1 //===-- LLParser.h - Parser Class -------------------------------*- C++ -*-===// 2 // 3 // The LLVM Compiler Infrastructure 4 // 5 // This file is distributed under the University of Illinois Open Source 6 // License. See LICENSE.TXT for details. 7 // 8 //===----------------------------------------------------------------------===// 9 // 10 // This file defines the parser class for .ll files. 11 // 12 //===----------------------------------------------------------------------===// 13 14 #ifndef LLVM_LIB_ASMPARSER_LLPARSER_H 15 #define LLVM_LIB_ASMPARSER_LLPARSER_H 16 17 #include "LLLexer.h" 18 #include "llvm/ADT/Optional.h" 19 #include "llvm/ADT/StringMap.h" 20 #include "llvm/IR/Attributes.h" 21 #include "llvm/IR/Instructions.h" 22 #include "llvm/IR/Module.h" 23 #include "llvm/IR/Operator.h" 24 #include "llvm/IR/Type.h" 25 #include "llvm/IR/ValueHandle.h" 26 #include <map> 27 28 namespace llvm { 29 class Module; 30 class OpaqueType; 31 class Function; 32 class Value; 33 class BasicBlock; 34 class Instruction; 35 class Constant; 36 class GlobalValue; 37 class Comdat; 38 class MDString; 39 class MDNode; 40 struct SlotMapping; 41 class StructType; 42 43 /// ValID - Represents a reference of a definition of some sort with no type. 44 /// There are several cases where we have to parse the value but where the 45 /// type can depend on later context. This may either be a numeric reference 46 /// or a symbolic (%var) reference. This is just a discriminated union. 47 struct ValID { 48 enum { 49 t_LocalID, t_GlobalID, // ID in UIntVal. 50 t_LocalName, t_GlobalName, // Name in StrVal. 51 t_APSInt, t_APFloat, // Value in APSIntVal/APFloatVal. 52 t_Null, t_Undef, t_Zero, t_None, // No value. 53 t_EmptyArray, // No value: [] 54 t_Constant, // Value in ConstantVal. 55 t_InlineAsm, // Value in FTy/StrVal/StrVal2/UIntVal. 56 t_ConstantStruct, // Value in ConstantStructElts. 57 t_PackedConstantStruct // Value in ConstantStructElts. 58 } Kind = t_LocalID; 59 60 LLLexer::LocTy Loc; 61 unsigned UIntVal; 62 FunctionType *FTy = nullptr; 63 std::string StrVal, StrVal2; 64 APSInt APSIntVal; 65 APFloat APFloatVal{0.0}; 66 Constant *ConstantVal; 67 std::unique_ptr<Constant *[]> ConstantStructElts; 68 69 ValID() = default; 70 ValID(const ValID &RHS) 71 : Kind(RHS.Kind), Loc(RHS.Loc), UIntVal(RHS.UIntVal), FTy(RHS.FTy), 72 StrVal(RHS.StrVal), StrVal2(RHS.StrVal2), APSIntVal(RHS.APSIntVal), 73 APFloatVal(RHS.APFloatVal), ConstantVal(RHS.ConstantVal) { 74 assert(!RHS.ConstantStructElts); 75 } 76 77 bool operator<(const ValID &RHS) const { 78 if (Kind == t_LocalID || Kind == t_GlobalID) 79 return UIntVal < RHS.UIntVal; 80 assert((Kind == t_LocalName || Kind == t_GlobalName || 81 Kind == t_ConstantStruct || Kind == t_PackedConstantStruct) && 82 "Ordering not defined for this ValID kind yet"); 83 return StrVal < RHS.StrVal; 84 } 85 }; 86 87 class LLParser { 88 public: 89 typedef LLLexer::LocTy LocTy; 90 private: 91 LLVMContext &Context; 92 LLLexer Lex; 93 Module *M; 94 SlotMapping *Slots; 95 96 // Instruction metadata resolution. Each instruction can have a list of 97 // MDRef info associated with them. 98 // 99 // The simpler approach of just creating temporary MDNodes and then calling 100 // RAUW on them when the definition is processed doesn't work because some 101 // instruction metadata kinds, such as dbg, get stored in the IR in an 102 // "optimized" format which doesn't participate in the normal value use 103 // lists. This means that RAUW doesn't work, even on temporary MDNodes 104 // which otherwise support RAUW. Instead, we defer resolving MDNode 105 // references until the definitions have been processed. 