CodeGenTBAA.cpp revision 162e1c1b487352434552147967c3dd296ebee2f7 
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/Mangle.h"
21#include "llvm/LLVMContext.h"
22#include "llvm/Metadata.h"
23#include "llvm/Constants.h"
24#include "llvm/Type.h"
25using namespace clang;
26using namespace CodeGen;
27
28CodeGenTBAA::CodeGenTBAA(ASTContext &Ctx, llvm::LLVMContext& VMContext,
29                         const LangOptions &Features, MangleContext &MContext)
30  : Context(Ctx), VMContext(VMContext), Features(Features), MContext(MContext),
31    Root(0), Char(0) {
32}
33
34CodeGenTBAA::~CodeGenTBAA() {
35}
36
37llvm::MDNode *CodeGenTBAA::getRoot() {
38  // Define the root of the tree. This identifies the tree, so that
39  // if our LLVM IR is linked with LLVM IR from a different front-end
40  // (or a different version of this front-end), their TBAA trees will
41  // remain distinct, and the optimizer will treat them conservatively.
42  if (!Root)
43    Root = getTBAAInfoForNamedType("Simple C/C++ TBAA", 0);
44
45  return Root;
46}
47
48llvm::MDNode *CodeGenTBAA::getChar() {
49  // Define the root of the tree for user-accessible memory. C and C++
50  // give special powers to char and certain similar types. However,
51  // these special powers only cover user-accessible memory, and doesn't
52  // include things like vtables.
53  if (!Char)
54    Char = getTBAAInfoForNamedType("omnipotent char", getRoot());
55
56  return Char;
57}
58
59/// getTBAAInfoForNamedType - Create a TBAA tree node with the given string
60/// as its identifier, and the given Parent node as its tree parent.
61llvm::MDNode *CodeGenTBAA::getTBAAInfoForNamedType(llvm::StringRef NameStr,
62                                                   llvm::MDNode *Parent,
63                                                   bool Readonly) {
64  // Currently there is only one flag defined - the readonly flag.
65  llvm::Value *Flags = 0;
66  if (Readonly)
67    Flags = llvm::ConstantInt::get(llvm::Type::getInt64Ty(VMContext), true);
68
69  // Set up the mdnode operand list.
70  llvm::Value *Ops[] = {
71    llvm::MDString::get(VMContext, NameStr),
72    Parent,
73    Flags
74  };
75
76  // Create the mdnode.
77  return llvm::MDNode::get(VMContext, Ops, llvm::array_lengthof(Ops) - !Flags);
78}
79
80static bool TypeHasMayAlias(QualType QTy) {
81  // Tagged types have declarations, and therefore may have attributes.
82  if (const TagType *TTy = dyn_cast<TagType>(QTy))
83    return TTy->getDecl()->hasAttr<MayAliasAttr>();
84
85  // Typedef types have declarations, and therefore may have attributes.
86  if (const TypedefType *TTy = dyn_cast<TypedefType>(QTy)) {
87    if (TTy->getDecl()->hasAttr<MayAliasAttr>())
88      return true;
89    // Also, their underlying types may have relevant attributes.
90    return TypeHasMayAlias(TTy->desugar());
91  }
92
93  return false;
94}
95
96llvm::MDNode *
97CodeGenTBAA::getTBAAInfo(QualType QTy) {
98  // If the type has the may_alias attribute (even on a typedef), it is
99  // effectively in the general char alias class.
100  if (TypeHasMayAlias(QTy))
101    return getChar();
102
103  const Type *Ty = Context.getCanonicalType(QTy).getTypePtr();
104
105  if (llvm::MDNode *N = MetadataCache[Ty])
106    return N;
107
108  // Handle builtin types.
109  if (const BuiltinType *BTy = dyn_cast<BuiltinType>(Ty)) {
110    switch (BTy->getKind()) {
111    // Character types are special and can alias anything.
112    // In C++, this technically only includes "char" and "unsigned char",
113    // and not "signed char". In C, it includes all three. For now,
114    // the risk of exploiting this detail in C++ seems likely to outweigh
115    // the benefit.
116    case BuiltinType::Char_U:
117    case BuiltinType::Char_S:
118    case BuiltinType::UChar:
119    case BuiltinType::SChar:
120      return getChar();
121
122    // Unsigned types can alias their corresponding signed types.
123    case BuiltinType::UShort:
124      return getTBAAInfo(Context.ShortTy);
125    case BuiltinType::UInt:
126      return getTBAAInfo(Context.IntTy);
127    case BuiltinType::ULong:
128      return getTBAAInfo(Context.LongTy);
129    case BuiltinType::ULongLong:
130      return getTBAAInfo(Context.LongLongTy);
131    case BuiltinType::UInt128:
132      return getTBAAInfo(Context.Int128Ty);
133
134    // Treat all other builtin types as distinct types. This includes
135    // treating wchar_t, char16_t, and char32_t as distinct from their
136    // "underlying types".
137    default:
138      return MetadataCache[Ty] =
139               getTBAAInfoForNamedType(BTy->getName(Features), getChar());
140    }
141  }
142
143  // Handle pointers.
144  // TODO: Implement C++'s type "similarity" and consider dis-"similar"
145  // pointers distinct.
146  if (Ty->isPointerType())
147    return MetadataCache[Ty] = getTBAAInfoForNamedType("any pointer",
148                                                       getChar());
149
150  // Enum types are distinct types. In C++ they have "underlying types",
151  // however they aren't related for TBAA.
152  if (const EnumType *ETy = dyn_cast<EnumType>(Ty)) {
153    // In C mode, two anonymous enums are compatible iff their members
154    // are the same -- see C99 6.2.7p1. For now, be conservative. We could
155    // theoretically implement this by combining information about all the
156    // members into a single identifying MDNode.
157    if (!Features.CPlusPlus &&
158        ETy->getDecl()->getTypedefNameForAnonDecl())
159      return MetadataCache[Ty] = getChar();
160
161    // In C++ mode, types have linkage, so we can rely on the ODR and
162    // on their mangled names, if they're external.
163    // TODO: Is there a way to get a program-wide unique name for a
164    // decl with local linkage or no linkage?
165    if (Features.CPlusPlus &&
166        ETy->getDecl()->getLinkage() != ExternalLinkage)
167      return MetadataCache[Ty] = getChar();
168
169    // TODO: This is using the RTTI name. Is there a better way to get
170    // a unique string for a type?
171    llvm::SmallString<256> OutName;
172    llvm::raw_svector_ostream Out(OutName);
173    MContext.mangleCXXRTTIName(QualType(ETy, 0), Out);
174    Out.flush();
175    return MetadataCache[Ty] = getTBAAInfoForNamedType(OutName, getChar());
176  }
177
178  // For now, handle any other kind of type conservatively.
179  return MetadataCache[Ty] = getChar();
180}
181