DataFlowSanitizer.cpp revision 263508 
1//===-- DataFlowSanitizer.cpp - dynamic data flow analysis ----------------===//
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/// \file
10/// This file is a part of DataFlowSanitizer, a generalised dynamic data flow
11/// analysis.
12///
13/// Unlike other Sanitizer tools, this tool is not designed to detect a specific
14/// class of bugs on its own.  Instead, it provides a generic dynamic data flow
15/// analysis framework to be used by clients to help detect application-specific
16/// issues within their own code.
17///
18/// The analysis is based on automatic propagation of data flow labels (also
19/// known as taint labels) through a program as it performs computation.  Each
20/// byte of application memory is backed by two bytes of shadow memory which
21/// hold the label.  On Linux/x86_64, memory is laid out as follows:
22///
23/// +--------------------+ 0x800000000000 (top of memory)
24/// | application memory |
25/// +--------------------+ 0x700000008000 (kAppAddr)
26/// |                    |
27/// |       unused       |
28/// |                    |
29/// +--------------------+ 0x200200000000 (kUnusedAddr)
30/// |    union table     |
31/// +--------------------+ 0x200000000000 (kUnionTableAddr)
32/// |   shadow memory    |
33/// +--------------------+ 0x000000010000 (kShadowAddr)
34/// | reserved by kernel |
35/// +--------------------+ 0x000000000000
36///
37/// To derive a shadow memory address from an application memory address,
38/// bits 44-46 are cleared to bring the address into the range
39/// [0x000000008000,0x100000000000).  Then the address is shifted left by 1 to
40/// account for the double byte representation of shadow labels and move the
41/// address into the shadow memory range.  See the function
42/// DataFlowSanitizer::getShadowAddress below.
43///
44/// For more information, please refer to the design document:
45/// http://clang.llvm.org/docs/DataFlowSanitizerDesign.html
46
47#include "llvm/Transforms/Instrumentation.h"
48#include "llvm/ADT/DenseMap.h"
49#include "llvm/ADT/DenseSet.h"
50#include "llvm/ADT/DepthFirstIterator.h"
51#include "llvm/ADT/StringExtras.h"
52#include "llvm/Analysis/ValueTracking.h"
53#include "llvm/IR/InlineAsm.h"
54#include "llvm/IR/IRBuilder.h"
55#include "llvm/IR/LLVMContext.h"
56#include "llvm/IR/MDBuilder.h"
57#include "llvm/IR/Type.h"
58#include "llvm/IR/Value.h"
59#include "llvm/InstVisitor.h"
60#include "llvm/Pass.h"
61#include "llvm/Support/CommandLine.h"
62#include "llvm/Transforms/Utils/BasicBlockUtils.h"
63#include "llvm/Transforms/Utils/Local.h"
64#include "llvm/Transforms/Utils/SpecialCaseList.h"
65#include <iterator>
66
67using namespace llvm;
68
69// The -dfsan-preserve-alignment flag controls whether this pass assumes that
70// alignment requirements provided by the input IR are correct.  For example,
71// if the input IR contains a load with alignment 8, this flag will cause
72// the shadow load to have alignment 16.  This flag is disabled by default as
73// we have unfortunately encountered too much code (including Clang itself;
74// see PR14291) which performs misaligned access.
75static cl::opt<bool> ClPreserveAlignment(
76    "dfsan-preserve-alignment",
77    cl::desc("respect alignment requirements provided by input IR"), cl::Hidden,
78    cl::init(false));
79
80// The ABI list file controls how shadow parameters are passed.  The pass treats
81// every function labelled "uninstrumented" in the ABI list file as conforming
82// to the "native" (i.e. unsanitized) ABI.  Unless the ABI list contains
83// additional annotations for those functions, a call to one of those functions
84// will produce a warning message, as the labelling behaviour of the function is
85// unknown.  The other supported annotations are "functional" and "discard",
86// which are described below under DataFlowSanitizer::WrapperKind.
87static cl::opt<std::string> ClABIListFile(
88    "dfsan-abilist",
89    cl::desc("File listing native ABI functions and how the pass treats them"),
90    cl::Hidden);
91
92// Controls whether the pass uses IA_Args or IA_TLS as the ABI for instrumented
93// functions (see DataFlowSanitizer::InstrumentedABI below).
94static cl::opt<bool> ClArgsABI(
95    "dfsan-args-abi",
96    cl::desc("Use the argument ABI rather than the TLS ABI"),
97    cl::Hidden);
98
99static cl::opt<bool> ClDebugNonzeroLabels(
100    "dfsan-debug-nonzero-labels",
101    cl::desc("Insert calls to __dfsan_nonzero_label on observing a parameter, "
102             "load or return with a nonzero label"),
103    cl::Hidden);
104
105namespace {
106
107class DataFlowSanitizer : public ModulePass {
108  friend struct DFSanFunction;
109  friend class DFSanVisitor;
110
111  enum {
112    ShadowWidth = 16
113  };
114
115  /// Which ABI should be used for instrumented functions?
116  enum InstrumentedABI {
117    /// Argument and return value labels are passed through additional
118    /// arguments and by modifying the return type.
119    IA_Args,
120
121    /// Argument and return value labels are passed through TLS variables
122    /// __dfsan_arg_tls and __dfsan_retval_tls.
123    IA_TLS
124  };
125
126  /// How should calls to uninstrumented functions be handled?
127  enum WrapperKind {
128    /// This function is present in an uninstrumented form but we don't know
129    /// how it should be handled.  Print a warning and call the function anyway.
