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_ASMPARSER_LLPARSER_H 15 #define LLVM_ASMPARSER_LLPARSER_H 16 17 #include "LLLexer.h" 18 #include "llvm/ADT/DenseMap.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 class StructType; 41 42 /// ValID - Represents a reference of a definition of some sort with no type. 43 /// There are several cases where we have to parse the value but where the 44 /// type can depend on later context. This may either be a numeric reference 45 /// or a symbolic (%var) reference. This is just a discriminated union. 46 struct ValID { 47 enum { 48 t_LocalID, t_GlobalID, // ID in UIntVal. 49 t_LocalName, t_GlobalName, // Name in StrVal. 50 t_APSInt, t_APFloat, // Value in APSIntVal/APFloatVal. 51 t_Null, t_Undef, t_Zero, // No value. 52 t_EmptyArray, // No value: [] 53 t_Constant, // Value in ConstantVal. 54 t_InlineAsm, // Value in StrVal/StrVal2/UIntVal. 55 t_MDNode, // Value in MDNodeVal. 56 t_MDString, // Value in MDStringVal. 57 t_ConstantStruct, // Value in ConstantStructElts. 58 t_PackedConstantStruct // Value in ConstantStructElts. 59 } Kind; 60 61 LLLexer::LocTy Loc; 62 unsigned UIntVal; 63 std::string StrVal, StrVal2; 64 APSInt APSIntVal; 65 APFloat APFloatVal; 66 Constant *ConstantVal; 67 MDNode *MDNodeVal; 68 MDString *MDStringVal; 69 Constant **ConstantStructElts; 70 71 ValID() : Kind(t_LocalID), APFloatVal(0.0) {} 72 ~ValID() { 73 if (Kind == t_ConstantStruct || Kind == t_PackedConstantStruct) 74 delete [] 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 95 // Instruction metadata resolution. Each instruction can have a list of 96 // MDRef info associated with them. 97 // 98 // The simpler approach of just creating temporary MDNodes and then calling 99 // RAUW on them when the definition is processed doesn't work because some 100 // instruction metadata kinds, such as dbg, get stored in the IR in an 101 // "optimized" format which doesn't participate in the normal value use 102 // lists. This means that RAUW doesn't work, even on temporary MDNodes 103 // which otherwise support RAUW. Instead, we defer resolving MDNode 104 // references until the definitions have been processed. 105 struct MDRef { 106 SMLoc Loc; 107 unsigned MDKind, MDSlot; 108 }; 109 DenseMap<Instruction*, std::vector<MDRef> > ForwardRefInstMetadata; 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::vector<std::pair<Type*, LocTy> > NumberedTypes; 117 118 std::vector<TrackingVH<MDNode> > NumberedMetadata; 119 std::map<unsigned, std::pair<TrackingVH<MDNode>, 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::vector<std::pair<ValID, GlobalValue*> > > 132 ForwardRefBlockAddresses; 133 134 // Attribute builder reference information. 135 std::map<Value*, std::vector<unsigned> > ForwardRefAttrGroups; 136 std::map<unsigned, AttrBuilder> NumberedAttrBuilders; 137 138 public: 139 LLParser(MemoryBuffer *F, SourceMgr &SM, SMDiagnostic &Err, Module *m) : 140 Context(m->getContext()), Lex(F, SM, Err, m->getContext()), 141 M(m) {} 142 bool Run(); 143 144 LLVMContext &getContext() { return Context; } 145 146 private: 147 148 bool Error(LocTy L, const Twine &Msg) const { 149 return Lex.Error(L, Msg); 150 } 151 bool TokError(const Twine &Msg) const { 152 return Error(Lex.getLoc(), Msg); 153 } 154 155 /// GetGlobalVal - Get a value with the specified name or ID, creating a 156 /// forward reference record if needed. This can return null if the value 157 /// exists but does not have the right type. 158 GlobalValue *GetGlobalVal(const std::string &N, Type *Ty, LocTy Loc); 159 GlobalValue *GetGlobalVal(unsigned ID, Type *Ty, LocTy Loc); 160 161 /// Get a Comdat with the specified name, creating a forward reference 162 /// record if needed. 