106 struct MDRef { 107 SMLoc Loc; 108 unsigned MDKind, MDSlot; 109 }; 110 111 SmallVector<Instruction*, 64> InstsWithTBAATag; 112 113 // Type resolution handling data structures. The location is set when we 114 // have processed a use of the type but not a definition yet. 115 StringMap<std::pair<Type*, LocTy> > NamedTypes; 116 std::map<unsigned, std::pair<Type*, LocTy> > NumberedTypes; 117 118 std::map<unsigned, TrackingMDNodeRef> NumberedMetadata; 119 std::map<unsigned, std::pair<TempMDTuple, LocTy>> ForwardRefMDNodes; 120 121 // Global Value reference information. 122 std::map<std::string, std::pair<GlobalValue*, LocTy> > ForwardRefVals; 123 std::map<unsigned, std::pair<GlobalValue*, LocTy> > ForwardRefValIDs; 124 std::vector<GlobalValue*> NumberedVals; 125 126 // Comdat forward reference information. 127 std::map<std::string, LocTy> ForwardRefComdats; 128 129 // References to blockaddress. The key is the function ValID, the value is 130 // a list of references to blocks in that function. 131 std::map<ValID, std::map<ValID, GlobalValue *>> ForwardRefBlockAddresses; 132 class PerFunctionState; 133 /// Reference to per-function state to allow basic blocks to be 134 /// forward-referenced by blockaddress instructions within the same 135 /// function. 136 PerFunctionState *BlockAddressPFS; 137 138 // Attribute builder reference information. 139 std::map<Value*, std::vector<unsigned> > ForwardRefAttrGroups; 140 std::map<unsigned, AttrBuilder> NumberedAttrBuilders; 141 142 public: 143 LLParser(StringRef F, SourceMgr &SM, SMDiagnostic &Err, Module *M, 144 SlotMapping *Slots = nullptr) 145 : Context(M->getContext()), Lex(F, SM, Err, M->getContext()), M(M), 146 Slots(Slots), BlockAddressPFS(nullptr) {} 147 bool Run(); 148 149 bool parseStandaloneConstantValue(Constant *&C, const SlotMapping *Slots); 150 151 bool parseTypeAtBeginning(Type *&Ty, unsigned &Read, 152 const SlotMapping *Slots); 153 154 LLVMContext &getContext() { return Context; } 155 156 private: 157 158 bool Error(LocTy L, const Twine &Msg) const { 159 return Lex.Error(L, Msg); 160 } 161 bool TokError(const Twine &Msg) const { 162 return Error(Lex.getLoc(), Msg); 163 } 164 165 /// Restore the internal name and slot mappings using the mappings that 166 /// were created at an earlier parsing stage. 167 void restoreParsingState(const SlotMapping *Slots); 168 169 /// GetGlobalVal - Get a value with the specified name or ID, creating a 170 /// forward reference record if needed. This can return null if the value 171 /// exists but does not have the right type. 172 GlobalValue *GetGlobalVal(const std::string &N, Type *Ty, LocTy Loc); 173 GlobalValue *GetGlobalVal(unsigned ID, Type *Ty, LocTy Loc); 174 175 /// Get a Comdat with the specified name, creating a forward reference 176 /// record if needed. 177 Comdat *getComdat(const std::string &N, LocTy Loc); 178 179 // Helper Routines. 