1 //===-- CodeGenTBAA.cpp - TBAA information for LLVM CodeGen ---------------===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 //
9 // This is the code that manages TBAA information and defines the TBAA policy
10 // for the optimizer to use. Relevant standards text includes:
11 //
12 //   C99 6.5p7
13 //   C++ [basic.lval] (p10 in n3126, p15 in some earlier versions)
14 //
15 //===----------------------------------------------------------------------===//
16 
17 #include "CodeGenTBAA.h"
18 #include "clang/AST/ASTContext.h"
19 #include "clang/AST/Attr.h"
20 #include "clang/AST/Mangle.h"
21 #include "clang/AST/RecordLayout.h"
22 #include "clang/Basic/CodeGenOptions.h"
23 #include "llvm/ADT/SmallSet.h"
24 #include "llvm/IR/Constants.h"
25 #include "llvm/IR/LLVMContext.h"
26 #include "llvm/IR/Metadata.h"
27 #include "llvm/IR/Module.h"
28 #include "llvm/IR/Type.h"
29 using namespace clang;
30 using namespace CodeGen;
31 
32 CodeGenTBAA::CodeGenTBAA(ASTContext &Ctx, llvm::Module &M,
33                          const CodeGenOptions &CGO,
34                          const LangOptions &Features, MangleContext &MContext)
35   : Context(Ctx), Module(M), CodeGenOpts(CGO),
36     Features(Features), MContext(MContext), MDHelper(M.getContext()),
37     Root(nullptr), Char(nullptr)
38 {}
39 
40 CodeGenTBAA::~CodeGenTBAA() {
41 }
42 
43 llvm::MDNode *CodeGenTBAA::getRoot() {
44   // Define the root of the tree. This identifies the tree, so that
45   // if our LLVM IR is linked with LLVM IR from a different front-end
46   // (or a different version of this front-end), their TBAA trees will
47   // remain distinct, and the optimizer will treat them conservatively.
48   if (!Root) {
49     if (Features.CPlusPlus)
50       Root = MDHelper.createTBAARoot("Simple C++ TBAA");
51     else
52       Root = MDHelper.createTBAARoot("Simple C/C++ TBAA");
53   }
54 
55   return Root;
56 }
57 
58 llvm::MDNode *CodeGenTBAA::createScalarTypeNode(StringRef Name,
59                                                 llvm::MDNode *Parent,
60                                                 uint64_t Size) {
61   if (CodeGenOpts.NewStructPathTBAA) {
62     llvm::Metadata *Id = MDHelper.createString(Name);
63     return MDHelper.createTBAATypeNode(Parent, Size, Id);
64   }
65   return MDHelper.createTBAAScalarTypeNode(Name, Parent);
66 }
67 
68 llvm::MDNode *CodeGenTBAA::getChar() {
69   // Define the root of the tree for user-accessible memory. C and C++
70   // give special powers to char and certain similar types. However,
71   // these special powers only cover user-accessible memory, and doesn't
72   // include things like vtables.
73   if (!Char)
74     Char = createScalarTypeNode("omnipotent char", getRoot(), /* Size= */ 1);
75 
76   return Char;
77 }
78 
79 static bool TypeHasMayAlias(QualType QTy) {
80   // Tagged types have declarations, and therefore may have attributes.
81   if (auto *TD = QTy->getAsTagDecl())
82     if (TD->hasAttr<MayAliasAttr>())
83       return true;
84 
85   // Also look for may_alias as a declaration attribute on a typedef.
86   // FIXME: We should follow GCC and model may_alias as a type attribute
87   // rather than as a declaration attribute.
88   while (auto *TT = QTy->getAs<TypedefType>()) {
89     if (TT->getDecl()->hasAttr<MayAliasAttr>())
90       return true;
91     QTy = TT->desugar();
92   }
93   return false;
94 }
95 
96 /// Check if the given type is a valid base type to be used in access tags.
97 static bool isValidBaseType(QualType QTy) {
98   if (QTy->isReferenceType())
99     return false;
100   if (const RecordType *TTy = QTy->getAs<RecordType>()) {
101     const RecordDecl *RD = TTy->getDecl()->getDefinition();
102     // Incomplete types are not valid base access types.