163 Comdat *getComdat(const std::string &N, LocTy Loc); 164 165 // Helper Routines. 166 bool ParseToken(lltok::Kind T, const char *ErrMsg); 167 bool EatIfPresent(lltok::Kind T) { 168 if (Lex.getKind() != T) return false; 169 Lex.Lex(); 170 return true; 171 } 172 173 FastMathFlags EatFastMathFlagsIfPresent() { 174 FastMathFlags FMF; 175 while (true) 176 switch (Lex.getKind()) { 177 case lltok::kw_fast: FMF.setUnsafeAlgebra(); Lex.Lex(); continue; 178 case lltok::kw_nnan: FMF.setNoNaNs(); Lex.Lex(); continue; 179 case lltok::kw_ninf: FMF.setNoInfs(); Lex.Lex(); continue; 180 case lltok::kw_nsz: FMF.setNoSignedZeros(); Lex.Lex(); continue; 181 case lltok::kw_arcp: FMF.setAllowReciprocal(); Lex.Lex(); continue; 182 default: return FMF; 183 } 184 return FMF; 185 } 186 187 bool ParseOptionalToken(lltok::Kind T, bool &Present, 188 LocTy *Loc = nullptr) { 189 if (Lex.getKind() != T) { 190 Present = false; 191 } else { 192 if (Loc) 193 *Loc = Lex.getLoc(); 194 Lex.Lex(); 195 Present = true; 196 } 197 return false; 198 } 199 bool ParseStringConstant(std::string &Result); 200 bool ParseUInt32(unsigned &Val); 201 bool ParseUInt32(unsigned &Val, LocTy &Loc) { 202 Loc = Lex.getLoc(); 203 return ParseUInt32(Val); 204 } 205 206 bool ParseTLSModel(GlobalVariable::ThreadLocalMode &TLM); 207 bool ParseOptionalThreadLocal(GlobalVariable::ThreadLocalMode &TLM); 208 bool parseOptionalUnnamedAddr(bool &UnnamedAddr) { 209 return ParseOptionalToken(lltok::kw_unnamed_addr, UnnamedAddr); 210 } 211 bool ParseOptionalAddrSpace(unsigned &AddrSpace); 212 bool ParseOptionalParamAttrs(AttrBuilder &B); 213 bool ParseOptionalReturnAttrs(AttrBuilder &B); 214 bool ParseOptionalLinkage(unsigned &Linkage, bool &HasLinkage); 215 bool ParseOptionalLinkage(unsigned &Linkage) { 216 bool HasLinkage; return ParseOptionalLinkage(Linkage, HasLinkage); 217 } 218 bool ParseOptionalVisibility(unsigned &Visibility); 219 bool ParseOptionalDLLStorageClass(unsigned &DLLStorageClass); 220 bool ParseOptionalCallingConv(CallingConv::ID &CC); 221 bool ParseOptionalAlignment(unsigned &Alignment); 222 bool ParseScopeAndOrdering(bool isAtomic, SynchronizationScope &Scope, 223 AtomicOrdering &Ordering); 224 bool ParseOrdering(AtomicOrdering &Ordering); 225 bool ParseOptionalStackAlignment(unsigned &Alignment); 226 bool ParseOptionalCommaAlign(unsigned &Alignment, bool &AteExtraComma); 227 bool ParseOptionalCommaInAlloca(bool &IsInAlloca); 228 bool ParseIndexList(SmallVectorImpl<unsigned> &Indices,bool &AteExtraComma); 229 bool ParseIndexList(SmallVectorImpl<unsigned> &Indices) { 230 bool AteExtraComma; 231 if (ParseIndexList(Indices, AteExtraComma)) return true; 232 if (AteExtraComma) 233 return TokError("expected index"); 234 return false; 235 } 236 237 // Top-Level Entities 238 bool ParseTopLevelEntities(); 239 bool ValidateEndOfModule(); 240 bool ParseTargetDefinition(); 241 bool ParseModuleAsm(); 242 bool ParseDepLibs(); // FIXME: Remove in 4.0. 243 bool ParseUnnamedType(); 244 bool ParseNamedType(); 245 bool ParseDeclare(); 246 bool ParseDefine(); 247 248 bool ParseGlobalType(bool &IsConstant); 249 bool ParseUnnamedGlobal(); 250 bool ParseNamedGlobal(); 251 bool ParseGlobal(const std::string &Name, LocTy Loc, unsigned Linkage, 252 bool HasLinkage, unsigned Visibility, 253 unsigned DLLStorageClass, 254 GlobalVariable::ThreadLocalMode TLM, bool UnnamedAddr); 255 bool ParseAlias(const std::string &Name, LocTy Loc, unsigned Visibility, 256 unsigned DLLStorageClass, 257 GlobalVariable::ThreadLocalMode TLM, bool UnnamedAddr); 258 bool parseComdat(); 259 bool ParseStandaloneMetadata(); 260 bool ParseNamedMetadata(); 261 bool ParseMDString(MDString *&Result); 262 bool ParseMDNodeID(MDNode *&Result); 263 bool ParseMDNodeID(MDNode *&Result, unsigned &SlotNo); 264 bool ParseUnnamedAttrGrp(); 265 bool ParseFnAttributeValuePairs(AttrBuilder &B, 266 std::vector<unsigned> &FwdRefAttrGrps, 267 bool inAttrGrp, LocTy &BuiltinLoc); 268 269 // Type Parsing. 