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/Support/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 MDString; 38 class MDNode; 39 class StructType; 40 41 /// ValID - Represents a reference of a definition of some sort with no type. 42 /// There are several cases where we have to parse the value but where the 43 /// type can depend on later context. This may either be a numeric reference 44 /// or a symbolic (%var) reference. This is just a discriminated union. 45 struct ValID { 46 enum { 47 t_LocalID, t_GlobalID, // ID in UIntVal. 48 t_LocalName, t_GlobalName, // Name in StrVal. 49 t_APSInt, t_APFloat, // Value in APSIntVal/APFloatVal. 50 t_Null, t_Undef, t_Zero, // No value. 51 t_EmptyArray, // No value: [] 52 t_Constant, // Value in ConstantVal. 53 t_InlineAsm, // Value in StrVal/StrVal2/UIntVal. 54 t_MDNode, // Value in MDNodeVal. 55 t_MDString, // Value in MDStringVal. 56 t_ConstantStruct, // Value in ConstantStructElts. 57 t_PackedConstantStruct // Value in ConstantStructElts. 58 } Kind; 59 60 LLLexer::LocTy Loc; 61 unsigned UIntVal; 62 std::string StrVal, StrVal2; 63 APSInt APSIntVal; 64 APFloat APFloatVal; 65 Constant *ConstantVal; 66 MDNode *MDNodeVal; 67 MDString *MDStringVal; 68 Constant **ConstantStructElts; 69 70 ValID() : Kind(t_LocalID), APFloatVal(0.0) {} 71 ~ValID() { 72 if (Kind == t_ConstantStruct || Kind == t_PackedConstantStruct) 73 delete [] ConstantStructElts; 74 } 75 76 bool operator<(const ValID &RHS) const { 77 if (Kind == t_LocalID || Kind == t_GlobalID) 78 return UIntVal < RHS.UIntVal; 79 assert((Kind == t_LocalName || Kind == t_GlobalName || 80 Kind == t_ConstantStruct || Kind == t_PackedConstantStruct) && 81 "Ordering not defined for this ValID kind yet"); 82 return StrVal < RHS.StrVal; 83 } 84 }; 85 86 class LLParser { 87 public: 88 typedef LLLexer::LocTy LocTy; 89 private: 90 LLVMContext &Context; 91 LLLexer Lex; 92 Module *M; 93 94 // Instruction metadata resolution. Each instruction can have a list of 95 // MDRef info associated with them. 96 // 97 // The simpler approach of just creating temporary MDNodes and then calling 98 // RAUW on them when the definition is processed doesn't work because some 99 // instruction metadata kinds, such as dbg, get stored in the IR in an 100 // "optimized" format which doesn't participate in the normal value use 101 // lists. This means that RAUW doesn't work, even on temporary MDNodes 102 // which otherwise support RAUW. Instead, we defer resolving MDNode 103 // references until the definitions have been processed. 104 struct MDRef { 105 SMLoc Loc; 106 unsigned MDKind, MDSlot; 107 }; 108 DenseMap<Instruction*, std::vector<MDRef> > ForwardRefInstMetadata; 109 110 // Type resolution handling data structures. The location is set when we 111 // have processed a use of the type but not a definition yet. 112 StringMap<std::pair<Type*, LocTy> > NamedTypes; 113 std::vector<std::pair<Type*, LocTy> > NumberedTypes; 114 115 std::vector<TrackingVH<MDNode> > NumberedMetadata; 116 std::map<unsigned, std::pair<TrackingVH<MDNode>, LocTy> > ForwardRefMDNodes; 117 118 // Global Value reference information. 119 std::map<std::string, std::pair<GlobalValue*, LocTy> > ForwardRefVals; 120 std::map<unsigned, std::pair<GlobalValue*, LocTy> > ForwardRefValIDs; 121 std::vector<GlobalValue*> NumberedVals; 122 123 // References to blockaddress. The key is the function ValID, the value is 124 // a list of references to blocks in that function. 125 std::map<ValID, std::vector<std::pair<ValID, GlobalValue*> > > 126 ForwardRefBlockAddresses; 127 128 // Attribute builder reference information. 