103     if (!RD)
104       return false;
105     if (RD->hasFlexibleArrayMember())
106       return false;
107     // RD can be struct, union, class, interface or enum.
108     // For now, we only handle struct and class.
109     if (RD->isStruct() || RD->isClass())
110       return true;
111   }
112   return false;
113 }
114 
115 llvm::MDNode *CodeGenTBAA::getTypeInfoHelper(const Type *Ty) {
116   uint64_t Size = Context.getTypeSizeInChars(Ty).getQuantity();
117 
118   // Handle builtin types.
119   if (const BuiltinType *BTy = dyn_cast<BuiltinType>(Ty)) {
120     switch (BTy->getKind()) {
121     // Character types are special and can alias anything.
122     // In C++, this technically only includes "char" and "unsigned char",
123     // and not "signed char". In C, it includes all three. For now,
124     // the risk of exploiting this detail in C++ seems likely to outweigh
125     // the benefit.
126     case BuiltinType::Char_U:
127     case BuiltinType::Char_S:
128     case BuiltinType::UChar:
129     case BuiltinType::SChar:
130       return getChar();
131 
132     // Unsigned types can alias their corresponding signed types.
133     case BuiltinType::UShort:
134       return getTypeInfo(Context.ShortTy);
135     case BuiltinType::UInt:
136       return getTypeInfo(Context.IntTy);
137     case BuiltinType::ULong:
138       return getTypeInfo(Context.LongTy);
139     case BuiltinType::ULongLong:
140       return getTypeInfo(Context.LongLongTy);
141     case BuiltinType::UInt128:
142       return getTypeInfo(Context.Int128Ty);
143 
144     case BuiltinType::UShortFract:
145       return getTypeInfo(Context.ShortFractTy);
146     case BuiltinType::UFract:
147       return getTypeInfo(Context.FractTy);
148     case BuiltinType::ULongFract:
149       return getTypeInfo(Context.LongFractTy);
150 
151     case BuiltinType::SatUShortFract:
152       return getTypeInfo(Context.SatShortFractTy);
153     case BuiltinType::SatUFract:
154       return getTypeInfo(Context.SatFractTy);
155     case BuiltinType::SatULongFract:
156       return getTypeInfo(Context.SatLongFractTy);
157 
158     case BuiltinType::UShortAccum:
159       return getTypeInfo(Context.ShortAccumTy);
160     case BuiltinType::UAccum:
161       return getTypeInfo(Context.AccumTy);
162     case BuiltinType::ULongAccum:
163       return getTypeInfo(Context.LongAccumTy);
164 
165     case BuiltinType::SatUShortAccum:
166       return getTypeInfo(Context.SatShortAccumTy);
167     case BuiltinType::SatUAccum:
168       return getTypeInfo(Context.SatAccumTy);
169     case BuiltinType::SatULongAccum:
170       return getTypeInfo(Context.SatLongAccumTy);
171 
172     // Treat all other builtin types as distinct types. This includes
173     // treating wchar_t, char16_t, and char32_t as distinct from their
174     // "underlying types".
175     default:
176       return createScalarTypeNode(BTy->getName(Features), getChar(), Size);
177     }
178   }
179 
180   // C++1z [basic.lval]p10: "If a program attempts to access the stored value of
181   // an object through a glvalue of other than one of the following types the
182   // behavior is undefined: [...] a char, unsigned char, or std::byte type."
183   if (Ty->isStdByteType())
184     return getChar();
185 
186   // Handle pointers and references.
187   // TODO: Implement C++'s type "similarity" and consider dis-"similar"
188   // pointers distinct.
189   if (Ty->isPointerType() || Ty->isReferenceType())
190     return createScalarTypeNode("any pointer", getChar(), Size);
191 
192   // Accesses to arrays are accesses to objects of their element types.
193   if (CodeGenOpts.NewStructPathTBAA && Ty->isArrayType())
194     return getTypeInfo(cast<ArrayType>(Ty)->getElementType());
195 
196   // Enum types are distinct types. In C++ they have "underlying types",
197   // however they aren't related for TBAA.
