraseFromParent(); |
| 1117 | |
| 1118 | if (Proto->isBinaryOp()) |
| 1119 | BinopPrecedence.erase(Proto->getOperatorName()); |
| 1120 | return nullptr; |
| 1121 | } |
| 1122 | |
| 1123 | //===----------------------------------------------------------------------===// |
| 1124 | // Top-Level parsing and JIT Driver |
| 1125 | //===----------------------------------------------------------------------===// |
| 1126 | |
| 1127 | class KaleidoscopeJIT { |
| 1128 | public: |
| 1129 | typedef ObjectLinkingLayer<> ObjLayerT; |
| 1130 | typedef IRCompileLayer<ObjLayerT> CompileLayerT; |
| 1131 | |
| 1132 | typedef CompileLayerT::ModuleSetHandleT ModuleHandleT; |
| 1133 | |
| 1134 | KaleidoscopeJIT() |
| 1135 | : TM(EngineBuilder().selectTarget()), |
| 1136 | Mang(TM->getDataLayout()), |
| 1137 | CompileLayer(ObjectLayer, SimpleCompiler(*TM)) {} |
| 1138 | |
| 1139 | ModuleHandleT addModule(std::unique_ptr<Module> M) { |
| 1140 | if (!M->getDataLayout()) |
| 1141 | M->setDataLayout(TM->getDataLayout()); |
| 1142 | |
| 1143 | // The LazyEmitLayer takes lists of modules, rather than single modules, so |
| 1144 | // we'll just build a single-element list. |
| 1145 | std::vector<std::unique_ptr<Module>> S; |
| 1146 | S.push_back(std::move(M)); |
| 1147 | |
| 1148 | // We need a memory manager to allocate memory and resolve symbols for this |
| 1149 | // new module. Create one that resolves symbols by looking back into the JIT. |
| 1150 | auto MM = createLookasideRTDyldMM<SectionMemoryManager>( |
| 1151 | [&](const std::string &S) { |
Lang Hames | dac276e | 2015-02-08 04:34:13 +0000 | [diff] [blame] | 1152 | return getMangledSymbolAddress(S); |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 1153 | }, |
| 1154 | [](const std::string &S) { return 0; } ); |
| 1155 | |
| 1156 | return CompileLayer.addModuleSet(std::move(S), std::move(MM)); |
| 1157 | } |
| 1158 | |
| 1159 | void removeModule(ModuleHandleT H) { CompileLayer.removeModuleSet(H); } |
| 1160 | |
Lang Hames | dac276e | 2015-02-08 04:34:13 +0000 | [diff] [blame] | 1161 | uint64_t getMangledSymbolAddress(const std::string &Name) { |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 1162 | return CompileLayer.getSymbolAddress(Name, false); |
| 1163 | } |
| 1164 | |
| 1165 | uint64_t getSymbolAddress(const std::string Name) { |
| 1166 | std::string MangledName; |
| 1167 | { |
| 1168 | raw_string_ostream MangledNameStream(MangledName); |
| 1169 | Mang.getNameWithPrefix(MangledNameStream, Name); |
| 1170 | } |
Lang Hames | dac276e | 2015-02-08 04:34:13 +0000 | [diff] [blame] | 1171 | return getMangledSymbolAddress(MangledName); |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 1172 | } |
| 1173 | |
| 1174 | private: |
| 1175 | |
| 1176 | std::unique_ptr<TargetMachine> TM; |
| 1177 | Mangler Mang; |
| 1178 | |
| 1179 | ObjLayerT ObjectLayer; |
| 1180 | CompileLayerT CompileLayer; |
| 1181 | }; |
| 1182 | |
David Blaikie | 9c4c23b | 2015-02-08 21:03:30 +0000 | [diff] [blame^] | 1183 | static std::unique_ptr<llvm::Module> IRGen(SessionContext &S, |
| 1184 | const FunctionAST &F) { |
David Blaikie | 1803dc2 | 2015-02-08 20:29:28 +0000 | [diff] [blame] | 1185 | IRGenContext C(S); |
| 1186 | auto LF = F.IRGen(C); |
| 1187 | if (!LF) |
| 1188 | return nullptr; |
| 1189 | #ifndef MINIMAL_STDERR_OUTPUT |
| 1190 | fprintf(stderr, "Read function definition:"); |
| 1191 | LF->dump(); |
| 1192 | #endif |
| 1193 | return C.takeM(); |
| 1194 | } |
| 1195 | |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 1196 | static void HandleDefinition(SessionContext &S, KaleidoscopeJIT &J) { |
| 1197 | if (auto F = ParseDefinition()) { |
David Blaikie | 9c4c23b | 2015-02-08 21:03:30 +0000 | [diff] [blame^] | 1198 | if (auto M = IRGen(S, *F)) { |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 1199 | S.addPrototypeAST(llvm::make_unique<PrototypeAST>(*F->Proto)); |
David Blaikie | 1803dc2 | 2015-02-08 20:29:28 +0000 | [diff] [blame] | 1200 | J.addModule(std::move(M)); |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 1201 | } |
| 1202 | } else { |
| 1203 | // Skip token for error recovery. |
| 1204 | getNextToken(); |
| 1205 | } |
| 1206 | } |
| 1207 | |
| 1208 | static void HandleExtern(SessionContext &S) { |
| 1209 | if (auto P = ParseExtern()) |
| 1210 | S.addPrototypeAST(std::move(P)); |
| 1211 | else { |
| 1212 | // Skip token for error recovery. |
| 1213 | getNextToken(); |
| 1214 | } |
| 1215 | } |
| 1216 | |
| 1217 | static void HandleTopLevelExpression(SessionContext &S, KaleidoscopeJIT &J) { |
