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/RecordLayout.h"
21 #include "clang/AST/Mangle.h"
22 #include "clang/Frontend/CodeGenOptions.h"
23 #include "llvm/LLVMContext.h"
24 #include "llvm/Metadata.h"
25 #include "llvm/Constants.h"
26 #include "llvm/Type.h"
27 using namespace clang;
28 using namespace CodeGen;
29 
30 CodeGenTBAA::CodeGenTBAA(ASTContext &Ctx, llvm::LLVMContext& VMContext,
31                          const CodeGenOptions &CGO,
32                          const LangOptions &Features, MangleContext &MContext)
33   : Context(Ctx), CodeGenOpts(CGO), Features(Features), MContext(MContext),
34     MDHelper(VMContext), Root(0), Char(0) {
35 }
36 
37 CodeGenTBAA::~CodeGenTBAA() {
38 }
39 
40 llvm::MDNode *CodeGenTBAA::getRoot() {
41   // Define the root of the tree. This identifies the tree, so that
42   // if our LLVM IR is linked with LLVM IR from a different front-end
43   // (or a different version of this front-end), their TBAA trees will
44   // remain distinct, and the optimizer will treat them conservatively.
45   if (!Root)
46     Root = MDHelper.createTBAARoot("Simple C/C++ TBAA");
47 
48   return Root;
49 }
50 
51 llvm::MDNode *CodeGenTBAA::getChar() {
52   // Define the root of the tree for user-accessible memory. C and C++
53   // give special powers to char and certain similar types. However,
54   // these special powers only cover user-accessible memory, and doesn't
55   // include things like vtables.
56   if (!Char)
57     Char = MDHelper.createTBAANode("omnipotent char", getRoot());
58 
59   return Char;
60 }
61 
62 static bool TypeHasMayAlias(QualType QTy) {
63   // Tagged types have declarations, and therefore may have attributes.
64   if (const TagType *TTy = dyn_cast<TagType>(QTy))
65     return TTy->getDecl()->hasAttr<MayAliasAttr>();
66 
67   // Typedef types have declarations, and therefore may have attributes.
68   if (const TypedefType *TTy = dyn_cast<TypedefType>(QTy)) {
69     if (TTy->getDecl()->hasAttr<MayAliasAttr>())
70       return true;
71     // Also, their underlying types may have relevant attributes.
72     return TypeHasMayAlias(TTy->desugar());
73   }
74 
75   return false;
76 }
77 
78 llvm::MDNode *
79 CodeGenTBAA::getTBAAInfo(QualType QTy) {
80   // At -O0 TBAA is not emitted for regular types.
81   if (CodeGenOpts.OptimizationLevel == 0 || CodeGenOpts.RelaxedAliasing)
82     return NULL;
83 
84   // If the type has the may_alias attribute (even on a typedef), it is
85   // effectively in the general char alias class.
86   if (TypeHasMayAlias(QTy))
87     return getChar();
88 
89   const Type *Ty = Context.getCanonicalType(QTy).getTypePtr();
90 
91   if (llvm::MDNode *N = MetadataCache[Ty])
92     return N;
93 
94   // Handle builtin types.
95   if (const BuiltinType *BTy = dyn_cast<BuiltinType>(Ty)) {
96     switch (BTy->getKind()) {
97     // Character types are special and can alias anything.
98     // In C++, this technically only includes "char" and "unsigned char",
99     // and not "signed char". In C, it includes all three. For now,
100     // the risk of exploiting this detail in C++ seems likely to outweigh
101     // the benefit.
102     case BuiltinType::Char_U:
103     case BuiltinType::Char_S:
104     case BuiltinType::UChar:
105     case BuiltinType::SChar:
106       return getChar();
107 
108     // Unsigned types can alias their corresponding signed types.
109     case BuiltinType::UShort:
110       return getTBAAInfo(Context.ShortTy);
111     case BuiltinType::UInt:
112       return getTBAAInfo(Context.IntTy);
113     case BuiltinType::ULong:
114       return getTBAAInfo(Context.LongTy);
115     case BuiltinType::ULongLong:
116       return getTBAAInfo(Context.LongLongTy);
117     case BuiltinType::UInt128:
118       return getTBAAInfo(Context.Int128Ty);
119 
120     // Treat all other builtin types as distinct types. This includes
121     // treating wchar_t, char16_t, and char32_t as distinct from their
122     // "underlying types".
123     default:
124       return MetadataCache[Ty] =
125         MDHelper.createTBAANode(BTy->getName(Features), getChar());
126     }
127   }
128 
129   // Handle pointers.
