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     bool ParseUInt64(uint64_t &Val);
206     bool ParseUInt64(uint64_t &Val, LocTy &Loc) {
207       Loc = Lex.getLoc();
208       return ParseUInt64(Val);
209     }
210 
211     bool ParseTLSModel(GlobalVariable::ThreadLocalMode &TLM);
212     bool ParseOptionalThreadLocal(GlobalVariable::ThreadLocalMode &TLM);
213     bool parseOptionalUnnamedAddr(bool &UnnamedAddr) {
214       return ParseOptionalToken(lltok::kw_unnamed_addr, UnnamedAddr);
215     }
216     bool ParseOptionalAddrSpace(unsigned &AddrSpace);
217     bool ParseOptionalParamAttrs(AttrBuilder &B);
218     bool ParseOptionalReturnAttrs(AttrBuilder &B);
219     bool ParseOptionalLinkage(unsigned &Linkage, bool &HasLinkage);
220     bool ParseOptionalLinkage(unsigned &Linkage) {
221       bool HasLinkage; return ParseOptionalLinkage(Linkage, HasLinkage);
222     }
223     bool ParseOptionalVisibility(unsigned &Visibility);
224     bool ParseOptionalDLLStorageClass(unsigned &DLLStorageClass);
225     bool ParseOptionalCallingConv(CallingConv::ID &CC);
226     bool ParseOptionalAlignment(unsigned &Alignment);
227     bool ParseOptionalDereferenceableBytes(uint64_t &Bytes);
228     bool ParseScopeAndOrdering(bool isAtomic, SynchronizationScope &Scope,
229                                AtomicOrdering &Ordering);
230     bool ParseOrdering(AtomicOrdering &Ordering);
231     bool ParseOptionalStackAlignment(unsigned &Alignment);
232     bool ParseOptionalCommaAlign(unsigned &Alignment, bool &AteExtraComma);
233     bool ParseOptionalCommaInAlloca(bool &IsInAlloca);
234     bool ParseIndexList(SmallVectorImpl<unsigned> &Indices,bool &AteExtraComma);
235     bool ParseIndexList(SmallVectorImpl<unsigned> &Indices) {
236       bool AteExtraComma;
237       if (ParseIndexList(Indices, AteExtraComma)) return true;
238       if (AteExtraComma)
239         return TokError("expected index");
240       return false;
241     }
242 
243     // Top-Level Entities
244     bool ParseTopLevelEntities();
245     bool ValidateEndOfModule();
246     bool ParseTargetDefinition();
247     bool ParseModuleAsm();
248     bool ParseDepLibs();        // FIXME: Remove in 4.0.
249     bool ParseUnnamedType();
250     bool ParseNamedType();
251     bool ParseDeclare();
252     bool ParseDefine();
253 
254     bool ParseGlobalType(bool &IsConstant);
255     bool ParseUnnamedGlobal();
256     bool ParseNamedGlobal();
257     bool ParseGlobal(const std::string &Name, LocTy Loc, unsigned Linkage,
258                      bool HasLinkage, unsigned Visibility,
259                      unsigned DLLStorageClass,
260                      GlobalVariable::ThreadLocalMode TLM, bool UnnamedAddr);
261     bool ParseAlias(const std::string &Name, LocTy Loc, unsigned Visibility,
262                     unsigned DLLStorageClass,
263                     GlobalVariable::ThreadLocalMode TLM, bool UnnamedAddr);
264     bool parseComdat();
265     bool ParseStandaloneMetadata();
266     bool ParseNamedMetadata();
267     bool ParseMDString(MDString *&Result);
268     bool ParseMDNodeID(MDNode *&Result);
269     bool ParseMDNodeID(MDNode *&Result, unsigned &SlotNo);
270     bool ParseUnnamedAttrGrp();
271     bool ParseFnAttributeValuePairs(AttrBuilder &B,
272                                     std::vector<unsigned> &FwdRefAttrGrps,
273                                     bool inAttrGrp, LocTy &BuiltinLoc);
274 
275     // Type Parsing.