129 std::map<Value*, std::vector<unsigned> > ForwardRefAttrGroups; 130 std::map<unsigned, AttrBuilder> NumberedAttrBuilders; 131 132 public: 133 LLParser(MemoryBuffer *F, SourceMgr &SM, SMDiagnostic &Err, Module *m) : 134 Context(m->getContext()), Lex(F, SM, Err, m->getContext()), 135 M(m) {} 136 bool Run(); 137 138 LLVMContext &getContext() { return Context; } 139 140 private: 141 142 bool Error(LocTy L, const Twine &Msg) const { 143 return Lex.Error(L, Msg); 144 } 145 bool TokError(const Twine &Msg) const { 146 return Error(Lex.getLoc(), Msg); 147 } 148 149 /// GetGlobalVal - Get a value with the specified name or ID, creating a 150 /// forward reference record if needed. This can return null if the value 151 /// exists but does not have the right type. 152 GlobalValue *GetGlobalVal(const std::string &N, Type *Ty, LocTy Loc); 153 GlobalValue *GetGlobalVal(unsigned ID, Type *Ty, LocTy Loc); 154 155 // Helper Routines. 156 bool ParseToken(lltok::Kind T, const char *ErrMsg); 157 bool EatIfPresent(lltok::Kind T) { 158 if (Lex.getKind() != T) return false; 159 Lex.Lex(); 160 return true; 161 } 162 163 FastMathFlags EatFastMathFlagsIfPresent() { 164 FastMathFlags FMF; 165 while (true) 166 switch (Lex.getKind()) { 167 case lltok::kw_fast: FMF.setUnsafeAlgebra(); Lex.Lex(); continue; 168 case lltok::kw_nnan: FMF.setNoNaNs(); Lex.Lex(); continue; 169 case lltok::kw_ninf: FMF.setNoInfs(); Lex.Lex(); continue; 170 case lltok::kw_nsz: FMF.setNoSignedZeros(); Lex.Lex(); continue; 171 case lltok::kw_arcp: FMF.setAllowReciprocal(); Lex.Lex(); continue; 172 default: return FMF; 173 } 174 return FMF; 175 } 176 177 bool ParseOptionalToken(lltok::Kind T, bool &Present, LocTy *Loc = 0) { 178 if (Lex.getKind() != T) { 179 Present = false; 180 } else { 181 if (Loc) 182 *Loc = Lex.getLoc(); 183 Lex.Lex(); 184 Present = true; 185 } 186 return false; 187 } 188 bool ParseStringConstant(std::string &Result); 189 bool ParseUInt32(unsigned &Val); 190 bool ParseUInt32(unsigned &Val, LocTy &Loc) { 191 Loc = Lex.getLoc(); 192 return ParseUInt32(Val); 193 } 194 195 bool ParseTLSModel(GlobalVariable::ThreadLocalMode &TLM); 196 bool ParseOptionalThreadLocal(GlobalVariable::ThreadLocalMode &TLM); 197 bool ParseOptionalAddrSpace(unsigned &AddrSpace); 198 bool ParseOptionalParamAttrs(AttrBuilder &B); 199 bool ParseOptionalReturnAttrs(AttrBuilder &B); 200 bool ParseOptionalLinkage(unsigned &Linkage, bool &HasLinkage); 201 bool ParseOptionalLinkage(unsigned &Linkage) { 202 bool HasLinkage; return ParseOptionalLinkage(Linkage, HasLinkage); 203 } 204 bool ParseOptionalVisibility(unsigned &Visibility); 205 bool ParseOptionalCallingConv(CallingConv::ID &CC); 206 bool ParseOptionalAlignment(unsigned &Alignment); 207 bool ParseScopeAndOrdering(bool isAtomic, SynchronizationScope &Scope, 208 AtomicOrdering &Ordering); 209 bool ParseOptionalStackAlignment(unsigned &Alignment); 210 bool ParseOptionalCommaAlign(unsigned &Alignment, bool &AteExtraComma); 211 bool ParseIndexList(SmallVectorImpl<unsigned> &Indices,bool &AteExtraComma); 212 bool ParseIndexList(SmallVectorImpl<unsigned> &Indices) { 213 bool AteExtraComma; 214 if (ParseIndexList(Indices, AteExtraComma)) return true; 215 if (AteExtraComma) 216 return TokError("expected index"); 217 return false; 218 } 219 220 // Top-Level Entities 221 bool ParseTopLevelEntities(); 222 bool ValidateEndOfModule(); 223 bool ParseTargetDefinition(); 224 bool ParseModuleAsm(); 225 bool ParseDepLibs(); // FIXME: Remove in 4.0. 