130   // TODO: Implement C++'s type "similarity" and consider dis-"similar"
131   // pointers distinct.
132   if (Ty->isPointerType())
133     return MetadataCache[Ty] = MDHelper.createTBAANode("any pointer",
134                                                        getChar());
135 
136   // Enum types are distinct types. In C++ they have "underlying types",
137   // however they aren't related for TBAA.
138   if (const EnumType *ETy = dyn_cast<EnumType>(Ty)) {
139     // In C mode, two anonymous enums are compatible iff their members
140     // are the same -- see C99 6.2.7p1. For now, be conservative. We could
141     // theoretically implement this by combining information about all the
142     // members into a single identifying MDNode.
143     if (!Features.CPlusPlus &&
144         ETy->getDecl()->getTypedefNameForAnonDecl())
145       return MetadataCache[Ty] = getChar();
146 
147     // In C++ mode, types have linkage, so we can rely on the ODR and
148     // on their mangled names, if they're external.
149     // TODO: Is there a way to get a program-wide unique name for a
150     // decl with local linkage or no linkage?
151     if (Features.CPlusPlus &&
152         ETy->getDecl()->getLinkage() != ExternalLinkage)
153       return MetadataCache[Ty] = getChar();
154 
155     // TODO: This is using the RTTI name. Is there a better way to get
156     // a unique string for a type?
157     SmallString<256> OutName;
158     llvm::raw_svector_ostream Out(OutName);
159     MContext.mangleCXXRTTIName(QualType(ETy, 0), Out);
160     Out.flush();
161     return MetadataCache[Ty] = MDHelper.createTBAANode(OutName, getChar());
162   }
163 
164   // For now, handle any other kind of type conservatively.
165   return MetadataCache[Ty] = getChar();
166 }
167 
168 llvm::MDNode *CodeGenTBAA::getTBAAInfoForVTablePtr() {
169   return MDHelper.createTBAANode("vtable pointer", getRoot());
170 }
171 
172 bool
173 CodeGenTBAA::CollectFields(uint64_t BaseOffset,
174                            QualType QTy,
175                            SmallVectorImpl<llvm::MDBuilder::TBAAStructField> &
176                              Fields,
177                            bool MayAlias) {
178   /* Things not handled yet include: C++ base classes, bitfields, */
179 
180   if (const RecordType *TTy = QTy->getAs<RecordType>()) {
181     const RecordDecl *RD = TTy->getDecl()->getDefinition();
182     if (RD->hasFlexibleArrayMember())
183       return false;
184 
185     // TODO: Handle C++ base classes.
186     if (const CXXRecordDecl *Decl = dyn_cast<CXXRecordDecl>(RD))
187       if (Decl->bases_begin() != Decl->bases_end())
188         return false;
189 
190     const ASTRecordLayout &Layout = Context.getASTRecordLayout(RD);
191 
192     unsigned idx = 0;
193     for (RecordDecl::field_iterator i = RD->field_begin(),
194          e = RD->field_end(); i != e; ++i, ++idx) {
195       uint64_t Offset = BaseOffset +
196                         Layout.getFieldOffset(idx) / Context.getCharWidth();
197       QualType FieldQTy = i->getType();
198       if (!CollectFields(Offset, FieldQTy, Fields,
199                          MayAlias || TypeHasMayAlias(FieldQTy)))
200         return false;
201     }
202     return true;
203   }
204 
205   /* Otherwise, treat whatever it is as a field. */
206   uint64_t Offset = BaseOffset;
207   uint64_t Size = Context.getTypeSizeInChars(QTy).getQuantity();
208   llvm::MDNode *TBAAInfo = MayAlias ? getChar() : getTBAAInfo(QTy);
209   Fields.push_back(llvm::MDBuilder::TBAAStructField(Offset, Size, TBAAInfo));
210   return true;
211 }
212 
213 llvm::MDNode *
214 CodeGenTBAA::getTBAAStructInfo(QualType QTy) {
215   const Type *Ty = Context.getCanonicalType(QTy).getTypePtr();
216 
217   if (llvm::MDNode *N = StructMetadataCache[Ty])
218     return N;
219 
220   SmallVector<llvm::MDBuilder::TBAAStructField, 4> Fields;
221   if (CollectFields(0, QTy, Fields, TypeHasMayAlias(QTy)))
222     return MDHelper.createTBAAStructNode(Fields);
223 
224   // For now, handle any other kind of type conservatively.
225   return StructMetadataCache[Ty] = NULL;
226 }
227