276     bool ParseType(Type *&Result, bool AllowVoid = false);
277     bool ParseType(Type *&Result, LocTy &Loc, bool AllowVoid = false) {
278       Loc = Lex.getLoc();
279       return ParseType(Result, AllowVoid);
280     }
281     bool ParseAnonStructType(Type *&Result, bool Packed);
282     bool ParseStructBody(SmallVectorImpl<Type*> &Body);
283     bool ParseStructDefinition(SMLoc TypeLoc, StringRef Name,
284                                std::pair<Type*, LocTy> &Entry,
285                                Type *&ResultTy);
286 
287     bool ParseArrayVectorType(Type *&Result, bool isVector);
288     bool ParseFunctionType(Type *&Result);
289 
290     // Function Semantic Analysis.
291     class PerFunctionState {
292       LLParser &P;
293       Function &F;
294       std::map<std::string, std::pair<Value*, LocTy> > ForwardRefVals;
295       std::map<unsigned, std::pair<Value*, LocTy> > ForwardRefValIDs;
296       std::vector<Value*> NumberedVals;
297 
298       /// FunctionNumber - If this is an unnamed function, this is the slot
299       /// number of it, otherwise it is -1.
300       int FunctionNumber;
301     public:
302       PerFunctionState(LLParser &p, Function &f, int FunctionNumber);
303       ~PerFunctionState();
304 
305       Function &getFunction() const { return F; }
306 
307       bool FinishFunction();
308 
309       /// GetVal - Get a value with the specified name or ID, creating a
310       /// forward reference record if needed.  This can return null if the value
311       /// exists but does not have the right type.
312       Value *GetVal(const std::string &Name, Type *Ty, LocTy Loc);
313       Value *GetVal(unsigned ID, Type *Ty, LocTy Loc);
314 
315       /// SetInstName - After an instruction is parsed and inserted into its
316       /// basic block, this installs its name.
317       bool SetInstName(int NameID, const std::string &NameStr, LocTy NameLoc,
318                        Instruction *Inst);
319 
320       /// GetBB - Get a basic block with the specified name or ID, creating a
321       /// forward reference record if needed.  This can return null if the value
322       /// is not a BasicBlock.
323       BasicBlock *GetBB(const std::string &Name, LocTy Loc);
324       BasicBlock *GetBB(unsigned ID, LocTy Loc);
325 
326       /// DefineBB - Define the specified basic block, which is either named or
327       /// unnamed.  If there is an error, this returns null otherwise it returns
328       /// the block being defined.
329       BasicBlock *DefineBB(const std::string &Name, LocTy Loc);
330     };
331 
332     bool ConvertValIDToValue(Type *Ty, ValID &ID, Value *&V,
333                              PerFunctionState *PFS);
334 
335     bool ParseValue(Type *Ty, Value *&V, PerFunctionState *PFS);
336     bool ParseValue(Type *Ty, Value *&V, PerFunctionState &PFS) {
337       return ParseValue(Ty, V, &PFS);
338     }
339     bool ParseValue(Type *Ty, Value *&V, LocTy &Loc,
340                     PerFunctionState &PFS) {
341       Loc = Lex.getLoc();
342       return ParseValue(Ty, V, &PFS);
343     }
344 
345     bool ParseTypeAndValue(Value *&V, PerFunctionState *PFS);
346     bool ParseTypeAndValue(Value *&V, PerFunctionState &PFS) {
347       return ParseTypeAndValue(V, &PFS);
348     }
349     bool ParseTypeAndValue(Value *&V, LocTy &Loc, PerFunctionState &PFS) {
350       Loc = Lex.getLoc();
351       return ParseTypeAndValue(V, PFS);
352     }
353     bool ParseTypeAndBasicBlock(BasicBlock *&BB, LocTy &Loc,
354                                 PerFunctionState &PFS);
355     bool ParseTypeAndBasicBlock(BasicBlock *&BB, PerFunctionState &PFS) {
356       LocTy Loc;
357       return ParseTypeAndBasicBlock(BB, Loc, PFS);
358     }
359 
360 
361     struct ParamInfo {
362       LocTy Loc;
363       Value *V;
364       AttributeSet Attrs;
365       ParamInfo(LocTy loc, Value *v, AttributeSet attrs)
366         : Loc(loc), V(v), Attrs(attrs) {}
367     };
368     bool ParseParameterList(SmallVectorImpl<ParamInfo> &ArgList,
369                             PerFunctionState &PFS);
370 
371     // Constant Parsing.