226 bool ParseUnnamedType(); 227 bool ParseNamedType(); 228 bool ParseDeclare(); 229 bool ParseDefine(); 230 231 bool ParseGlobalType(bool &IsConstant); 232 bool ParseUnnamedGlobal(); 233 bool ParseNamedGlobal(); 234 bool ParseGlobal(const std::string &Name, LocTy Loc, unsigned Linkage, 235 bool HasLinkage, unsigned Visibility); 236 bool ParseAlias(const std::string &Name, LocTy Loc, unsigned Visibility); 237 bool ParseStandaloneMetadata(); 238 bool ParseNamedMetadata(); 239 bool ParseMDString(MDString *&Result); 240 bool ParseMDNodeID(MDNode *&Result); 241 bool ParseMDNodeID(MDNode *&Result, unsigned &SlotNo); 242 bool ParseUnnamedAttrGrp(); 243 bool ParseFnAttributeValuePairs(AttrBuilder &B, 244 std::vector<unsigned> &FwdRefAttrGrps, 245 bool inAttrGrp, LocTy &BuiltinLoc); 246 247 // Type Parsing. 248 bool ParseType(Type *&Result, bool AllowVoid = false); 249 bool ParseType(Type *&Result, LocTy &Loc, bool AllowVoid = false) { 250 Loc = Lex.getLoc(); 251 return ParseType(Result, AllowVoid); 252 } 253 bool ParseAnonStructType(Type *&Result, bool Packed); 254 bool ParseStructBody(SmallVectorImpl<Type*> &Body); 255 bool ParseStructDefinition(SMLoc TypeLoc, StringRef Name, 256 std::pair<Type*, LocTy> &Entry, 257 Type *&ResultTy); 258 259 bool ParseArrayVectorType(Type *&Result, bool isVector); 260 bool ParseFunctionType(Type *&Result); 261 262 // Function Semantic Analysis. 263 class PerFunctionState { 264 LLParser &P; 265 Function &F; 266 std::map<std::string, std::pair<Value*, LocTy> > ForwardRefVals; 267 std::map<unsigned, std::pair<Value*, LocTy> > ForwardRefValIDs; 268 std::vector<Value*> NumberedVals; 269 270 /// FunctionNumber - If this is an unnamed function, this is the slot 271 /// number of it, otherwise it is -1. 272 int FunctionNumber; 273 public: 274 PerFunctionState(LLParser &p, Function &f, int FunctionNumber); 275 ~PerFunctionState(); 276 277 Function &getFunction() const { return F; } 278 279 bool FinishFunction(); 280 281 /// GetVal - Get a value with the specified name or ID, creating a 282 /// forward reference record if needed. This can return null if the value 283 /// exists but does not have the right type. 284 Value *GetVal(const std::string &Name, Type *Ty, LocTy Loc); 285 Value *GetVal(unsigned ID, Type *Ty, LocTy Loc); 286 287 /// SetInstName - After an instruction is parsed and inserted into its 288 /// basic block, this installs its name. 289 bool SetInstName(int NameID, const std::string &NameStr, LocTy NameLoc, 290 Instruction *Inst); 291 292 /// GetBB - Get a basic block with the specified name or ID, creating a 293 /// forward reference record if needed. This can return null if the value 294 /// is not a BasicBlock. 295 BasicBlock *GetBB(const std::string &Name, LocTy Loc); 296 BasicBlock *GetBB(unsigned ID, LocTy Loc); 297 298 /// DefineBB - Define the specified basic block, which is either named or 299 /// unnamed. If there is an error, this returns null otherwise it returns 300 /// the block being defined. 301 BasicBlock *DefineBB(const std::string &Name, LocTy Loc); 302 }; 303 304 bool ConvertValIDToValue(Type *Ty, ValID &ID, Value *&V, 305 PerFunctionState *PFS); 306 307 bool ParseValue(Type *Ty, Value *&V, PerFunctionState *PFS); 308 bool ParseValue(Type *Ty, Value *&V, PerFunctionState &PFS) { 309 return ParseValue(Ty, V, &PFS); 310 } 311 bool ParseValue(Type *Ty, Value *&V, LocTy &Loc, 312 PerFunctionState &PFS) { 313 Loc = Lex.getLoc(); 314 return ParseValue(Ty, V, &PFS); 315 } 316 317 bool ParseTypeAndValue(Value *&V, PerFunctionState *PFS); 318 bool ParseTypeAndValue(Value *&V, PerFunctionState &PFS) { 319 return ParseTypeAndValue(V, &PFS); 320 } 321 bool ParseTypeAndValue(Value *&V, LocTy &Loc, PerFunctionState &PFS) { 322 Loc = Lex.getLoc(); 323 return ParseTypeAndValue(V, PFS); 324 } 325 bool ParseTypeAndBasicBlock(BasicBlock *&BB, LocTy &Loc, 326 PerFunctionState &PFS); 327 bool ParseTypeAndBasicBlock(BasicBlock *&BB, PerFunctionState &PFS) { 328 LocTy Loc; 329 return ParseTypeAndBasicBlock(BB, Loc, PFS); 330 } 331 332 333 struct ParamInfo { 334 LocTy Loc; 335 Value *V; 336 AttributeSet Attrs; 337 ParamInfo(LocTy loc, Value *v, AttributeSet attrs) 338 : Loc(loc), V(v), Attrs(attrs) {} 339 }; 340 bool ParseParameterList(SmallVectorImpl<ParamInfo> &ArgList, 341 PerFunctionState &PFS); 342 343 // Constant Parsing. 