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(0), Char(0) {
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     Root = MDHelper.createTBAARoot("Simple C/C++ TBAA");
49 
50   return Root;
51 }
52 
53 llvm::MDNode *CodeGenTBAA::getChar() {
54   // Define the root of the tree for user-accessible memory. C and C++
55   // give special powers to char and certain similar types. However,
56   // these special powers only cover user-accessible memory, and doesn't
57   // include things like vtables.
58   if (!Char)
59     Char = MDHelper.createTBAANode("omnipotent char", getRoot());
60 
61   return Char;
62 }
63 
64 static bool TypeHasMayAlias(QualType QTy) {
65   // Tagged types have declarations, and therefore may have attributes.
66   if (const TagType *TTy = dyn_cast<TagType>(QTy))
67     return TTy->getDecl()->hasAttr<MayAliasAttr>();
68 
69   // Typedef types have declarations, and therefore may have attributes.
70   if (const TypedefType *TTy = dyn_cast<TypedefType>(QTy)) {
71     if (TTy->getDecl()->hasAttr<MayAliasAttr>())
72       return true;
73     // Also, their underlying types may have relevant attributes.
74     return TypeHasMayAlias(TTy->desugar());
75   }
76 
77   return false;
78 }
79 
80 llvm::MDNode *
81 CodeGenTBAA::getTBAAInfo(QualType QTy) {
82   // At -O0 TBAA is not emitted for regular types.
83   if (CodeGenOpts.OptimizationLevel == 0 || CodeGenOpts.RelaxedAliasing)
84     return NULL;
85 
86   // If the type has the may_alias attribute (even on a typedef), it is
87   // effectively in the general char alias class.
88   if (TypeHasMayAlias(QTy))
89     return getChar();
90 
91   const Type *Ty = Context.getCanonicalType(QTy).getTypePtr();
92 
93   if (llvm::MDNode *N = MetadataCache[Ty])
94     return N;
95 
96   // Handle builtin types.
97   if (const BuiltinType *BTy = dyn_cast<BuiltinType>(Ty)) {
98     switch (BTy->getKind()) {
99     // Character types are special and can alias anything.
100     // In C++, this technically only includes "char" and "unsigned char",
101     // and not "signed char". In C, it includes all three. For now,
102     // the risk of exploiting this detail in C++ seems likely to outweigh
103     // the benefit.
104     case BuiltinType::Char_U:
105     case BuiltinType::Char_S:
106     case BuiltinType::UChar:
107     case BuiltinType::SChar:
108       return getChar();
109 
110     // Unsigned types can alias their corresponding signed types.
111     case BuiltinType::UShort:
112       return getTBAAInfo(Context.ShortTy);
113     case BuiltinType::UInt:
114       return getTBAAInfo(Context.IntTy);
115     case BuiltinType::ULong:
116       return getTBAAInfo(Context.LongTy);
117     case BuiltinType::ULongLong:
118       return getTBAAInfo(Context.LongLongTy);
119     case BuiltinType::UInt128:
120       return getTBAAInfo(Context.Int128Ty);
121 
122     // Treat all other builtin types as distinct types. This includes
123     // treating wchar_t, char16_t, and char32_t as distinct from their
124     // "underlying types".
125     default:
126       return MetadataCache[Ty] =
127         MDHelper.createTBAANode(BTy->getName(Features), getChar());
128     }
129   }
130 
131   // Handle pointers.
132   // TODO: Implement C++'s type "similarity" and consider dis-"similar"
133   // pointers distinct.
134   if (Ty->isPointerType())
135     return MetadataCache[Ty] = MDHelper.createTBAANode("any pointer",
136                                                        getChar());
137 
138   // Enum types are distinct types. In C++ they have "underlying types",
139   // however they aren't related for TBAA.
140   if (const EnumType *ETy = dyn_cast<EnumType>(Ty)) {
141     // In C mode, two anonymous enums are compatible iff their members
142     // are the same -- see C99 6.2.7p1. For now, be conservative. We could
143     // theoretically implement this by combining information about all the
144     // members into a single identifying MDNode.