372     bool ParseValID(ValID &ID, PerFunctionState *PFS = nullptr);
373     bool ParseGlobalValue(Type *Ty, Constant *&V);
374     bool ParseGlobalTypeAndValue(Constant *&V);
375     bool ParseGlobalValueVector(SmallVectorImpl<Constant*> &Elts);
376     bool parseOptionalComdat(Comdat *&C);
377     bool ParseMetadataListValue(ValID &ID, PerFunctionState *PFS);
378     bool ParseMetadataValue(ValID &ID, PerFunctionState *PFS);
379     bool ParseMDNodeVector(SmallVectorImpl<Value*> &, PerFunctionState *PFS);
380     bool ParseInstructionMetadata(Instruction *Inst, PerFunctionState *PFS);
381 
382     // Function Parsing.
383     struct ArgInfo {
384       LocTy Loc;
385       Type *Ty;
386       AttributeSet Attrs;
387       std::string Name;
388       ArgInfo(LocTy L, Type *ty, AttributeSet Attr, const std::string &N)
389         : Loc(L), Ty(ty), Attrs(Attr), Name(N) {}
390     };
391     bool ParseArgumentList(SmallVectorImpl<ArgInfo> &ArgList, bool &isVarArg);
392     bool ParseFunctionHeader(Function *&Fn, bool isDefine);
393     bool ParseFunctionBody(Function &Fn);
394     bool ParseBasicBlock(PerFunctionState &PFS);
395 
396     enum TailCallType { TCT_None, TCT_Tail, TCT_MustTail };
397 
398     // Instruction Parsing.  Each instruction parsing routine can return with a
399     // normal result, an error result, or return having eaten an extra comma.
400     enum InstResult { InstNormal = 0, InstError = 1, InstExtraComma = 2 };
401     int ParseInstruction(Instruction *&Inst, BasicBlock *BB,
402                          PerFunctionState &PFS);
403     bool ParseCmpPredicate(unsigned &Pred, unsigned Opc);
404 
405     bool ParseRet(Instruction *&Inst, BasicBlock *BB, PerFunctionState &PFS);
406     bool ParseBr(Instruction *&Inst, PerFunctionState &PFS);
407     bool ParseSwitch(Instruction *&Inst, PerFunctionState &PFS);
408     bool ParseIndirectBr(Instruction *&Inst, PerFunctionState &PFS);
409     bool ParseInvoke(Instruction *&Inst, PerFunctionState &PFS);
410     bool ParseResume(Instruction *&Inst, PerFunctionState &PFS);
411 
412     bool ParseArithmetic(Instruction *&I, PerFunctionState &PFS, unsigned Opc,
413                          unsigned OperandType);
414     bool ParseLogical(Instruction *&I, PerFunctionState &PFS, unsigned Opc);
415     bool ParseCompare(Instruction *&I, PerFunctionState &PFS, unsigned Opc);
416     bool ParseCast(Instruction *&I, PerFunctionState &PFS, unsigned Opc);
417     bool ParseSelect(Instruction *&I, PerFunctionState &PFS);
418     bool ParseVA_Arg(Instruction *&I, PerFunctionState &PFS);
419     bool ParseExtractElement(Instruction *&I, PerFunctionState &PFS);
420     bool ParseInsertElement(Instruction *&I, PerFunctionState &PFS);
421     bool ParseShuffleVector(Instruction *&I, PerFunctionState &PFS);
422     int ParsePHI(Instruction *&I, PerFunctionState &PFS);
423     bool ParseLandingPad(Instruction *&I, PerFunctionState &PFS);
424     bool ParseCall(Instruction *&I, PerFunctionState &PFS,
425                    CallInst::TailCallKind IsTail);
426     int ParseAlloc(Instruction *&I, PerFunctionState &PFS);
427     int ParseLoad(Instruction *&I, PerFunctionState &PFS);
428     int ParseStore(Instruction *&I, PerFunctionState &PFS);
429     int ParseCmpXchg(Instruction *&I, PerFunctionState &PFS);
430     int ParseAtomicRMW(Instruction *&I, PerFunctionState &PFS);
431     int ParseFence(Instruction *&I, PerFunctionState &PFS);
432     int ParseGetElementPtr(Instruction *&I, PerFunctionState &PFS);
433     int ParseExtractValue(Instruction *&I, PerFunctionState &PFS);
434     int ParseInsertValue(Instruction *&I, PerFunctionState &PFS);
435 
436     bool ResolveForwardRefBlockAddresses(Function *TheFn,
437                              std::vector<std::pair<ValID, GlobalValue*> > &Refs,
438                                          PerFunctionState *PFS);
439   };
440 } // End llvm namespace
441 
442 #endif
443