130    /// Don't label the return value.
131    WK_Warning,
132
133    /// This function does not write to (user-accessible) memory, and its return
134    /// value is unlabelled.
135    WK_Discard,
136
137    /// This function does not write to (user-accessible) memory, and the label
138    /// of its return value is the union of the label of its arguments.
139    WK_Functional,
140
141    /// Instead of calling the function, a custom wrapper __dfsw_F is called,
142    /// where F is the name of the function.  This function may wrap the
143    /// original function or provide its own implementation.  This is similar to
144    /// the IA_Args ABI, except that IA_Args uses a struct return type to
145    /// pass the return value shadow in a register, while WK_Custom uses an
146    /// extra pointer argument to return the shadow.  This allows the wrapped
147    /// form of the function type to be expressed in C.
148    WK_Custom
149  };
150
151  DataLayout *DL;
152  Module *Mod;
153  LLVMContext *Ctx;
154  IntegerType *ShadowTy;
155  PointerType *ShadowPtrTy;
156  IntegerType *IntptrTy;
157  ConstantInt *ZeroShadow;
158  ConstantInt *ShadowPtrMask;
159  ConstantInt *ShadowPtrMul;
160  Constant *ArgTLS;
161  Constant *RetvalTLS;
162  void *(*GetArgTLSPtr)();
163  void *(*GetRetvalTLSPtr)();
164  Constant *GetArgTLS;
165  Constant *GetRetvalTLS;
166  FunctionType *DFSanUnionFnTy;
167  FunctionType *DFSanUnionLoadFnTy;
168  FunctionType *DFSanUnimplementedFnTy;
169  FunctionType *DFSanSetLabelFnTy;
170  FunctionType *DFSanNonzeroLabelFnTy;
171  Constant *DFSanUnionFn;
172  Constant *DFSanUnionLoadFn;
173  Constant *DFSanUnimplementedFn;
174  Constant *DFSanSetLabelFn;
175  Constant *DFSanNonzeroLabelFn;
176  MDNode *ColdCallWeights;
177  OwningPtr<SpecialCaseList> ABIList;
178  DenseMap<Value *, Function *> UnwrappedFnMap;
179  AttributeSet ReadOnlyNoneAttrs;
180
181  Value *getShadowAddress(Value *Addr, Instruction *Pos);
182  Value *combineShadows(Value *V1, Value *V2, Instruction *Pos);
183  bool isInstrumented(const Function *F);
184  bool isInstrumented(const GlobalAlias *GA);
185  FunctionType *getArgsFunctionType(FunctionType *T);
186  FunctionType *getTrampolineFunctionType(FunctionType *T);
187  FunctionType *getCustomFunctionType(FunctionType *T);
188  InstrumentedABI getInstrumentedABI();
189  WrapperKind getWrapperKind(Function *F);
190  void addGlobalNamePrefix(GlobalValue *GV);
191  Function *buildWrapperFunction(Function *F, StringRef NewFName,
192                                 GlobalValue::LinkageTypes NewFLink,
193                                 FunctionType *NewFT);
194  Constant *getOrBuildTrampolineFunction(FunctionType *FT, StringRef FName);
195
196 public:
197  DataFlowSanitizer(StringRef ABIListFile = StringRef(),
198                    void *(*getArgTLS)() = 0, void *(*getRetValTLS)() = 0);
199  static char ID;
200  bool doInitialization(Module &M);
201  bool runOnModule(Module &M);
202};
203
204struct DFSanFunction {
205  DataFlowSanitizer &DFS;
206  Function *F;
207  DataFlowSanitizer::InstrumentedABI IA;
208  bool IsNativeABI;
209  Value *ArgTLSPtr;
210  Value *RetvalTLSPtr;
211  AllocaInst *LabelReturnAlloca;
212  DenseMap<Value *, Value *> ValShadowMap;
213  DenseMap<AllocaInst *, AllocaInst *> AllocaShadowMap;
214  std::vector<std::pair<PHINode *, PHINode *> > PHIFixups;
215  DenseSet<Instruction *> SkipInsts;
216  DenseSet<Value *> NonZeroChecks;
217
218  DFSanFunction(DataFlowSanitizer &DFS, Function *F, bool IsNativeABI)
219      : DFS(DFS), F(F), IA(DFS.getInstrumentedABI()),
220        IsNativeABI(IsNativeABI), ArgTLSPtr(0), RetvalTLSPtr(0),
221        LabelReturnAlloca(0) {}
222  Value *getArgTLSPtr();
223  Value *getArgTLS(unsigned Index, Instruction *Pos);
224  Value *getRetvalTLS();
225  Value *getShadow(Value *V);
226  void setShadow(Instruction *I, Value *Shadow);
227  Value *combineOperandShadows(Instruction *Inst);
228  Value *loadShadow(Value *ShadowAddr, uint64_t Size, uint64_t Align,
229                    Instruction *Pos);
230  void storeShadow(Value *Addr, uint64_t Size, uint64_t Align, Value *Shadow,
231                   Instruction *Pos);
232};
233
234class DFSanVisitor : public InstVisitor<DFSanVisitor> {
235 public:
236  DFSanFunction &DFSF;
237  DFSanVisitor(DFSanFunction &DFSF) : DFSF(DFSF) {}
238
239  void visitOperandShadowInst(Instruction &I);
240
241  void visitBinaryOperator(BinaryOperator &BO);
242  void visitCastInst(CastInst &CI);
243  void visitCmpInst(CmpInst &CI);