145     if (!Features.CPlusPlus &&
146         ETy->getDecl()->getTypedefNameForAnonDecl())
147       return MetadataCache[Ty] = getChar();
148 
149     // In C++ mode, types have linkage, so we can rely on the ODR and
150     // on their mangled names, if they're external.
151     // TODO: Is there a way to get a program-wide unique name for a
152     // decl with local linkage or no linkage?
153     if (Features.CPlusPlus &&
154         ETy->getDecl()->getLinkage() != ExternalLinkage)
155       return MetadataCache[Ty] = getChar();
156 
157     // TODO: This is using the RTTI name. Is there a better way to get
158     // a unique string for a type?
159     SmallString<256> OutName;
160     llvm::raw_svector_ostream Out(OutName);
161     MContext.mangleCXXRTTIName(QualType(ETy, 0), Out);
162     Out.flush();
163     return MetadataCache[Ty] = MDHelper.createTBAANode(OutName, getChar());
164   }
165 
166   // For now, handle any other kind of type conservatively.
167   return MetadataCache[Ty] = getChar();
168 }
169 
170 llvm::MDNode *CodeGenTBAA::getTBAAInfoForVTablePtr() {
171   return MDHelper.createTBAANode("vtable pointer", getRoot());
172 }
173 
174 bool
175 CodeGenTBAA::CollectFields(uint64_t BaseOffset,
176                            QualType QTy,
177                            SmallVectorImpl<llvm::MDBuilder::TBAAStructField> &
178                              Fields,
179                            bool MayAlias) {
180   /* Things not handled yet include: C++ base classes, bitfields, */
181 
182   if (const RecordType *TTy = QTy->getAs<RecordType>()) {
183     const RecordDecl *RD = TTy->getDecl()->getDefinition();
184     if (RD->hasFlexibleArrayMember())
185       return false;
186 
187     // TODO: Handle C++ base classes.
188     if (const CXXRecordDecl *Decl = dyn_cast<CXXRecordDecl>(RD))
189       if (Decl->bases_begin() != Decl->bases_end())
190         return false;
191 
192     const ASTRecordLayout &Layout = Context.getASTRecordLayout(RD);
193 
194     unsigned idx = 0;
195     for (RecordDecl::field_iterator i = RD->field_begin(),
196          e = RD->field_end(); i != e; ++i, ++idx) {
197       uint64_t Offset = BaseOffset +
198                         Layout.getFieldOffset(idx) / Context.getCharWidth();
199       QualType FieldQTy = i->getType();
200       if (!CollectFields(Offset, FieldQTy, Fields,
201                          MayAlias || TypeHasMayAlias(FieldQTy)))
202         return false;
203     }
204     return true;
205   }
206 
207   /* Otherwise, treat whatever it is as a field. */
208   uint64_t Offset = BaseOffset;
209   uint64_t Size = Context.getTypeSizeInChars(QTy).getQuantity();
210   llvm::MDNode *TBAAInfo = MayAlias ? getChar() : getTBAAInfo(QTy);
211   Fields.push_back(llvm::MDBuilder::TBAAStructField(Offset, Size, TBAAInfo));
212   return true;
213 }
214 
215 llvm::MDNode *
216 CodeGenTBAA::getTBAAStructInfo(QualType QTy) {
217   const Type *Ty = Context.getCanonicalType(QTy).getTypePtr();
218 
219   if (llvm::MDNode *N = StructMetadataCache[Ty])
220     return N;
221 
222   SmallVector<llvm::MDBuilder::TBAAStructField, 4> Fields;
223   if (CollectFields(0, QTy, Fields, TypeHasMayAlias(QTy)))
224     return MDHelper.createTBAAStructNode(Fields);
225 
226   // For now, handle any other kind of type conservatively.
227   return StructMetadataCache[Ty] = NULL;
228 }
229 
230 /// Check if the given type can be handled by path-aware TBAA.
