CodeGenTBAA.cpp revision 224d75972a836b06e2ca708d1eafdac6f762c487 
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 "Mangle.h"
20#include "clang/AST/ASTContext.h"
21#include "llvm/LLVMContext.h"
22#include "llvm/Metadata.h"
23using namespace clang;
24using namespace CodeGen;
25
26CodeGenTBAA::CodeGenTBAA(ASTContext &Ctx, llvm::LLVMContext& VMContext,
27                         const LangOptions &Features, MangleContext &MContext)
28  : Context(Ctx), VMContext(VMContext), Features(Features), MContext(MContext),
29    Root(0), Char(0) {
30}
31
32CodeGenTBAA::~CodeGenTBAA() {
33}
34
35llvm::MDNode *CodeGenTBAA::getRoot() {
36  // Define the root of the tree. This identifies the tree, so that
37  // if our LLVM IR is linked with LLVM IR from a different front-end
38  // (or a different version of this front-end), their TBAA trees will
39  // remain distinct, and the optimizer will treat them conservatively.
40  if (!Root)
41    Root = getTBAAInfoForNamedType("Simple C/C++ TBAA", 0);
42
43  return Root;
44}
45
46llvm::MDNode *CodeGenTBAA::getChar() {
47  // Define the root of the tree for user-accessible memory. C and C++
48  // give special powers to char and certain similar types. However,
49  // these special powers only cover user-accessible memory, and doesn't
50  // include things like vtables.
51  if (!Char)
52    Char = getTBAAInfoForNamedType("omnipotent char", getRoot());
53
54  return Char;
55}
56
57/// getTBAAInfoForNamedType - Create a TBAA tree node with the given string
58/// as its identifier, and the given Parent node as its tree parent.
59llvm::MDNode *CodeGenTBAA::getTBAAInfoForNamedType(llvm::StringRef NameStr,
60                                                   llvm::MDNode *Parent) {
61  llvm::Value *Ops[] = {
62    llvm::MDString::get(VMContext, NameStr),
63    Parent
64  };
65
66  return llvm::MDNode::get(VMContext, Ops, llvm::array_lengthof(Ops));
67}
68
69llvm::MDNode *
70CodeGenTBAA::getTBAAInfo(QualType QTy) {
71  Type *Ty = Context.getCanonicalType(QTy).getTypePtr();
72
73  if (llvm::MDNode *N = MetadataCache[Ty])
74    return N;
75
76  // Handle builtin types.
77  if (const BuiltinType *BTy = dyn_cast<BuiltinType>(Ty)) {
78    switch (BTy->getKind()) {
79    // Character types are special and can alias anything.
80    // In C++, this technically only includes "char" and "unsigned char",
81    // and not "signed char". In C, it includes all three. For now,
82    // the risk of exploiting this detail in C++ seems likely to outweigh
83    // the benefit.
84    case BuiltinType::Char_U:
85    case BuiltinType::Char_S:
86    case BuiltinType::UChar:
87    case BuiltinType::SChar:
88      return getChar();
89
90    // Unsigned types can alias their corresponding signed types.
91    case BuiltinType::UShort:
92      return getTBAAInfo(Context.ShortTy);
93    case BuiltinType::UInt:
94      return getTBAAInfo(Context.IntTy);
95    case BuiltinType::ULong:
96      return getTBAAInfo(Context.LongTy);
97    case BuiltinType::ULongLong:
98      return getTBAAInfo(Context.LongLongTy);
99    case BuiltinType::UInt128:
100      return getTBAAInfo(Context.Int128Ty);
101
102    // Treat all other builtin types as distinct types. This includes
103    // treating wchar_t, char16_t, and char32_t as distinct from their
104    // "underlying types".
105    default:
106      return MetadataCache[Ty] =
107               getTBAAInfoForNamedType(BTy->getName(Features), getChar());
108    }
109  }
110
111  // Handle pointers.
112  // TODO: Implement C++'s type "similarity" and consider dis-"similar"
113  // pointers distinct.
114  if (Ty->isPointerType())
115    return MetadataCache[Ty] = getTBAAInfoForNamedType("any pointer",
116                                                       getChar());
117
118  // Enum types are distinct types. In C++ they have "underlying types",
119  // however they aren't related for TBAA.
120  if (const EnumType *ETy = dyn_cast<EnumType>(Ty)) {
121    // In C mode, two anonymous enums are compatible iff their members
122    // are the same -- see C99 6.2.7p1. For now, be conservative. We could
123    // theoretically implement this by combining information about all the
124    // members into a single identifying MDNode.
125    if (!Features.CPlusPlus &&
126        ETy->getDecl()->getTypedefForAnonDecl())
127      return MetadataCache[Ty] = getChar();
128
129    // In C++ mode, types have linkage, so we can rely on the ODR and
130    // on their mangled names, if they're external.
131    // TODO: Is there a way to get a program-wide unique name for a
132    // decl with local linkage or no linkage?
133    if (Features.CPlusPlus &&
134        ETy->getDecl()->getLinkage() != ExternalLinkage)
135      return MetadataCache[Ty] = getChar();
136
137    // TODO: This is using the RTTI name. Is there a better way to get
138    // a unique string for a type?
139    llvm::SmallString<256> OutName;
140    MContext.mangleCXXRTTIName(QualType(ETy, 0), OutName);
141    return MetadataCache[Ty] = getTBAAInfoForNamedType(OutName, getChar());
142  }
143
144  // For now, handle any other kind of type conservatively.
145  return MetadataCache[Ty] = getChar();
146}
147