244  void visitGetElementPtrInst(GetElementPtrInst &GEPI);
245  void visitLoadInst(LoadInst &LI);
246  void visitStoreInst(StoreInst &SI);
247  void visitReturnInst(ReturnInst &RI);
248  void visitCallSite(CallSite CS);
249  void visitPHINode(PHINode &PN);
250  void visitExtractElementInst(ExtractElementInst &I);
251  void visitInsertElementInst(InsertElementInst &I);
252  void visitShuffleVectorInst(ShuffleVectorInst &I);
253  void visitExtractValueInst(ExtractValueInst &I);
254  void visitInsertValueInst(InsertValueInst &I);
255  void visitAllocaInst(AllocaInst &I);
256  void visitSelectInst(SelectInst &I);
257  void visitMemSetInst(MemSetInst &I);
258  void visitMemTransferInst(MemTransferInst &I);
259};
260
261}
262
263char DataFlowSanitizer::ID;
264INITIALIZE_PASS(DataFlowSanitizer, "dfsan",
265                "DataFlowSanitizer: dynamic data flow analysis.", false, false)
266
267ModulePass *llvm::createDataFlowSanitizerPass(StringRef ABIListFile,
268                                              void *(*getArgTLS)(),
269                                              void *(*getRetValTLS)()) {
270  return new DataFlowSanitizer(ABIListFile, getArgTLS, getRetValTLS);
271}
272
273DataFlowSanitizer::DataFlowSanitizer(StringRef ABIListFile,
274                                     void *(*getArgTLS)(),
275                                     void *(*getRetValTLS)())
276    : ModulePass(ID), GetArgTLSPtr(getArgTLS), GetRetvalTLSPtr(getRetValTLS),
277      ABIList(SpecialCaseList::createOrDie(ABIListFile.empty() ? ClABIListFile
278                                                               : ABIListFile)) {
279}
280
281FunctionType *DataFlowSanitizer::getArgsFunctionType(FunctionType *T) {
282  llvm::SmallVector<Type *, 4> ArgTypes;
283  std::copy(T->param_begin(), T->param_end(), std::back_inserter(ArgTypes));
284  for (unsigned i = 0, e = T->getNumParams(); i != e; ++i)
285    ArgTypes.push_back(ShadowTy);
286  if (T->isVarArg())
287    ArgTypes.push_back(ShadowPtrTy);
288  Type *RetType = T->getReturnType();
289  if (!RetType->isVoidTy())
290    RetType = StructType::get(RetType, ShadowTy, (Type *)0);
291  return FunctionType::get(RetType, ArgTypes, T->isVarArg());
292}
293
294FunctionType *DataFlowSanitizer::getTrampolineFunctionType(FunctionType *T) {
295  assert(!T->isVarArg());
296  llvm::SmallVector<Type *, 4> ArgTypes;
297  ArgTypes.push_back(T->getPointerTo());
298  std::copy(T->param_begin(), T->param_end(), std::back_inserter(ArgTypes));
299  for (unsigned i = 0, e = T->getNumParams(); i != e; ++i)
300    ArgTypes.push_back(ShadowTy);
301  Type *RetType = T->getReturnType();
302  if (!RetType->isVoidTy())
303    ArgTypes.push_back(ShadowPtrTy);
304  return FunctionType::get(T->getReturnType(), ArgTypes, false);
305}
306
307FunctionType *DataFlowSanitizer::getCustomFunctionType(FunctionType *T) {
308  assert(!T->isVarArg());
309  llvm::SmallVector<Type *, 4> ArgTypes;
310  for (FunctionType::param_iterator i = T->param_begin(), e = T->param_end();
311       i != e; ++i) {
312    FunctionType *FT;
313    if (isa<PointerType>(*i) && (FT = dyn_cast<FunctionType>(cast<PointerType>(
314                                     *i)->getElementType()))) {
315      ArgTypes.push_back(getTrampolineFunctionType(FT)->getPointerTo());
316      ArgTypes.push_back(Type::getInt8PtrTy(*Ctx));
317    } else {
318      ArgTypes.push_back(*i);
319    }
320  }
321  for (unsigned i = 0, e = T->getNumParams(); i != e; ++i)
322    ArgTypes.push_back(ShadowTy);
323  Type *RetType = T->getReturnType();
324  if (!RetType->isVoidTy())
325    ArgTypes.push_back(ShadowPtrTy);
326  return FunctionType::get(T->getReturnType(), ArgTypes, false);
327}
328
329bool DataFlowSanitizer::doInitialization(Module &M) {
330  DL = getAnalysisIfAvailable<DataLayout>();
331  if (!DL)
332    return false;
333
334  Mod = &M;
335  Ctx = &M.getContext();
336  ShadowTy = IntegerType::get(*Ctx, ShadowWidth);
337  ShadowPtrTy = PointerType::getUnqual(ShadowTy);
338  IntptrTy = DL->getIntPtrType(*Ctx);
339  ZeroShadow = ConstantInt::getSigned(ShadowTy, 0);
340  ShadowPtrMask = ConstantInt::getSigned(IntptrTy, ~0x700000000000LL);
341  ShadowPtrMul = ConstantInt::getSigned(IntptrTy, ShadowWidth / 8);
342
343  Type *DFSanUnionArgs[2] = { ShadowTy, ShadowTy };
344  DFSanUnionFnTy =
345      FunctionType::get(ShadowTy, DFSanUnionArgs, /*isVarArg=*/ false);
346  Type *DFSanUnionLoadArgs[2] = { ShadowPtrTy, IntptrTy };
347  DFSanUnionLoadFnTy =
348      FunctionType::get(ShadowTy, DFSanUnionLoadArgs, /*isVarArg=*/ false);