198   if (const EnumType *ETy = dyn_cast<EnumType>(Ty)) {
199     // In C++ mode, types have linkage, so we can rely on the ODR and
200     // on their mangled names, if they're external.
201     // TODO: Is there a way to get a program-wide unique name for a
202     // decl with local linkage or no linkage?
203     if (!Features.CPlusPlus || !ETy->getDecl()->isExternallyVisible())
204       return getChar();
205 
206     SmallString<256> OutName;
207     llvm::raw_svector_ostream Out(OutName);
208     MContext.mangleTypeName(QualType(ETy, 0), Out);
209     return createScalarTypeNode(OutName, getChar(), Size);
210   }
211 
212   // For now, handle any other kind of type conservatively.
213   return getChar();
214 }
215 
216 llvm::MDNode *CodeGenTBAA::getTypeInfo(QualType QTy) {
217   // At -O0 or relaxed aliasing, TBAA is not emitted for regular types.
218   if (CodeGenOpts.OptimizationLevel == 0 || CodeGenOpts.RelaxedAliasing)
219     return nullptr;
220 
221   // If the type has the may_alias attribute (even on a typedef), it is
222   // effectively in the general char alias class.
223   if (TypeHasMayAlias(QTy))
224     return getChar();
225 
226   // We need this function to not fall back to returning the "omnipotent char"
227   // type node for aggregate and union types. Otherwise, any dereference of an
228   // aggregate will result into the may-alias access descriptor, meaning all
229   // subsequent accesses to direct and indirect members of that aggregate will
230   // be considered may-alias too.
231   // TODO: Combine getTypeInfo() and getBaseTypeInfo() into a single function.
232   if (isValidBaseType(QTy))
233     return getBaseTypeInfo(QTy);
234 
235   const Type *Ty = Context.getCanonicalType(QTy).getTypePtr();
236   if (llvm::MDNode *N = MetadataCache[Ty])
237     return N;
238 
239   // Note that the following helper call is allowed to add new nodes to the
240   // cache, which invalidates all its previously obtained iterators. So we
241   // first generate the node for the type and then add that node to the cache.
242   llvm::MDNode *TypeNode = getTypeInfoHelper(Ty);
243   return MetadataCache[Ty] = TypeNode;
244 }
245 
246 TBAAAccessInfo CodeGenTBAA::getAccessInfo(QualType AccessType) {
247   // Pointee values may have incomplete types, but they shall never be
248   // dereferenced.
249   if (AccessType->isIncompleteType())
250     return TBAAAccessInfo::getIncompleteInfo();
251 
252   if (TypeHasMayAlias(AccessType))
253     return TBAAAccessInfo::getMayAliasInfo();
254 
255   uint64_t Size = Context.getTypeSizeInChars(AccessType).getQuantity();
256   return TBAAAccessInfo(getTypeInfo(AccessType), Size);
257 }
258 
259 TBAAAccessInfo CodeGenTBAA::getVTablePtrAccessInfo(llvm::Type *VTablePtrType) {
260   llvm::DataLayout DL(&Module);
261   unsigned Size = DL.getPointerTypeSize(VTablePtrType);
262   return TBAAAccessInfo(createScalarTypeNode("vtable pointer", getRoot(), Size),
263                         Size);
264 }
265 
266 bool
267 CodeGenTBAA::CollectFields(uint64_t BaseOffset,
268                            QualType QTy,
269                            SmallVectorImpl<llvm::MDBuilder::TBAAStructField> &
270                              Fields,
271                            bool MayAlias) {
272   /* Things not handled yet include: C++ base classes, bitfields, */
273 
274   if (const RecordType *TTy = QTy->getAs<RecordType>()) {
275     const RecordDecl *RD = TTy->getDecl()->getDefinition();
276     if (RD->hasFlexibleArrayMember())
277       return false;
278 
279     // TODO: Handle C++ base classes.