180 bool ParseToken(lltok::Kind T, const char *ErrMsg); 181 bool EatIfPresent(lltok::Kind T) { 182 if (Lex.getKind() != T) return false; 183 Lex.Lex(); 184 return true; 185 } 186 187 FastMathFlags EatFastMathFlagsIfPresent() { 188 FastMathFlags FMF; 189 while (true) 190 switch (Lex.getKind()) { 191 case lltok::kw_fast: FMF.setUnsafeAlgebra(); Lex.Lex(); continue; 192 case lltok::kw_nnan: FMF.setNoNaNs(); Lex.Lex(); continue; 193 case lltok::kw_ninf: FMF.setNoInfs(); Lex.Lex(); continue; 194 case lltok::kw_nsz: FMF.setNoSignedZeros(); Lex.Lex(); continue; 195 case lltok::kw_arcp: FMF.setAllowReciprocal(); Lex.Lex(); continue; 196 default: return FMF; 197 } 198 return FMF; 199 } 200 201 bool ParseOptionalToken(lltok::Kind T, bool &Present, 202 LocTy *Loc = nullptr) { 203 if (Lex.getKind() != T) { 204 Present = false; 205 } else { 206 if (Loc) 207 *Loc = Lex.getLoc(); 208 Lex.Lex(); 209 Present = true; 210 } 211 return false; 212 } 213 bool ParseStringConstant(std::string &Result); 214 bool ParseUInt32(unsigned &Val); 215 bool ParseUInt32(unsigned &Val, LocTy &Loc) { 216 Loc = Lex.getLoc(); 217 return ParseUInt32(Val); 218 } 219 bool ParseUInt64(uint64_t &Val); 220 bool ParseUInt64(uint64_t &Val, LocTy &Loc) { 221 Loc = Lex.getLoc(); 222 return ParseUInt64(Val); 223 } 224 225 bool ParseStringAttribute(AttrBuilder &B); 226 227 bool ParseTLSModel(GlobalVariable::ThreadLocalMode &TLM); 228 bool ParseOptionalThreadLocal(GlobalVariable::ThreadLocalMode &TLM); 229 bool ParseOptionalUnnamedAddr(GlobalVariable::UnnamedAddr &UnnamedAddr); 230 bool ParseOptionalAddrSpace(unsigned &AddrSpace); 231 bool ParseOptionalParamAttrs(AttrBuilder &B); 232 bool ParseOptionalReturnAttrs(AttrBuilder &B); 233 bool ParseOptionalLinkage(unsigned &Linkage, bool &HasLinkage, 234 unsigned &Visibility, unsigned &DLLStorageClass); 235 void ParseOptionalVisibility(unsigned &Visibility); 236 void ParseOptionalDLLStorageClass(unsigned &DLLStorageClass); 237 bool ParseOptionalCallingConv(unsigned &CC); 238 bool ParseOptionalAlignment(unsigned &Alignment); 239 bool ParseOptionalDerefAttrBytes(lltok::Kind AttrKind, uint64_t &Bytes); 240 bool ParseScopeAndOrdering(bool isAtomic, SynchronizationScope &Scope, 241 AtomicOrdering &Ordering); 242 bool ParseOrdering(AtomicOrdering &Ordering); 243 bool ParseOptionalStackAlignment(unsigned &Alignment); 244 bool ParseOptionalCommaAlign(unsigned &Alignment, bool &AteExtraComma); 245 bool ParseOptionalCommaInAlloca(bool &IsInAlloca); 246 bool parseAllocSizeArguments(unsigned &ElemSizeArg, 247 Optional<unsigned> &HowManyArg); 248 bool ParseIndexList(SmallVectorImpl<unsigned> &Indices, 249 bool &AteExtraComma); 250 bool ParseIndexList(SmallVectorImpl<unsigned> &Indices) { 251 bool AteExtraComma; 252 if (ParseIndexList(Indices, AteExtraComma)) return true; 253 if (AteExtraComma) 254 return TokError("expected index"); 255 return false; 256 } 257 258 // Top-Level Entities 259 bool ParseTopLevelEntities(); 260 bool ValidateEndOfModule(); 261 bool ParseTargetDefinition(); 262 bool ParseModuleAsm(); 263 bool ParseSourceFileName(); 264 bool ParseDepLibs(); // FIXME: Remove in 4.0. 