349  DFSanUnimplementedFnTy = FunctionType::get(
350      Type::getVoidTy(*Ctx), Type::getInt8PtrTy(*Ctx), /*isVarArg=*/false);
351  Type *DFSanSetLabelArgs[3] = { ShadowTy, Type::getInt8PtrTy(*Ctx), IntptrTy };
352  DFSanSetLabelFnTy = FunctionType::get(Type::getVoidTy(*Ctx),
353                                        DFSanSetLabelArgs, /*isVarArg=*/false);
354  DFSanNonzeroLabelFnTy = FunctionType::get(
355      Type::getVoidTy(*Ctx), ArrayRef<Type *>(), /*isVarArg=*/false);
356
357  if (GetArgTLSPtr) {
358    Type *ArgTLSTy = ArrayType::get(ShadowTy, 64);
359    ArgTLS = 0;
360    GetArgTLS = ConstantExpr::getIntToPtr(
361        ConstantInt::get(IntptrTy, uintptr_t(GetArgTLSPtr)),
362        PointerType::getUnqual(
363            FunctionType::get(PointerType::getUnqual(ArgTLSTy), (Type *)0)));
364  }
365  if (GetRetvalTLSPtr) {
366    RetvalTLS = 0;
367    GetRetvalTLS = ConstantExpr::getIntToPtr(
368        ConstantInt::get(IntptrTy, uintptr_t(GetRetvalTLSPtr)),
369        PointerType::getUnqual(
370            FunctionType::get(PointerType::getUnqual(ShadowTy), (Type *)0)));
371  }
372
373  ColdCallWeights = MDBuilder(*Ctx).createBranchWeights(1, 1000);
374  return true;
375}
376
377bool DataFlowSanitizer::isInstrumented(const Function *F) {
378  return !ABIList->isIn(*F, "uninstrumented");
379}
380
381bool DataFlowSanitizer::isInstrumented(const GlobalAlias *GA) {
382  return !ABIList->isIn(*GA, "uninstrumented");
383}
384
385DataFlowSanitizer::InstrumentedABI DataFlowSanitizer::getInstrumentedABI() {
386  return ClArgsABI ? IA_Args : IA_TLS;
387}
388
389DataFlowSanitizer::WrapperKind DataFlowSanitizer::getWrapperKind(Function *F) {
390  if (ABIList->isIn(*F, "functional"))
391    return WK_Functional;
392  if (ABIList->isIn(*F, "discard"))
393    return WK_Discard;
394  if (ABIList->isIn(*F, "custom"))
395    return WK_Custom;
396
397  return WK_Warning;
398}
399
400void DataFlowSanitizer::addGlobalNamePrefix(GlobalValue *GV) {
401  std::string GVName = GV->getName(), Prefix = "dfs$";
402  GV->setName(Prefix + GVName);
403
404  // Try to change the name of the function in module inline asm.  We only do
405  // this for specific asm directives, currently only ".symver", to try to avoid
406  // corrupting asm which happens to contain the symbol name as a substring.
407  // Note that the substitution for .symver assumes that the versioned symbol
408  // also has an instrumented name.
409  std::string Asm = GV->getParent()->getModuleInlineAsm();
410  std::string SearchStr = ".symver " + GVName + ",";
411  size_t Pos = Asm.find(SearchStr);
412  if (Pos != std::string::npos) {
413    Asm.replace(Pos, SearchStr.size(),
414                ".symver " + Prefix + GVName + "," + Prefix);
415    GV->getParent()->setModuleInlineAsm(Asm);
416  }
417}
418
419Function *
420DataFlowSanitizer::buildWrapperFunction(Function *F, StringRef NewFName,
421                                        GlobalValue::LinkageTypes NewFLink,
422                                        FunctionType *NewFT) {
423  FunctionType *FT = F->getFunctionType();
424  Function *NewF = Function::Create(NewFT, NewFLink, NewFName,
425                                    F->getParent());
426  NewF->copyAttributesFrom(F);
427  NewF->removeAttributes(
428      AttributeSet::ReturnIndex,
429      AttributeFuncs::typeIncompatible(NewFT->getReturnType(),
430                                       AttributeSet::ReturnIndex));
431
432  BasicBlock *BB = BasicBlock::Create(*Ctx, "entry", NewF);
433  std::vector<Value *> Args;
434  unsigned n = FT->getNumParams();
435  for (Function::arg_iterator ai = NewF->arg_begin(); n != 0; ++ai, --n)
436    Args.push_back(&*ai);
437  CallInst *CI = CallInst::Create(F, Args, "", BB);
438  if (FT->getReturnType()->isVoidTy())
439    ReturnInst::Create(*Ctx, BB);
440  else
441    ReturnInst::Create(*Ctx, CI, BB);
442
443  return NewF;
444}
445
446Constant *DataFlowSanitizer::getOrBuildTrampolineFunction(FunctionType *FT,
447                                                          StringRef FName) {
448  FunctionType *FTT = getTrampolineFunctionType(FT);
449  Constant *C = Mod->getOrInsertFunction(FName, FTT);
450  Function *F = dyn_cast<Function>(C);
451  if (F && F->isDeclaration()) {
452    F->setLinkage(GlobalValue::LinkOnceODRLinkage);
453    BasicBlock *BB = BasicBlock::Create(*Ctx, "entry", F);
454    std::vector<Value *> Args;
455    Function::arg_iterator AI = F->arg_begin(); ++AI;
456    for (unsigned N = FT->getNumParams(); N != 0; ++AI, --N)