280     if (const CXXRecordDecl *Decl = dyn_cast<CXXRecordDecl>(RD))
281       if (Decl->bases_begin() != Decl->bases_end())
282         return false;
283 
284     const ASTRecordLayout &Layout = Context.getASTRecordLayout(RD);
285 
286     unsigned idx = 0;
287     for (RecordDecl::field_iterator i = RD->field_begin(),
288          e = RD->field_end(); i != e; ++i, ++idx) {
289       if ((*i)->isZeroSize(Context) || (*i)->isUnnamedBitfield())
290         continue;
291       uint64_t Offset = BaseOffset +
292                         Layout.getFieldOffset(idx) / Context.getCharWidth();
293       QualType FieldQTy = i->getType();
294       if (!CollectFields(Offset, FieldQTy, Fields,
295                          MayAlias || TypeHasMayAlias(FieldQTy)))
296         return false;
297     }
298     return true;
299   }
300 
301   /* Otherwise, treat whatever it is as a field. */
302   uint64_t Offset = BaseOffset;
303   uint64_t Size = Context.getTypeSizeInChars(QTy).getQuantity();
304   llvm::MDNode *TBAAType = MayAlias ? getChar() : getTypeInfo(QTy);
305   llvm::MDNode *TBAATag = getAccessTagInfo(TBAAAccessInfo(TBAAType, Size));
306   Fields.push_back(llvm::MDBuilder::TBAAStructField(Offset, Size, TBAATag));
307   return true;
308 }
309 
310 llvm::MDNode *
311 CodeGenTBAA::getTBAAStructInfo(QualType QTy) {
312   const Type *Ty = Context.getCanonicalType(QTy).getTypePtr();
313 
314   if (llvm::MDNode *N = StructMetadataCache[Ty])
315     return N;
316 
317   SmallVector<llvm::MDBuilder::TBAAStructField, 4> Fields;
318   if (CollectFields(0, QTy, Fields, TypeHasMayAlias(QTy)))
319     return MDHelper.createTBAAStructNode(Fields);
320 
321   // For now, handle any other kind of type conservatively.
322   return StructMetadataCache[Ty] = nullptr;
323 }
324 
325 llvm::MDNode *CodeGenTBAA::getBaseTypeInfoHelper(const Type *Ty) {
326   if (auto *TTy = dyn_cast<RecordType>(Ty)) {
327     const RecordDecl *RD = TTy->getDecl()->getDefinition();
328     const ASTRecordLayout &Layout = Context.getASTRecordLayout(RD);
329     SmallVector<llvm::MDBuilder::TBAAStructField, 4> Fields;
330     for (FieldDecl *Field : RD->fields()) {
331       if (Field->isZeroSize(Context) || Field->isUnnamedBitfield())
332         continue;
333       QualType FieldQTy = Field->getType();
334       llvm::MDNode *TypeNode = isValidBaseType(FieldQTy) ?
335           getBaseTypeInfo(FieldQTy) : getTypeInfo(FieldQTy);
336       if (!TypeNode)
337         return BaseTypeMetadataCache[Ty] = nullptr;
338 
339       uint64_t BitOffset = Layout.getFieldOffset(Field->getFieldIndex());
340       uint64_t Offset = Context.toCharUnitsFromBits(BitOffset).getQuantity();
341       uint64_t Size = Context.getTypeSizeInChars(FieldQTy).getQuantity();
342       Fields.push_back(llvm::MDBuilder::TBAAStructField(Offset, Size,
343                                                         TypeNode));
344     }
345 
346     SmallString<256> OutName;
347     if (Features.CPlusPlus) {
348       // Don't use the mangler for C code.
349       llvm::raw_svector_ostream Out(OutName);
350       MContext.mangleTypeName(QualType(Ty, 0), Out);
351     } else {
352       OutName = RD->getName();
353     }
354 
355     if (CodeGenOpts.NewStructPathTBAA) {
356       llvm::MDNode *Parent = getChar();
357       uint64_t Size = Context.getTypeSizeInChars(Ty).getQuantity();
358       llvm::Metadata *Id = MDHelper.createString(OutName);
359       return MDHelper.createTBAATypeNode(Parent, Size, Id, Fields);
360     }
361 
362     // Create the struct type node with a vector of pairs (offset, type).