265 bool ParseUnnamedType(); 266 bool ParseNamedType(); 267 bool ParseDeclare(); 268 bool ParseDefine(); 269 270 bool ParseGlobalType(bool &IsConstant); 271 bool ParseUnnamedGlobal(); 272 bool ParseNamedGlobal(); 273 bool ParseGlobal(const std::string &Name, LocTy Loc, unsigned Linkage, 274 bool HasLinkage, unsigned Visibility, 275 unsigned DLLStorageClass, 276 GlobalVariable::ThreadLocalMode TLM, 277 GlobalVariable::UnnamedAddr UnnamedAddr); 278 bool parseIndirectSymbol(const std::string &Name, LocTy Loc, 279 unsigned Linkage, unsigned Visibility, 280 unsigned DLLStorageClass, 281 GlobalVariable::ThreadLocalMode TLM, 282 GlobalVariable::UnnamedAddr UnnamedAddr); 283 bool parseComdat(); 284 bool ParseStandaloneMetadata(); 285 bool ParseNamedMetadata(); 286 bool ParseMDString(MDString *&Result); 287 bool ParseMDNodeID(MDNode *&Result); 288 bool ParseUnnamedAttrGrp(); 289 bool ParseFnAttributeValuePairs(AttrBuilder &B, 290 std::vector<unsigned> &FwdRefAttrGrps, 291 bool inAttrGrp, LocTy &BuiltinLoc); 292 293 // Type Parsing. 294 bool ParseType(Type *&Result, const Twine &Msg, bool AllowVoid = false); 295 bool ParseType(Type *&Result, bool AllowVoid = false) { 296 return ParseType(Result, "expected type", AllowVoid); 297 } 298 bool ParseType(Type *&Result, const Twine &Msg, LocTy &Loc, 299 bool AllowVoid = false) { 300 Loc = Lex.getLoc(); 301 return ParseType(Result, Msg, AllowVoid); 302 } 303 bool ParseType(Type *&Result, LocTy &Loc, bool AllowVoid = false) { 304 Loc = Lex.getLoc(); 305 return ParseType(Result, AllowVoid); 306 } 307 bool ParseAnonStructType(Type *&Result, bool Packed); 308 bool ParseStructBody(SmallVectorImpl<Type*> &Body); 309 bool ParseStructDefinition(SMLoc TypeLoc, StringRef Name, 310 std::pair<Type*, LocTy> &Entry, 311 Type *&ResultTy); 312 313 bool ParseArrayVectorType(Type *&Result, bool isVector); 314 bool ParseFunctionType(Type *&Result); 315 316 // Function Semantic Analysis. 317 class PerFunctionState { 318 LLParser &P; 319 Function &F; 320 std::map<std::string, std::pair<Value*, LocTy> > ForwardRefVals; 321 std::map<unsigned, std::pair<Value*, LocTy> > ForwardRefValIDs; 322 std::vector<Value*> NumberedVals; 323 324 /// FunctionNumber - If this is an unnamed function, this is the slot 325 /// number of it, otherwise it is -1. 326 int FunctionNumber; 327 public: 328 PerFunctionState(LLParser &p, Function &f, int FunctionNumber); 329 ~PerFunctionState(); 330 331 Function &getFunction() const { return F; } 332 333 bool FinishFunction(); 334 335 /// GetVal - Get a value with the specified name or ID, creating a 336 /// forward reference record if needed. This can return null if the value 337 /// exists but does not have the right type. 338 Value *GetVal(const std::string &Name, Type *Ty, LocTy Loc); 339 Value *GetVal(unsigned ID, Type *Ty, LocTy Loc); 340 341 /// SetInstName - After an instruction is parsed and inserted into its 342 /// basic block, this installs its name. 343 bool SetInstName(int NameID, const std::string &NameStr, LocTy NameLoc, 344 Instruction *Inst); 345 346 /// GetBB - Get a basic block with the specified name or ID, creating a 347 /// forward reference record if needed. This can return null if the value 348 /// is not a BasicBlock. 349 BasicBlock *GetBB(const std::string &Name, LocTy Loc); 350 BasicBlock *GetBB(unsigned ID, LocTy Loc); 351 352 /// DefineBB - Define the specified basic block, which is either named or 353 /// unnamed. If there is an error, this returns null otherwise it returns 354 /// the block being defined. 