270 bool ParseType(Type *&Result, bool AllowVoid = false); 271 bool ParseType(Type *&Result, LocTy &Loc, bool AllowVoid = false) { 272 Loc = Lex.getLoc(); 273 return ParseType(Result, AllowVoid); 274 } 275 bool ParseAnonStructType(Type *&Result, bool Packed); 276 bool ParseStructBody(SmallVectorImpl<Type*> &Body); 277 bool ParseStructDefinition(SMLoc TypeLoc, StringRef Name, 278 std::pair<Type*, LocTy> &Entry, 279 Type *&ResultTy); 280 281 bool ParseArrayVectorType(Type *&Result, bool isVector); 282 bool ParseFunctionType(Type *&Result); 283 284 // Function Semantic Analysis. 285 class PerFunctionState { 286 LLParser &P; 287 Function &F; 288 std::map<std::string, std::pair<Value*, LocTy> > ForwardRefVals; 289 std::map<unsigned, std::pair<Value*, LocTy> > ForwardRefValIDs; 290 std::vector<Value*> NumberedVals; 291 292 /// FunctionNumber - If this is an unnamed function, this is the slot 293 /// number of it, otherwise it is -1. 294 int FunctionNumber; 295 public: 296 PerFunctionState(LLParser &p, Function &f, int FunctionNumber); 297 ~PerFunctionState(); 298 299 Function &getFunction() const { return F; } 300 301 bool FinishFunction(); 302 303 /// GetVal - Get a value with the specified name or ID, creating a 304 /// forward reference record if needed. This can return null if the value 305 /// exists but does not have the right type. 306 Value *GetVal(const std::string &Name, Type *Ty, LocTy Loc); 307 Value *GetVal(unsigned ID, Type *Ty, LocTy Loc); 308 309 /// SetInstName - After an instruction is parsed and inserted into its 310 /// basic block, this installs its name. 311 bool SetInstName(int NameID, const std::string &NameStr, LocTy NameLoc, 312 Instruction *Inst); 313 314 /// GetBB - Get a basic block with the specified name or ID, creating a 315 /// forward reference record if needed. This can return null if the value 316 /// is not a BasicBlock. 317 BasicBlock *GetBB(const std::string &Name, LocTy Loc); 318 BasicBlock *GetBB(unsigned ID, LocTy Loc); 319 320 /// DefineBB - Define the specified basic block, which is either named or 321 /// unnamed. If there is an error, this returns null otherwise it returns 322 /// the block being defined. 323 BasicBlock *DefineBB(const std::string &Name, LocTy Loc); 324 }; 325 326 bool ConvertValIDToValue(Type *Ty, ValID &ID, Value *&V, 327 PerFunctionState *PFS); 328 329 bool ParseValue(Type *Ty, Value *&V, PerFunctionState *PFS); 330 bool ParseValue(Type *Ty, Value *&V, PerFunctionState &PFS) { 331 return ParseValue(Ty, V, &PFS); 332 } 333 bool ParseValue(Type *Ty, Value *&V, LocTy &Loc, 334 PerFunctionState &PFS) { 335 Loc = Lex.getLoc(); 336 return ParseValue(Ty, V, &PFS); 337 } 338 339 bool ParseTypeAndValue(Value *&V, PerFunctionState *PFS); 340 bool ParseTypeAndValue(Value *&V, PerFunctionState &PFS) { 341 return ParseTypeAndValue(V, &PFS); 342 } 343 bool ParseTypeAndValue(Value *&V, LocTy &Loc, PerFunctionState &PFS) { 344 Loc = Lex.getLoc(); 345 return ParseTypeAndValue(V, PFS); 346 } 347 bool ParseTypeAndBasicBlock(BasicBlock *&BB, LocTy &Loc, 348 PerFunctionState &PFS); 349 bool ParseTypeAndBasicBlock(BasicBlock *&BB, PerFunctionState &PFS) { 350 LocTy Loc; 351 return ParseTypeAndBasicBlock(BB, Loc, PFS); 352 } 353 354 355 struct ParamInfo { 356 LocTy Loc; 357 Value *V; 358 AttributeSet Attrs; 359 ParamInfo(LocTy loc, Value *v, AttributeSet attrs) 360 : Loc(loc), V(v), Attrs(attrs) {} 361 }; 362 bool ParseParameterList(SmallVectorImpl<ParamInfo> &ArgList, 363 PerFunctionState &PFS); 364 365 // Constant Parsing. 