1 //===--- CodeGenTypes.cpp - TBAA information for LLVM CodeGen -------------===//
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 is the code that manages TBAA information and defines the TBAA policy
11 // for the optimizer to use. Relevant standards text includes:
12 //
13 //   C99 6.5p7
14 //   C++ [basic.lval] (p10 in n3126, p15 in some earlier versions)
15 //
16 //===----------------------------------------------------------------------===//
17 
18 #include "CodeGenTBAA.h"
19 #include "clang/AST/ASTContext.h"
20 #include "clang/AST/Attr.h"
21 #include "clang/AST/Mangle.h"
22 #include "clang/AST/RecordLayout.h"
23 #include "clang/Frontend/CodeGenOptions.h"
24 #include "llvm/ADT/SmallSet.h"
25 #include "llvm/IR/Constants.h"
26 #include "llvm/IR/LLVMContext.h"
27 #include "llvm/IR/Metadata.h"
28 #include "llvm/IR/Type.h"
29 using namespace clang;
30 using namespace CodeGen;
31 
32 CodeGenTBAA::CodeGenTBAA(ASTContext &Ctx, llvm::LLVMContext& VMContext,
33                          const CodeGenOptions &CGO,
34                          const LangOptions &Features, MangleContext &MContext)
35   : Context(Ctx), CodeGenOpts(CGO), Features(Features), MContext(MContext),
36     MDHelper(VMContext), Root(nullptr), Char(nullptr) {
37 }
38 
39 CodeGenTBAA::~CodeGenTBAA() {
40 }
41 
42 llvm::MDNode *CodeGenTBAA::getRoot() {
43   // Define the root of the tree. This identifies the tree, so that
44   // if our LLVM IR is linked with LLVM IR from a different front-end
45   // (or a different version of this front-end), their TBAA trees will
46   // remain distinct, and the optimizer will treat them conservatively.
47   if (!Root) {
48     if (Features.CPlusPlus)
49       Root = MDHelper.createTBAARoot("Simple C++ TBAA");
50     else
51       Root = MDHelper.createTBAARoot("Simple C/C++ TBAA");
52   }
53 
54   return Root;
55 }
56 
57 // For both scalar TBAA and struct-path aware TBAA, the scalar type has the
58 // same format: name, parent node, and offset.
59 llvm::MDNode *CodeGenTBAA::createTBAAScalarType(StringRef Name,
60                                                 llvm::MDNode *Parent) {
61   return MDHelper.createTBAAScalarTypeNode(Name, Parent);
62 }
63 
64 llvm::MDNode *CodeGenTBAA::getChar() {
65   // Define the root of the tree for user-accessible memory. C and C++
66   // give special powers to char and certain similar types. However,
67   // these special powers only cover user-accessible memory, and doesn't
68   // include things like vtables.
69   if (!Char)
70     Char = createTBAAScalarType("omnipotent char", getRoot());
71 
72   return Char;
73 }
74 
75 static bool TypeHasMayAlias(QualType QTy) {
76   // Tagged types have declarations, and therefore may have attributes.
77   if (const TagType *TTy = dyn_cast<TagType>(QTy))
78     return TTy->getDecl()->hasAttr<MayAliasAttr>();
79 
80   // Typedef types have declarations, and therefore may have attributes.
81   if (const TypedefType *TTy = dyn_cast<TypedefType>(QTy)) {
82     if (TTy->getDecl()->hasAttr<MayAliasAttr>())
83       return true;
84     // Also, their underlying types may have relevant attributes.
85     return TypeHasMayAlias(TTy->desugar());
86   }
87 
88   return false;
89 }
90 
91 llvm::MDNode *CodeGenTBAA::getTypeInfo(QualType QTy) {
92   // At -O0 or relaxed aliasing, TBAA is not emitted for regular types.
93   if (CodeGenOpts.OptimizationLevel == 0 || CodeGenOpts.RelaxedAliasing)
94     return nullptr;
95 
96   // If the type has the may_alias attribute (even on a typedef), it is
97   // effectively in the general char alias class.
98   if (TypeHasMayAlias(QTy))
99     return getChar();
100 
101   const Type *Ty = Context.getCanonicalType(QTy).getTypePtr();
102 
103   if (llvm::MDNode *N = MetadataCache[Ty])
104     return N;
105 
106   // Handle builtin types.
107   if (const BuiltinType *BTy = dyn_cast<BuiltinType>(Ty)) {
108     switch (BTy->getKind()) {
109     // Character types are special and can alias anything.
110     // In C++, this technically only includes "char" and "unsigned char",
111     // and not "signed char". In C, it includes all three. For now,
112     // the risk of exploiting this detail in C++ seems likely to outweigh
113     // the benefit.