457      Args.push_back(&*AI);
458    CallInst *CI =
459        CallInst::Create(&F->getArgumentList().front(), Args, "", BB);
460    ReturnInst *RI;
461    if (FT->getReturnType()->isVoidTy())
462      RI = ReturnInst::Create(*Ctx, BB);
463    else
464      RI = ReturnInst::Create(*Ctx, CI, BB);
465
466    DFSanFunction DFSF(*this, F, /*IsNativeABI=*/true);
467    Function::arg_iterator ValAI = F->arg_begin(), ShadowAI = AI; ++ValAI;
468    for (unsigned N = FT->getNumParams(); N != 0; ++ValAI, ++ShadowAI, --N)
469      DFSF.ValShadowMap[ValAI] = ShadowAI;
470    DFSanVisitor(DFSF).visitCallInst(*CI);
471    if (!FT->getReturnType()->isVoidTy())
472      new StoreInst(DFSF.getShadow(RI->getReturnValue()),
473                    &F->getArgumentList().back(), RI);
474  }
475
476  return C;
477}
478
479bool DataFlowSanitizer::runOnModule(Module &M) {
480  if (!DL)
481    return false;
482
483  if (ABIList->isIn(M, "skip"))
484    return false;
485
486  if (!GetArgTLSPtr) {
487    Type *ArgTLSTy = ArrayType::get(ShadowTy, 64);
488    ArgTLS = Mod->getOrInsertGlobal("__dfsan_arg_tls", ArgTLSTy);
489    if (GlobalVariable *G = dyn_cast<GlobalVariable>(ArgTLS))
490      G->setThreadLocalMode(GlobalVariable::InitialExecTLSModel);
491  }
492  if (!GetRetvalTLSPtr) {
493    RetvalTLS = Mod->getOrInsertGlobal("__dfsan_retval_tls", ShadowTy);
494    if (GlobalVariable *G = dyn_cast<GlobalVariable>(RetvalTLS))
495      G->setThreadLocalMode(GlobalVariable::InitialExecTLSModel);
496  }
497
498  DFSanUnionFn = Mod->getOrInsertFunction("__dfsan_union", DFSanUnionFnTy);
499  if (Function *F = dyn_cast<Function>(DFSanUnionFn)) {
500    F->addAttribute(AttributeSet::FunctionIndex, Attribute::ReadNone);
501    F->addAttribute(AttributeSet::ReturnIndex, Attribute::ZExt);
502    F->addAttribute(1, Attribute::ZExt);
503    F->addAttribute(2, Attribute::ZExt);
504  }
505  DFSanUnionLoadFn =
506      Mod->getOrInsertFunction("__dfsan_union_load", DFSanUnionLoadFnTy);
507  if (Function *F = dyn_cast<Function>(DFSanUnionLoadFn)) {
508    F->addAttribute(AttributeSet::ReturnIndex, Attribute::ZExt);
509  }
510  DFSanUnimplementedFn =
511      Mod->getOrInsertFunction("__dfsan_unimplemented", DFSanUnimplementedFnTy);
512  DFSanSetLabelFn =
513      Mod->getOrInsertFunction("__dfsan_set_label", DFSanSetLabelFnTy);
514  if (Function *F = dyn_cast<Function>(DFSanSetLabelFn)) {
515    F->addAttribute(1, Attribute::ZExt);
516  }
517  DFSanNonzeroLabelFn =
518      Mod->getOrInsertFunction("__dfsan_nonzero_label", DFSanNonzeroLabelFnTy);
519
520  std::vector<Function *> FnsToInstrument;
521  llvm::SmallPtrSet<Function *, 2> FnsWithNativeABI;
522  for (Module::iterator i = M.begin(), e = M.end(); i != e; ++i) {
523    if (!i->isIntrinsic() &&
524        i != DFSanUnionFn &&
525        i != DFSanUnionLoadFn &&
526        i != DFSanUnimplementedFn &&
527        i != DFSanSetLabelFn &&
528        i != DFSanNonzeroLabelFn)
529      FnsToInstrument.push_back(&*i);
530  }
531
532  // Give function aliases prefixes when necessary, and build wrappers where the
533  // instrumentedness is inconsistent.
534  for (Module::alias_iterator i = M.alias_begin(), e = M.alias_end(); i != e;) {
535    GlobalAlias *GA = &*i;
536    ++i;
537    // Don't stop on weak.  We assume people aren't playing games with the
538    // instrumentedness of overridden weak aliases.
539    if (Function *F = dyn_cast<Function>(
540            GA->resolveAliasedGlobal(/*stopOnWeak=*/false))) {
541      bool GAInst = isInstrumented(GA), FInst = isInstrumented(F);
542      if (GAInst && FInst) {
543        addGlobalNamePrefix(GA);
544      } else if (GAInst != FInst) {
545        // Non-instrumented alias of an instrumented function, or vice versa.
546        // Replace the alias with a native-ABI wrapper of the aliasee.  The pass
547        // below will take care of instrumenting it.
548        Function *NewF =
549            buildWrapperFunction(F, "", GA->getLinkage(), F->getFunctionType());
550        GA->replaceAllUsesWith(NewF);
551        NewF->takeName(GA);
552        GA->eraseFromParent();
553        FnsToInstrument.push_back(NewF);
554      }
555    }
556  }
557
558  AttrBuilder B;
559  B.addAttribute(Attribute::ReadOnly).addAttribute(Attribute::ReadNone);
560  ReadOnlyNoneAttrs = AttributeSet::get(*Ctx, AttributeSet::FunctionIndex, B);
561
562  // First, change the ABI of every function in the module.  ABI-listed
563  // functions keep their original ABI and get a wrapper function.