355 BasicBlock *DefineBB(const std::string &Name, LocTy Loc); 356 357 bool resolveForwardRefBlockAddresses(); 358 }; 359 360 bool ConvertValIDToValue(Type *Ty, ValID &ID, Value *&V, 361 PerFunctionState *PFS); 362 363 bool parseConstantValue(Type *Ty, Constant *&C); 364 bool ParseValue(Type *Ty, Value *&V, PerFunctionState *PFS); 365 bool ParseValue(Type *Ty, Value *&V, PerFunctionState &PFS) { 366 return ParseValue(Ty, V, &PFS); 367 } 368 369 bool ParseValue(Type *Ty, Value *&V, LocTy &Loc, 370 PerFunctionState &PFS) { 371 Loc = Lex.getLoc(); 372 return ParseValue(Ty, V, &PFS); 373 } 374 375 bool ParseTypeAndValue(Value *&V, PerFunctionState *PFS); 376 bool ParseTypeAndValue(Value *&V, PerFunctionState &PFS) { 377 return ParseTypeAndValue(V, &PFS); 378 } 379 bool ParseTypeAndValue(Value *&V, LocTy &Loc, PerFunctionState &PFS) { 380 Loc = Lex.getLoc(); 381 return ParseTypeAndValue(V, PFS); 382 } 383 bool ParseTypeAndBasicBlock(BasicBlock *&BB, LocTy &Loc, 384 PerFunctionState &PFS); 385 bool ParseTypeAndBasicBlock(BasicBlock *&BB, PerFunctionState &PFS) { 386 LocTy Loc; 387 return ParseTypeAndBasicBlock(BB, Loc, PFS); 388 } 389 390 391 struct ParamInfo { 392 LocTy Loc; 393 Value *V; 394 AttributeSet Attrs; 395 ParamInfo(LocTy loc, Value *v, AttributeSet attrs) 396 : Loc(loc), V(v), Attrs(attrs) {} 397 }; 398 bool ParseParameterList(SmallVectorImpl<ParamInfo> &ArgList, 399 PerFunctionState &PFS, 400 bool IsMustTailCall = false, 401 bool InVarArgsFunc = false); 402 403 bool 404 ParseOptionalOperandBundles(SmallVectorImpl<OperandBundleDef> &BundleList, 405 PerFunctionState &PFS); 406 407 bool ParseExceptionArgs(SmallVectorImpl<Value *> &Args, 408 PerFunctionState &PFS); 409 410 // Constant Parsing. 411 bool ParseValID(ValID &ID, PerFunctionState *PFS = nullptr); 412 bool ParseGlobalValue(Type *Ty, Constant *&V); 413 bool ParseGlobalTypeAndValue(Constant *&V); 414 bool ParseGlobalValueVector(SmallVectorImpl<Constant *> &Elts); 415 bool parseOptionalComdat(StringRef GlobalName, Comdat *&C); 416 bool ParseMetadataAsValue(Value *&V, PerFunctionState &PFS); 417 bool ParseValueAsMetadata(Metadata *&MD, const Twine &TypeMsg, 418 PerFunctionState *PFS); 419 bool ParseMetadata(Metadata *&MD, PerFunctionState *PFS); 420 bool ParseMDTuple(MDNode *&MD, bool IsDistinct = false); 421 bool ParseMDNode(MDNode *&MD); 422 bool ParseMDNodeTail(MDNode *&MD); 423 bool ParseMDNodeVector(SmallVectorImpl<Metadata *> &MDs); 424 bool ParseMetadataAttachment(unsigned &Kind, MDNode *&MD); 425 bool ParseInstructionMetadata(Instruction &Inst); 426 bool ParseGlobalObjectMetadataAttachment(GlobalObject &GO); 427 bool ParseOptionalFunctionMetadata(Function &F); 428 429 template <class FieldTy> 430 bool ParseMDField(LocTy Loc, StringRef Name, FieldTy &Result); 431 template <class FieldTy> bool ParseMDField(StringRef Name, FieldTy &Result); 432 template <class ParserTy> 433 bool ParseMDFieldsImplBody(ParserTy parseField); 434 template <class ParserTy> 435 bool ParseMDFieldsImpl(ParserTy parseField, LocTy &ClosingLoc); 436 bool ParseSpecializedMDNode(MDNode *&N, bool IsDistinct = false); 437 438 #define HANDLE_SPECIALIZED_MDNODE_LEAF(CLASS) \ 439 bool Parse##CLASS(MDNode *&Result, bool IsDistinct); 440 #include "llvm/IR/Metadata.def" 441 442 // Function Parsing. 