366 bool ParseValID(ValID &ID, PerFunctionState *PFS = nullptr); 367 bool ParseGlobalValue(Type *Ty, Constant *&V); 368 bool ParseGlobalTypeAndValue(Constant *&V); 369 bool ParseGlobalValueVector(SmallVectorImpl<Constant*> &Elts); 370 bool parseOptionalComdat(Comdat *&C); 371 bool ParseMetadataListValue(ValID &ID, PerFunctionState *PFS); 372 bool ParseMetadataValue(ValID &ID, PerFunctionState *PFS); 373 bool ParseMDNodeVector(SmallVectorImpl<Value*> &, PerFunctionState *PFS); 374 bool ParseInstructionMetadata(Instruction *Inst, PerFunctionState *PFS); 375 376 // Function Parsing. 377 struct ArgInfo { 378 LocTy Loc; 379 Type *Ty; 380 AttributeSet Attrs; 381 std::string Name; 382 ArgInfo(LocTy L, Type *ty, AttributeSet Attr, const std::string &N) 383 : Loc(L), Ty(ty), Attrs(Attr), Name(N) {} 384 }; 385 bool ParseArgumentList(SmallVectorImpl<ArgInfo> &ArgList, bool &isVarArg); 386 bool ParseFunctionHeader(Function *&Fn, bool isDefine); 387 bool ParseFunctionBody(Function &Fn); 388 bool ParseBasicBlock(PerFunctionState &PFS); 389 390 enum TailCallType { TCT_None, TCT_Tail, TCT_MustTail }; 391 392 // Instruction Parsing. Each instruction parsing routine can return with a 393 // normal result, an error result, or return having eaten an extra comma. 394 enum InstResult { InstNormal = 0, InstError = 1, InstExtraComma = 2 }; 395 int ParseInstruction(Instruction *&Inst, BasicBlock *BB, 396 PerFunctionState &PFS); 397 bool ParseCmpPredicate(unsigned &Pred, unsigned Opc); 398 399 bool ParseRet(Instruction *&Inst, BasicBlock *BB, PerFunctionState &PFS); 400 bool ParseBr(Instruction *&Inst, PerFunctionState &PFS); 401 bool ParseSwitch(Instruction *&Inst, PerFunctionState &PFS); 402 bool ParseIndirectBr(Instruction *&Inst, PerFunctionState &PFS); 403 bool ParseInvoke(Instruction *&Inst, PerFunctionState &PFS); 404 bool ParseResume(Instruction *&Inst, PerFunctionState &PFS); 405 406 bool ParseArithmetic(Instruction *&I, PerFunctionState &PFS, unsigned Opc, 407 unsigned OperandType); 408 bool ParseLogical(Instruction *&I, PerFunctionState &PFS, unsigned Opc); 409 bool ParseCompare(Instruction *&I, PerFunctionState &PFS, unsigned Opc); 410 bool ParseCast(Instruction *&I, PerFunctionState &PFS, unsigned Opc); 411 bool ParseSelect(Instruction *&I, PerFunctionState &PFS); 412 bool ParseVA_Arg(Instruction *&I, PerFunctionState &PFS); 413 bool ParseExtractElement(Instruction *&I, PerFunctionState &PFS); 414 bool ParseInsertElement(Instruction *&I, PerFunctionState &PFS); 415 bool ParseShuffleVector(Instruction *&I, PerFunctionState &PFS); 416 int ParsePHI(Instruction *&I, PerFunctionState &PFS); 417 bool ParseLandingPad(Instruction *&I, PerFunctionState &PFS); 418 bool ParseCall(Instruction *&I, PerFunctionState &PFS, 419 CallInst::TailCallKind IsTail); 420 int ParseAlloc(Instruction *&I, PerFunctionState &PFS); 421 int ParseLoad(Instruction *&I, PerFunctionState &PFS); 422 int ParseStore(Instruction *&I, PerFunctionState &PFS); 423 int ParseCmpXchg(Instruction *&I, PerFunctionState &PFS); 424 int ParseAtomicRMW(Instruction *&I, PerFunctionState &PFS); 425 int ParseFence(Instruction *&I, PerFunctionState &PFS); 426 int ParseGetElementPtr(Instruction *&I, PerFunctionState &PFS); 427 int ParseExtractValue(Instruction *&I, PerFunctionState &PFS); 428 int ParseInsertValue(Instruction *&I, PerFunctionState &PFS); 429 430 bool ResolveForwardRefBlockAddresses(Function *TheFn, 431 std::vector<std::pair<ValID, GlobalValue*> > &Refs, 432 PerFunctionState *PFS); 433 }; 434 } // End llvm namespace 435 436 #endif 437