114     case BuiltinType::Char_U:
115     case BuiltinType::Char_S:
116     case BuiltinType::UChar:
117     case BuiltinType::SChar:
118       return getChar();
119 
120     // Unsigned types can alias their corresponding signed types.
121     case BuiltinType::UShort:
122       return getTypeInfo(Context.ShortTy);
123     case BuiltinType::UInt:
124       return getTypeInfo(Context.IntTy);
125     case BuiltinType::ULong:
126       return getTypeInfo(Context.LongTy);
127     case BuiltinType::ULongLong:
128       return getTypeInfo(Context.LongLongTy);
129     case BuiltinType::UInt128:
130       return getTypeInfo(Context.Int128Ty);
131 
132     // Treat all other builtin types as distinct types. This includes
133     // treating wchar_t, char16_t, and char32_t as distinct from their
134     // "underlying types".
135     default:
136       return MetadataCache[Ty] =
137         createTBAAScalarType(BTy->getName(Features), getChar());
138     }
139   }
140 
141   // C++1z [basic.lval]p10: "If a program attempts to access the stored value of
142   // an object through a glvalue of other than one of the following types the
143   // behavior is undefined: [...] a char, unsigned char, or std::byte type."
144   if (Ty->isStdByteType())
145     return MetadataCache[Ty] = getChar();
146 
147   // Handle pointers and references.
148   // TODO: Implement C++'s type "similarity" and consider dis-"similar"
149   // pointers distinct.
150   if (Ty->isPointerType() || Ty->isReferenceType())
151     return MetadataCache[Ty] = createTBAAScalarType("any pointer",
152                                                     getChar());
153 
154   // Enum types are distinct types. In C++ they have "underlying types",
155   // however they aren't related for TBAA.
156   if (const EnumType *ETy = dyn_cast<EnumType>(Ty)) {
157     // In C++ mode, types have linkage, so we can rely on the ODR and
158     // on their mangled names, if they're external.
159     // TODO: Is there a way to get a program-wide unique name for a
160     // decl with local linkage or no linkage?
161     if (!Features.CPlusPlus || !ETy->getDecl()->isExternallyVisible())
162       return MetadataCache[Ty] = getChar();
163 
164     SmallString<256> OutName;
165     llvm::raw_svector_ostream Out(OutName);
166     MContext.mangleTypeName(QualType(ETy, 0), Out);
167     return MetadataCache[Ty] = createTBAAScalarType(OutName, getChar());
168   }
169 
170   // For now, handle any other kind of type conservatively.
171   return MetadataCache[Ty] = getChar();
172 }
173 
174 TBAAAccessInfo CodeGenTBAA::getVTablePtrAccessInfo() {
175   return TBAAAccessInfo(createTBAAScalarType("vtable pointer", getRoot()));
176 }
177 
178 bool
179 CodeGenTBAA::CollectFields(uint64_t BaseOffset,
180                            QualType QTy,
181                            SmallVectorImpl<llvm::MDBuilder::TBAAStructField> &
182                              Fields,
183                            bool MayAlias) {
184   /* Things not handled yet include: C++ base classes, bitfields, */
185 
186   if (const RecordType *TTy = QTy->getAs<RecordType>()) {
187     const RecordDecl *RD = TTy->getDecl()->getDefinition();
188     if (RD->hasFlexibleArrayMember())
189       return false;
190 
191     // TODO: Handle C++ base classes.
192     if (const CXXRecordDecl *Decl = dyn_cast<CXXRecordDecl>(RD))
193       if (Decl->bases_begin() != Decl->bases_end())
194         return false;
195 
196     const ASTRecordLayout &Layout = Context.getASTRecordLayout(RD);
197 
198     unsigned idx = 0;
199     for (RecordDecl::field_iterator i = RD->field_begin(),
200          e = RD->field_end(); i != e; ++i, ++idx) {
201       uint64_t Offset = BaseOffset +
202                         Layout.getFieldOffset(idx) / Context.getCharWidth();
203       QualType FieldQTy = i->getType();
204       if (!CollectFields(Offset, FieldQTy, Fields,
205                          MayAlias || TypeHasMayAlias(FieldQTy)))
206         return false;
207     }
208     return true;
209   }
210 
211   /* Otherwise, treat whatever it is as a field. */
212   uint64_t Offset = BaseOffset;
213   uint64_t Size = Context.getTypeSizeInChars(QTy).getQuantity();
214   llvm::MDNode *TBAAType = MayAlias ? getChar() : getTypeInfo(QTy);
215   llvm::MDNode *TBAATag = getAccessTagInfo(TBAAAccessInfo(TBAAType));
216   Fields.push_back(llvm::MDBuilder::TBAAStructField(Offset, Size, TBAATag));
217   return true;
218 }
219 
220 llvm::MDNode *
221 CodeGenTBAA::getTBAAStructInfo(QualType QTy) {
222   const Type *Ty = Context.getCanonicalType(QTy).getTypePtr();
223 
224   if (llvm::MDNode *N = StructMetadataCache[Ty])
225     return N;
226 
227   SmallVector<llvm::MDBuilder::TBAAStructField, 4> Fields;
228   if (CollectFields(0, QTy, Fields, TypeHasMayAlias(QTy)))
229     return MDHelper.createTBAAStructNode(Fields);
230 
231   // For now, handle any other kind of type conservatively.