564  for (std::vector<Function *>::iterator i = FnsToInstrument.begin(),
565                                         e = FnsToInstrument.end();
566       i != e; ++i) {
567    Function &F = **i;
568    FunctionType *FT = F.getFunctionType();
569
570    bool IsZeroArgsVoidRet = (FT->getNumParams() == 0 && !FT->isVarArg() &&
571                              FT->getReturnType()->isVoidTy());
572
573    if (isInstrumented(&F)) {
574      // Instrumented functions get a 'dfs$' prefix.  This allows us to more
575      // easily identify cases of mismatching ABIs.
576      if (getInstrumentedABI() == IA_Args && !IsZeroArgsVoidRet) {
577        FunctionType *NewFT = getArgsFunctionType(FT);
578        Function *NewF = Function::Create(NewFT, F.getLinkage(), "", &M);
579        NewF->copyAttributesFrom(&F);
580        NewF->removeAttributes(
581            AttributeSet::ReturnIndex,
582            AttributeFuncs::typeIncompatible(NewFT->getReturnType(),
583                                             AttributeSet::ReturnIndex));
584        for (Function::arg_iterator FArg = F.arg_begin(),
585                                    NewFArg = NewF->arg_begin(),
586                                    FArgEnd = F.arg_end();
587             FArg != FArgEnd; ++FArg, ++NewFArg) {
588          FArg->replaceAllUsesWith(NewFArg);
589        }
590        NewF->getBasicBlockList().splice(NewF->begin(), F.getBasicBlockList());
591
592        for (Function::use_iterator ui = F.use_begin(), ue = F.use_end();
593             ui != ue;) {
594          BlockAddress *BA = dyn_cast<BlockAddress>(ui.getUse().getUser());
595          ++ui;
596          if (BA) {
597            BA->replaceAllUsesWith(
598                BlockAddress::get(NewF, BA->getBasicBlock()));
599            delete BA;
600          }
601        }
602        F.replaceAllUsesWith(
603            ConstantExpr::getBitCast(NewF, PointerType::getUnqual(FT)));
604        NewF->takeName(&F);
605        F.eraseFromParent();
606        *i = NewF;
607        addGlobalNamePrefix(NewF);
608      } else {
609        addGlobalNamePrefix(&F);
610      }
611               // Hopefully, nobody will try to indirectly call a vararg
612               // function... yet.
613    } else if (FT->isVarArg()) {
614      UnwrappedFnMap[&F] = &F;
615      *i = 0;
616    } else if (!IsZeroArgsVoidRet || getWrapperKind(&F) == WK_Custom) {
617      // Build a wrapper function for F.  The wrapper simply calls F, and is
618      // added to FnsToInstrument so that any instrumentation according to its
619      // WrapperKind is done in the second pass below.
620      FunctionType *NewFT = getInstrumentedABI() == IA_Args
621                                ? getArgsFunctionType(FT)
622                                : FT;
623      Function *NewF = buildWrapperFunction(
624          &F, std::string("dfsw$") + std::string(F.getName()),
625          GlobalValue::LinkOnceODRLinkage, NewFT);
626      if (getInstrumentedABI() == IA_TLS)
627        NewF->removeAttributes(AttributeSet::FunctionIndex, ReadOnlyNoneAttrs);
628
629      Value *WrappedFnCst =
630          ConstantExpr::getBitCast(NewF, PointerType::getUnqual(FT));
631      F.replaceAllUsesWith(WrappedFnCst);
632      UnwrappedFnMap[WrappedFnCst] = &F;
633      *i = NewF;
634
635      if (!F.isDeclaration()) {
636        // This function is probably defining an interposition of an
637        // uninstrumented function and hence needs to keep the original ABI.
638        // But any functions it may call need to use the instrumented ABI, so
639        // we instrument it in a mode which preserves the original ABI.
640        FnsWithNativeABI.insert(&F);
641
642        // This code needs to rebuild the iterators, as they may be invalidated
643        // by the push_back, taking care that the new range does not include
644        // any functions added by this code.
645        size_t N = i - FnsToInstrument.begin(),
646               Count = e - FnsToInstrument.begin();
647        FnsToInstrument.push_back(&F);
648        i = FnsToInstrument.begin() + N;
649        e = FnsToInstrument.begin() + Count;
650      }
651    }
652  }
653
654  for (std::vector<Function *>::iterator i = FnsToInstrument.begin(),
655                                         e = FnsToInstrument.end();
656       i != e; ++i) {
657    if (!*i || (*i)->isDeclaration())
658      continue;
659
660    removeUnreachableBlocks(**i);
661
662    DFSanFunction DFSF(*this, *i, FnsWithNativeABI.count(*i));
663
664    // DFSanVisitor may create new basic blocks, which confuses df_iterator.
665    // Build a copy of the list before iterating over it.
666    llvm::SmallVector<BasicBlock *, 4> BBList;
667    std::copy(df_begin(&(*i)->getEntryBlock()), df_end(&(*i)->getEntryBlock()),
668              std::back_inserter(BBList));
669
670    for (llvm::SmallVector<BasicBlock *, 4>::iterator i = BBList.begin(),
671                                                      e = BBList.end();
672         i != e; ++i) {
673      Instruction *Inst = &(*i)->front();
674      while (1) {
675        // DFSanVisitor may split the current basic block, changing the current
676        // instruction's next pointer and moving the next instruction to the
677        // tail block from which we should continue.