| 1218 | // Evaluate a top-level expression into an anonymous function. |
| 1219 | if (auto F = ParseTopLevelExpr()) { |
| 1220 | IRGenContext C(S); |
| 1221 | if (auto ExprFunc = F->IRGen(C)) { |
| 1222 | #ifndef MINIMAL_STDERR_OUTPUT |
Lang Hames | be9df34 | 2015-02-08 19:14:56 +0000 | [diff] [blame] | 1223 | std::cerr << "Expression function:\n"; |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 1224 | ExprFunc->dump(); |
| 1225 | #endif |
| 1226 | // Add the CodeGen'd module to the JIT. Keep a handle to it: We can remove |
| 1227 | // this module as soon as we've executed Function ExprFunc. |
| 1228 | auto H = J.addModule(C.takeM()); |
| 1229 | |
| 1230 | // Get the address of the JIT'd function in memory. |
| 1231 | uint64_t ExprFuncAddr = J.getSymbolAddress("__anon_expr"); |
| 1232 | |
| 1233 | // Cast it to the right type (takes no arguments, returns a double) so we |
| 1234 | // can call it as a native function. |
| 1235 | double (*FP)() = (double (*)())(intptr_t)ExprFuncAddr; |
| 1236 | #ifdef MINIMAL_STDERR_OUTPUT |
| 1237 | FP(); |
| 1238 | #else |
Lang Hames | be9df34 | 2015-02-08 19:14:56 +0000 | [diff] [blame] | 1239 | std::cerr << "Evaluated to " << FP() << "\n"; |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 1240 | #endif |
| 1241 | |
| 1242 | // Remove the function. |
| 1243 | J.removeModule(H); |
| 1244 | } |
| 1245 | } else { |
| 1246 | // Skip token for error recovery. |
| 1247 | getNextToken(); |
| 1248 | } |
| 1249 | } |
| 1250 | |
| 1251 | /// top ::= definition | external | expression | ';' |
| 1252 | static void MainLoop() { |
| 1253 | KaleidoscopeJIT J; |
| 1254 | SessionContext S(getGlobalContext()); |
| 1255 | |
| 1256 | while (1) { |
| 1257 | #ifndef MINIMAL_STDERR_OUTPUT |
Lang Hames | be9df34 | 2015-02-08 19:14:56 +0000 | [diff] [blame] | 1258 | std::cerr << "ready> "; |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 1259 | #endif |
| 1260 | switch (CurTok) { |
| 1261 | case tok_eof: return; |
| 1262 | case ';': getNextToken(); break; // ignore top-level semicolons. |
| 1263 | case tok_def: HandleDefinition(S, J); break; |
| 1264 | case tok_extern: HandleExtern(S); break; |
| 1265 | default: HandleTopLevelExpression(S, J); break; |
| 1266 | } |
| 1267 | } |
| 1268 | } |
| 1269 | |
| 1270 | //===----------------------------------------------------------------------===// |
| 1271 | // "Library" functions that can be "extern'd" from user code. |
| 1272 | //===----------------------------------------------------------------------===// |
| 1273 | |
| 1274 | /// putchard - putchar that takes a double and returns 0. |
| 1275 | extern "C" |
| 1276 | double putchard(double X) { |
| 1277 | putchar((char)X); |
| 1278 | return 0; |
| 1279 | } |
| 1280 | |
| 1281 | /// printd - printf that takes a double prints it as "%f\n", returning 0. |
| 1282 | extern "C" |
| 1283 | double printd(double X) { |
| 1284 | printf("%f", X); |
| 1285 | return 0; |
| 1286 | } |
| 1287 | |
| 1288 | extern "C" |
| 1289 | double printlf() { |
| 1290 | printf("\n"); |
| 1291 | return 0; |
| 1292 | } |
| 1293 | |
| 1294 | //===----------------------------------------------------------------------===// |
| 1295 | // Main driver code. |
| 1296 | //===----------------------------------------------------------------------===// |
| 1297 | |
| 1298 | int main() { |
| 1299 | InitializeNativeTarget(); |
| 1300 | InitializeNativeTargetAsmPrinter(); |
| 1301 | InitializeNativeTargetAsmParser(); |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 1302 | |
| 1303 | // Install standard binary operators. |
| 1304 | // 1 is lowest precedence. |
| 1305 | BinopPrecedence['='] = 2; |
| 1306 | BinopPrecedence['<'] = 10; |
| 1307 | BinopPrecedence['+'] = 20; |
| 1308 | BinopPrecedence['-'] = 20; |
| 1309 | BinopPrecedence['/'] = 40; |
| 1310 | BinopPrecedence['*'] = 40; // highest. |
| 1311 | |
| 1312 | // Prime the first token. |
| 1313 | #ifndef MINIMAL_STDERR_OUTPUT |
Lang Hames | be9df34 | 2015-02-08 19:14:56 +0000 | [diff] [blame] | 1314 | std::cerr << "ready> "; |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 1315 | #endif |
| 1316 | getNextToken(); |
| 1317 | |
Lang Hames | be9df34 | 2015-02-08 19:14:56 +0000 | [diff] [blame] | 1318 | std::cerr << std::fixed; |
| 1319 | |
Lang Hames | d855e45 | 2015-02-06 22:52:04 +0000 | [diff] [blame] | 1320 | // Run the main "interpreter loop" now. |
| 1321 | MainLoop(); |
| 1322 | |
| 1323 | return 0; |
| 1324 | } |
| 1325 | |