443 struct ArgInfo { 444 LocTy Loc; 445 Type *Ty; 446 AttributeSet Attrs; 447 std::string Name; 448 ArgInfo(LocTy L, Type *ty, AttributeSet Attr, const std::string &N) 449 : Loc(L), Ty(ty), Attrs(Attr), Name(N) {} 450 }; 451 bool ParseArgumentList(SmallVectorImpl<ArgInfo> &ArgList, bool &isVarArg); 452 bool ParseFunctionHeader(Function *&Fn, bool isDefine); 453 bool ParseFunctionBody(Function &Fn); 454 bool ParseBasicBlock(PerFunctionState &PFS); 455 456 enum TailCallType { TCT_None, TCT_Tail, TCT_MustTail }; 457 458 // Instruction Parsing. Each instruction parsing routine can return with a 459 // normal result, an error result, or return having eaten an extra comma. 460 enum InstResult { InstNormal = 0, InstError = 1, InstExtraComma = 2 }; 461 int ParseInstruction(Instruction *&Inst, BasicBlock *BB, 462 PerFunctionState &PFS); 463 bool ParseCmpPredicate(unsigned &Pred, unsigned Opc); 464 465 bool ParseRet(Instruction *&Inst, BasicBlock *BB, PerFunctionState &PFS); 466 bool ParseBr(Instruction *&Inst, PerFunctionState &PFS); 467 bool ParseSwitch(Instruction *&Inst, PerFunctionState &PFS); 468 bool ParseIndirectBr(Instruction *&Inst, PerFunctionState &PFS); 469 bool ParseInvoke(Instruction *&Inst, PerFunctionState &PFS); 470 bool ParseResume(Instruction *&Inst, PerFunctionState &PFS); 471 bool ParseCleanupRet(Instruction *&Inst, PerFunctionState &PFS); 472 bool ParseCatchRet(Instruction *&Inst, PerFunctionState &PFS); 473 bool ParseCatchSwitch(Instruction *&Inst, PerFunctionState &PFS); 474 bool ParseCatchPad(Instruction *&Inst, PerFunctionState &PFS); 475 bool ParseCleanupPad(Instruction *&Inst, PerFunctionState &PFS); 476 477 bool ParseArithmetic(Instruction *&I, PerFunctionState &PFS, unsigned Opc, 478 unsigned OperandType); 479 bool ParseLogical(Instruction *&I, PerFunctionState &PFS, unsigned Opc); 480 bool ParseCompare(Instruction *&I, PerFunctionState &PFS, unsigned Opc); 481 bool ParseCast(Instruction *&I, PerFunctionState &PFS, unsigned Opc); 482 bool ParseSelect(Instruction *&I, PerFunctionState &PFS); 483 bool ParseVA_Arg(Instruction *&I, PerFunctionState &PFS); 484 bool ParseExtractElement(Instruction *&I, PerFunctionState &PFS); 485 bool ParseInsertElement(Instruction *&I, PerFunctionState &PFS); 486 bool ParseShuffleVector(Instruction *&I, PerFunctionState &PFS); 487 int ParsePHI(Instruction *&I, PerFunctionState &PFS); 488 bool ParseLandingPad(Instruction *&I, PerFunctionState &PFS); 489 bool ParseCall(Instruction *&I, PerFunctionState &PFS, 490 CallInst::TailCallKind IsTail); 491 int ParseAlloc(Instruction *&I, PerFunctionState &PFS); 492 int ParseLoad(Instruction *&I, PerFunctionState &PFS); 493 int ParseStore(Instruction *&I, PerFunctionState &PFS); 494 int ParseCmpXchg(Instruction *&I, PerFunctionState &PFS); 495 int ParseAtomicRMW(Instruction *&I, PerFunctionState &PFS); 496 int ParseFence(Instruction *&I, PerFunctionState &PFS); 497 int ParseGetElementPtr(Instruction *&I, PerFunctionState &PFS); 498 int ParseExtractValue(Instruction *&I, PerFunctionState &PFS); 499 int ParseInsertValue(Instruction *&I, PerFunctionState &PFS); 500 501 // Use-list order directives. 502 bool ParseUseListOrder(PerFunctionState *PFS = nullptr); 503 bool ParseUseListOrderBB(); 504 bool ParseUseListOrderIndexes(SmallVectorImpl<unsigned> &Indexes); 505 bool sortUseListOrder(Value *V, ArrayRef<unsigned> Indexes, SMLoc Loc); 506 }; 507 } // End llvm namespace 508 509 #endif 510