678        Instruction *Next = Inst->getNextNode();
679        // DFSanVisitor may delete Inst, so keep track of whether it was a
680        // terminator.
681        bool IsTerminator = isa<TerminatorInst>(Inst);
682        if (!DFSF.SkipInsts.count(Inst))
683          DFSanVisitor(DFSF).visit(Inst);
684        if (IsTerminator)
685          break;
686        Inst = Next;
687      }
688    }
689
690    // We will not necessarily be able to compute the shadow for every phi node
691    // until we have visited every block.  Therefore, the code that handles phi
692    // nodes adds them to the PHIFixups list so that they can be properly
693    // handled here.
694    for (std::vector<std::pair<PHINode *, PHINode *> >::iterator
695             i = DFSF.PHIFixups.begin(),
696             e = DFSF.PHIFixups.end();
697         i != e; ++i) {
698      for (unsigned val = 0, n = i->first->getNumIncomingValues(); val != n;
699           ++val) {
700        i->second->setIncomingValue(
701            val, DFSF.getShadow(i->first->getIncomingValue(val)));
702      }
703    }
704
705    // -dfsan-debug-nonzero-labels will split the CFG in all kinds of crazy
706    // places (i.e. instructions in basic blocks we haven't even begun visiting
707    // yet).  To make our life easier, do this work in a pass after the main
708    // instrumentation.
709    if (ClDebugNonzeroLabels) {
710      for (DenseSet<Value *>::iterator i = DFSF.NonZeroChecks.begin(),
711                                       e = DFSF.NonZeroChecks.end();
712           i != e; ++i) {
713        Instruction *Pos;
714        if (Instruction *I = dyn_cast<Instruction>(*i))
715          Pos = I->getNextNode();
716        else
717          Pos = DFSF.F->getEntryBlock().begin();
718        while (isa<PHINode>(Pos) || isa<AllocaInst>(Pos))
719          Pos = Pos->getNextNode();
720        IRBuilder<> IRB(Pos);
721        Instruction *NeInst = cast<Instruction>(
722            IRB.CreateICmpNE(*i, DFSF.DFS.ZeroShadow));
723        BranchInst *BI = cast<BranchInst>(SplitBlockAndInsertIfThen(
724            NeInst, /*Unreachable=*/ false, ColdCallWeights));
725        IRBuilder<> ThenIRB(BI);
726        ThenIRB.CreateCall(DFSF.DFS.DFSanNonzeroLabelFn);
727      }
728    }
729  }
730
731  return false;
732}
733
734Value *DFSanFunction::getArgTLSPtr() {
735  if (ArgTLSPtr)
736    return ArgTLSPtr;
737  if (DFS.ArgTLS)
738    return ArgTLSPtr = DFS.ArgTLS;
739
740  IRBuilder<> IRB(F->getEntryBlock().begin());
741  return ArgTLSPtr = IRB.CreateCall(DFS.GetArgTLS);
742}
743
744Value *DFSanFunction::getRetvalTLS() {
745  if (RetvalTLSPtr)
746    return RetvalTLSPtr;
747  if (DFS.RetvalTLS)
748    return RetvalTLSPtr = DFS.RetvalTLS;
749
750  IRBuilder<> IRB(F->getEntryBlock().begin());
751  return RetvalTLSPtr = IRB.CreateCall(DFS.GetRetvalTLS);
752}
753
754Value *DFSanFunction::getArgTLS(unsigned Idx, Instruction *Pos) {
755  IRBuilder<> IRB(Pos);
756  return IRB.CreateConstGEP2_64(getArgTLSPtr(), 0, Idx);
757}
758
759Value *DFSanFunction::getShadow(Value *V) {
760  if (!isa<Argument>(V) && !isa<Instruction>(V))
761    return DFS.ZeroShadow;
762  Value *&Shadow = ValShadowMap[V];
763  if (!Shadow) {
764    if (Argument *A = dyn_cast<Argument>(V)) {
765      if (IsNativeABI)
766        return DFS.ZeroShadow;
767      switch (IA) {
768      case DataFlowSanitizer::IA_TLS: {
769        Value *ArgTLSPtr = getArgTLSPtr();
770        Instruction *ArgTLSPos =
771            DFS.ArgTLS ? &*F->getEntryBlock().begin()
772                       : cast<Instruction>(ArgTLSPtr)->getNextNode();
773        IRBuilder<> IRB(ArgTLSPos);
774        Shadow = IRB.CreateLoad(getArgTLS(A->getArgNo(), ArgTLSPos));
775        break;
776      }
777      case DataFlowSanitizer::IA_Args: {
778        unsigned ArgIdx = A->getArgNo() + F->getArgumentList().size() / 2;
779        Function::arg_iterator i = F->arg_begin();
780        while (ArgIdx--)
781          ++i;
782        Shadow = i;
783        assert(Shadow->getType() == DFS.ShadowTy);
784        break;
785      }
786      }
787      NonZeroChecks.insert(Shadow);
788    } else {
789      Shadow = DFS.ZeroShadow;
790    }
791  }
792  return Shadow;
793}
794
795void DFSanFunction::setShadow(Instruction *I, Value *Shadow) {
796  assert(!ValShadowMap.count(I));
797  assert(Shadow->getType() == DFS.ShadowTy);
798  ValShadowMap[I] = Shadow;
799}
800
801Value *DataFlowSanitizer::getShadowAddress(Value *Addr, Instruction *Pos) {
802  assert(Addr != RetvalTLS && "Reinstrumenting?");
803  IRBuilder<> IRB(Pos);
804  return IRB.CreateIntToPtr(
805      IRB.CreateMul(
806          IRB.CreateAnd(IRB.CreatePtrToInt(Addr, IntptrTy), ShadowPtrMask),
807          ShadowPtrMul),
808      ShadowPtrTy);
809}
810
811// Generates IR to compute the union of the two given shadows, inserting it
812// before Pos.  Returns the computed union Value.