344 bool ParseValID(ValID &ID, PerFunctionState *PFS = NULL); 345 bool ParseGlobalValue(Type *Ty, Constant *&V); 346 bool ParseGlobalTypeAndValue(Constant *&V); 347 bool ParseGlobalValueVector(SmallVectorImpl<Constant*> &Elts); 348 bool ParseMetadataListValue(ValID &ID, PerFunctionState *PFS); 349 bool ParseMetadataValue(ValID &ID, PerFunctionState *PFS); 350 bool ParseMDNodeVector(SmallVectorImpl<Value*> &, PerFunctionState *PFS); 351 bool ParseInstructionMetadata(Instruction *Inst, PerFunctionState *PFS); 352 353 // Function Parsing. 354 struct ArgInfo { 355 LocTy Loc; 356 Type *Ty; 357 AttributeSet Attrs; 358 std::string Name; 359 ArgInfo(LocTy L, Type *ty, AttributeSet Attr, const std::string &N) 360 : Loc(L), Ty(ty), Attrs(Attr), Name(N) {} 361 }; 362 bool ParseArgumentList(SmallVectorImpl<ArgInfo> &ArgList, bool &isVarArg); 363 bool ParseFunctionHeader(Function *&Fn, bool isDefine); 364 bool ParseFunctionBody(Function &Fn); 365 bool ParseBasicBlock(PerFunctionState &PFS); 366 367 // Instruction Parsing. Each instruction parsing routine can return with a 368 // normal result, an error result, or return having eaten an extra comma. 369 enum InstResult { InstNormal = 0, InstError = 1, InstExtraComma = 2 }; 370 int ParseInstruction(Instruction *&Inst, BasicBlock *BB, 371 PerFunctionState &PFS); 372 bool ParseCmpPredicate(unsigned &Pred, unsigned Opc); 373 374 bool ParseRet(Instruction *&Inst, BasicBlock *BB, PerFunctionState &PFS); 375 bool ParseBr(Instruction *&Inst, PerFunctionState &PFS); 376 bool ParseSwitch(Instruction *&Inst, PerFunctionState &PFS); 377 bool ParseIndirectBr(Instruction *&Inst, PerFunctionState &PFS); 378 bool ParseInvoke(Instruction *&Inst, PerFunctionState &PFS); 379 bool ParseResume(Instruction *&Inst, PerFunctionState &PFS); 380 381 bool ParseArithmetic(Instruction *&I, PerFunctionState &PFS, unsigned Opc, 382 unsigned OperandType); 383 bool ParseLogical(Instruction *&I, PerFunctionState &PFS, unsigned Opc); 384 bool ParseCompare(Instruction *&I, PerFunctionState &PFS, unsigned Opc); 385 bool ParseCast(Instruction *&I, PerFunctionState &PFS, unsigned Opc); 386 bool ParseSelect(Instruction *&I, PerFunctionState &PFS); 387 bool ParseVA_Arg(Instruction *&I, PerFunctionState &PFS); 388 bool ParseExtractElement(Instruction *&I, PerFunctionState &PFS); 389 bool ParseInsertElement(Instruction *&I, PerFunctionState &PFS); 390 bool ParseShuffleVector(Instruction *&I, PerFunctionState &PFS); 391 int ParsePHI(Instruction *&I, PerFunctionState &PFS); 392 bool ParseLandingPad(Instruction *&I, PerFunctionState &PFS); 393 bool ParseCall(Instruction *&I, PerFunctionState &PFS, bool isTail); 394 int ParseAlloc(Instruction *&I, PerFunctionState &PFS); 395 int ParseLoad(Instruction *&I, PerFunctionState &PFS); 396 int ParseStore(Instruction *&I, PerFunctionState &PFS); 397 int ParseCmpXchg(Instruction *&I, PerFunctionState &PFS); 398 int ParseAtomicRMW(Instruction *&I, PerFunctionState &PFS); 399 int ParseFence(Instruction *&I, PerFunctionState &PFS); 400 int ParseGetElementPtr(Instruction *&I, PerFunctionState &PFS); 401 int ParseExtractValue(Instruction *&I, PerFunctionState &PFS); 402 int ParseInsertValue(Instruction *&I, PerFunctionState &PFS); 403 404 bool ResolveForwardRefBlockAddresses(Function *TheFn, 405 std::vector<std::pair<ValID, GlobalValue*> > &Refs, 406 PerFunctionState *PFS); 407 }; 408 } // End llvm namespace 409 410 #endif 411