232   return StructMetadataCache[Ty] = nullptr;
233 }
234 
235 /// Check if the given type is a valid base type to be used in access tags.
236 static bool isValidBaseType(QualType QTy) {
237   if (const RecordType *TTy = QTy->getAs<RecordType>()) {
238     const RecordDecl *RD = TTy->getDecl()->getDefinition();
239     if (RD->hasFlexibleArrayMember())
240       return false;
241     // RD can be struct, union, class, interface or enum.
242     // For now, we only handle struct and class.
243     if (RD->isStruct() || RD->isClass())
244       return true;
245   }
246   return false;
247 }
248 
249 llvm::MDNode *CodeGenTBAA::getBaseTypeInfo(QualType QTy) {
250   if (!isValidBaseType(QTy))
251     return nullptr;
252 
253   const Type *Ty = Context.getCanonicalType(QTy).getTypePtr();
254   if (llvm::MDNode *N = BaseTypeMetadataCache[Ty])
255     return N;
256 
257   if (const RecordType *TTy = QTy->getAs<RecordType>()) {
258     const RecordDecl *RD = TTy->getDecl()->getDefinition();
259 
260     const ASTRecordLayout &Layout = Context.getASTRecordLayout(RD);
261     SmallVector <std::pair<llvm::MDNode*, uint64_t>, 4> Fields;
262     unsigned idx = 0;
263     for (RecordDecl::field_iterator i = RD->field_begin(),
264          e = RD->field_end(); i != e; ++i, ++idx) {
265       QualType FieldQTy = i->getType();
266       llvm::MDNode *FieldNode = isValidBaseType(FieldQTy) ?
267           getBaseTypeInfo(FieldQTy) : getTypeInfo(FieldQTy);
268       if (!FieldNode)
269         return BaseTypeMetadataCache[Ty] = nullptr;
270       Fields.push_back(std::make_pair(
271           FieldNode, Layout.getFieldOffset(idx) / Context.getCharWidth()));
272     }
273 
274     SmallString<256> OutName;
275     if (Features.CPlusPlus) {
276       // Don't use the mangler for C code.
277       llvm::raw_svector_ostream Out(OutName);
278       MContext.mangleTypeName(QualType(Ty, 0), Out);
279     } else {
280       OutName = RD->getName();
281     }
282     // Create the struct type node with a vector of pairs (offset, type).
283     return BaseTypeMetadataCache[Ty] =
284       MDHelper.createTBAAStructTypeNode(OutName, Fields);
285   }
286 
287   return BaseTypeMetadataCache[Ty] = nullptr;
288 }
289 
290 llvm::MDNode *CodeGenTBAA::getAccessTagInfo(TBAAAccessInfo Info) {
291   if (!Info.AccessType)
292     return nullptr;
293 
294   if (!CodeGenOpts.StructPathTBAA)
295     Info = TBAAAccessInfo(Info.AccessType);
296 
297   llvm::MDNode *&N = AccessTagMetadataCache[Info];
298   if (N)
299     return N;
300 
301   if (!Info.BaseType) {
302     Info.BaseType = Info.AccessType;
303     assert(!Info.Offset && "Nonzero offset for an access with no base type!");
304   }
305   return N = MDHelper.createTBAAStructTagNode(Info.BaseType, Info.AccessType,
306                                               Info.Offset);
307 }
308 
309 TBAAAccessInfo CodeGenTBAA::getMayAliasAccessInfo() {
310   return TBAAAccessInfo(getChar());
311 }
312 
313 TBAAAccessInfo CodeGenTBAA::mergeTBAAInfoForCast(TBAAAccessInfo SourceInfo,
314                                                  TBAAAccessInfo TargetInfo) {
315   TBAAAccessInfo MayAliasInfo = getMayAliasAccessInfo();
316   if (SourceInfo == MayAliasInfo || TargetInfo == MayAliasInfo)
317     return MayAliasInfo;
318   return TargetInfo;
319 }
320