231 static bool isTBAAPathStruct(QualType QTy) {
232   if (const RecordType *TTy = QTy->getAs<RecordType>()) {
233     const RecordDecl *RD = TTy->getDecl()->getDefinition();
234     // RD can be struct, union, class, interface or enum.
235     // For now, we only handle struct.
236     if (RD->isStruct() && !RD->hasFlexibleArrayMember())
237       return true;
238   }
239   return false;
240 }
241 
242 llvm::MDNode *
243 CodeGenTBAA::getTBAAStructTypeInfo(QualType QTy) {
244   const Type *Ty = Context.getCanonicalType(QTy).getTypePtr();
245   assert(isTBAAPathStruct(QTy));
246 
247   if (llvm::MDNode *N = StructTypeMetadataCache[Ty])
248     return N;
249 
250   if (const RecordType *TTy = QTy->getAs<RecordType>()) {
251     const RecordDecl *RD = TTy->getDecl()->getDefinition();
252 
253     const ASTRecordLayout &Layout = Context.getASTRecordLayout(RD);
254     SmallVector <std::pair<uint64_t, llvm::MDNode*>, 4> Fields;
255     // To reduce the size of MDNode for a given struct type, we only output
256     // once for all the fields with the same scalar types.
257     // Offsets for scalar fields in the type DAG are not used.
258     llvm::SmallSet <llvm::MDNode*, 4> ScalarFieldTypes;
259     unsigned idx = 0;
260     for (RecordDecl::field_iterator i = RD->field_begin(),
261          e = RD->field_end(); i != e; ++i, ++idx) {
262       QualType FieldQTy = i->getType();
263       llvm::MDNode *FieldNode;
264       if (isTBAAPathStruct(FieldQTy))
265         FieldNode = getTBAAStructTypeInfo(FieldQTy);
266       else {
267         FieldNode = getTBAAInfo(FieldQTy);
268         // Ignore this field if the type already exists.
269         if (ScalarFieldTypes.count(FieldNode))
270           continue;
271         ScalarFieldTypes.insert(FieldNode);
272        }
273       if (!FieldNode)
274         return StructTypeMetadataCache[Ty] = NULL;
275       Fields.push_back(std::make_pair(
276           Layout.getFieldOffset(idx) / Context.getCharWidth(), FieldNode));
277     }
278 
279     // TODO: This is using the RTTI name. Is there a better way to get
280     // a unique string for a type?
281     SmallString<256> OutName;
282     llvm::raw_svector_ostream Out(OutName);
283     MContext.mangleCXXRTTIName(QualType(Ty, 0), Out);
284     Out.flush();
285     // Create the struct type node with a vector of pairs (offset, type).
286     return StructTypeMetadataCache[Ty] =
287       MDHelper.createTBAAStructTypeNode(OutName, Fields);
288   }
289 
290   return StructMetadataCache[Ty] = NULL;
291 }
292 
293 llvm::MDNode *
294 CodeGenTBAA::getTBAAStructTagInfo(QualType BaseQTy, llvm::MDNode *AccessNode,
295                                   uint64_t Offset) {
296   if (!CodeGenOpts.StructPathTBAA)
297     return AccessNode;
298 
299   const Type *BTy = Context.getCanonicalType(BaseQTy).getTypePtr();
300   TBAAPathTag PathTag = TBAAPathTag(BTy, AccessNode, Offset);
301   if (llvm::MDNode *N = StructTagMetadataCache[PathTag])
302     return N;
303 
304   llvm::MDNode *BNode = 0;
305   if (isTBAAPathStruct(BaseQTy))
306     BNode  = getTBAAStructTypeInfo(BaseQTy);
307   if (!BNode)
308     return StructTagMetadataCache[PathTag] = AccessNode;
309 
310   return StructTagMetadataCache[PathTag] =
311     MDHelper.createTBAAStructTagNode(BNode, AccessNode, Offset);
312 }
313