363     SmallVector<std::pair<llvm::MDNode*, uint64_t>, 4> OffsetsAndTypes;
364     for (const auto &Field : Fields)
365         OffsetsAndTypes.push_back(std::make_pair(Field.Type, Field.Offset));
366     return MDHelper.createTBAAStructTypeNode(OutName, OffsetsAndTypes);
367   }
368 
369   return nullptr;
370 }
371 
372 llvm::MDNode *CodeGenTBAA::getBaseTypeInfo(QualType QTy) {
373   if (!isValidBaseType(QTy))
374     return nullptr;
375 
376   const Type *Ty = Context.getCanonicalType(QTy).getTypePtr();
377   if (llvm::MDNode *N = BaseTypeMetadataCache[Ty])
378     return N;
379 
380   // Note that the following helper call is allowed to add new nodes to the
381   // cache, which invalidates all its previously obtained iterators. So we
382   // first generate the node for the type and then add that node to the cache.
383   llvm::MDNode *TypeNode = getBaseTypeInfoHelper(Ty);
384   return BaseTypeMetadataCache[Ty] = TypeNode;
385 }
386 
387 llvm::MDNode *CodeGenTBAA::getAccessTagInfo(TBAAAccessInfo Info) {
388   assert(!Info.isIncomplete() && "Access to an object of an incomplete type!");
389 
390   if (Info.isMayAlias())
391     Info = TBAAAccessInfo(getChar(), Info.Size);
392 
393   if (!Info.AccessType)
394     return nullptr;
395 
396   if (!CodeGenOpts.StructPathTBAA)
397     Info = TBAAAccessInfo(Info.AccessType, Info.Size);
398 
399   llvm::MDNode *&N = AccessTagMetadataCache[Info];
400   if (N)
401     return N;
402 
403   if (!Info.BaseType) {
404     Info.BaseType = Info.AccessType;
405     assert(!Info.Offset && "Nonzero offset for an access with no base type!");
406   }
407   if (CodeGenOpts.NewStructPathTBAA) {
408     return N = MDHelper.createTBAAAccessTag(Info.BaseType, Info.AccessType,
409                                             Info.Offset, Info.Size);
410   }
411   return N = MDHelper.createTBAAStructTagNode(Info.BaseType, Info.AccessType,
412                                               Info.Offset);
413 }
414 
415 TBAAAccessInfo CodeGenTBAA::mergeTBAAInfoForCast(TBAAAccessInfo SourceInfo,
416                                                  TBAAAccessInfo TargetInfo) {
417   if (SourceInfo.isMayAlias() || TargetInfo.isMayAlias())
418     return TBAAAccessInfo::getMayAliasInfo();
419   return TargetInfo;
420 }
421 
422 TBAAAccessInfo
423 CodeGenTBAA::mergeTBAAInfoForConditionalOperator(TBAAAccessInfo InfoA,
424                                                  TBAAAccessInfo InfoB) {
425   if (InfoA == InfoB)
426     return InfoA;
427 
428   if (!InfoA || !InfoB)
429     return TBAAAccessInfo();
430 
431   if (InfoA.isMayAlias() || InfoB.isMayAlias())
432     return TBAAAccessInfo::getMayAliasInfo();
433 
434   // TODO: Implement the rest of the logic here. For example, two accesses
435   // with same final access types result in an access to an object of that final
436   // access type regardless of their base types.
437   return TBAAAccessInfo::getMayAliasInfo();
438 }
439 
440 TBAAAccessInfo
441 CodeGenTBAA::mergeTBAAInfoForMemoryTransfer(TBAAAccessInfo DestInfo,
442                                             TBAAAccessInfo SrcInfo) {
443   if (DestInfo == SrcInfo)
444     return DestInfo;
445 
446   if (!DestInfo || !SrcInfo)
447     return TBAAAccessInfo();
448 
449   if (DestInfo.isMayAlias() || SrcInfo.isMayAlias())
450     return TBAAAccessInfo::getMayAliasInfo();
451 
452   // TODO: Implement the rest of the logic here. For example, two accesses
453   // with same final access types result in an access to an object of that final
454   // access type regardless of their base types.
455   return TBAAAccessInfo::getMayAliasInfo();
456 }
457