813Value *DataFlowSanitizer::combineShadows(Value *V1, Value *V2,
814                                         Instruction *Pos) {
815  if (V1 == ZeroShadow)
816    return V2;
817  if (V2 == ZeroShadow)
818    return V1;
819  if (V1 == V2)
820    return V1;
821  IRBuilder<> IRB(Pos);
822  BasicBlock *Head = Pos->getParent();
823  Value *Ne = IRB.CreateICmpNE(V1, V2);
824  Instruction *NeInst = dyn_cast<Instruction>(Ne);
825  if (NeInst) {
826    BranchInst *BI = cast<BranchInst>(SplitBlockAndInsertIfThen(
827        NeInst, /*Unreachable=*/ false, ColdCallWeights));
828    IRBuilder<> ThenIRB(BI);
829    CallInst *Call = ThenIRB.CreateCall2(DFSanUnionFn, V1, V2);
830    Call->addAttribute(AttributeSet::ReturnIndex, Attribute::ZExt);
831    Call->addAttribute(1, Attribute::ZExt);
832    Call->addAttribute(2, Attribute::ZExt);
833
834    BasicBlock *Tail = BI->getSuccessor(0);
835    PHINode *Phi = PHINode::Create(ShadowTy, 2, "", Tail->begin());
836    Phi->addIncoming(Call, Call->getParent());
837    Phi->addIncoming(V1, Head);
838    Pos = Phi;
839    return Phi;
840  } else {
841    assert(0 && "todo");
842    return 0;
843  }
844}
845
846// A convenience function which folds the shadows of each of the operands
847// of the provided instruction Inst, inserting the IR before Inst.  Returns
848// the computed union Value.
849Value *DFSanFunction::combineOperandShadows(Instruction *Inst) {
850  if (Inst->getNumOperands() == 0)
851    return DFS.ZeroShadow;
852
853  Value *Shadow = getShadow(Inst->getOperand(0));
854  for (unsigned i = 1, n = Inst->getNumOperands(); i != n; ++i) {
855    Shadow = DFS.combineShadows(Shadow, getShadow(Inst->getOperand(i)), Inst);
856  }
857  return Shadow;
858}
859
860void DFSanVisitor::visitOperandShadowInst(Instruction &I) {
861  Value *CombinedShadow = DFSF.combineOperandShadows(&I);
862  DFSF.setShadow(&I, CombinedShadow);
863}
864
865// Generates IR to load shadow corresponding to bytes [Addr, Addr+Size), where
866// Addr has alignment Align, and take the union of each of those shadows.
867Value *DFSanFunction::loadShadow(Value *Addr, uint64_t Size, uint64_t Align,
868                                 Instruction *Pos) {
869  if (AllocaInst *AI = dyn_cast<AllocaInst>(Addr)) {
870    llvm::DenseMap<AllocaInst *, AllocaInst *>::iterator i =
871        AllocaShadowMap.find(AI);
872    if (i != AllocaShadowMap.end()) {
873      IRBuilder<> IRB(Pos);
874      return IRB.CreateLoad(i->second);
875    }
876  }
877
878  uint64_t ShadowAlign = Align * DFS.ShadowWidth / 8;
879  SmallVector<Value *, 2> Objs;
880  GetUnderlyingObjects(Addr, Objs, DFS.DL);
881  bool AllConstants = true;
882  for (SmallVector<Value *, 2>::iterator i = Objs.begin(), e = Objs.end();
883       i != e; ++i) {
884    if (isa<Function>(*i) || isa<BlockAddress>(*i))
885      continue;
886    if (isa<GlobalVariable>(*i) && cast<GlobalVariable>(*i)->isConstant())
887      continue;
888
889    AllConstants = false;
890    break;
891  }
892  if (AllConstants)
893    return DFS.ZeroShadow;
894
895  Value *ShadowAddr = DFS.getShadowAddress(Addr, Pos);
896  switch (Size) {
897  case 0:
898    return DFS.ZeroShadow;
899  case 1: {
900    LoadInst *LI = new LoadInst(ShadowAddr, "", Pos);
901    LI->setAlignment(ShadowAlign);
902    return LI;
903  }
904  case 2: {
905    IRBuilder<> IRB(Pos);
906    Value *ShadowAddr1 =
907        IRB.CreateGEP(ShadowAddr, ConstantInt::get(DFS.IntptrTy, 1));
908    return DFS.combineShadows(IRB.CreateAlignedLoad(ShadowAddr, ShadowAlign),
909                              IRB.CreateAlignedLoad(ShadowAddr1, ShadowAlign),
910                              Pos);
911  }
912  }
913  if (Size % (64 / DFS.ShadowWidth) == 0) {
914    // Fast path for the common case where each byte has identical shadow: load
915    // shadow 64 bits at a time, fall out to a __dfsan_union_load call if any
916    // shadow is non-equal.
917    BasicBlock *FallbackBB = BasicBlock::Create(*DFS.Ctx, "", F);
918    IRBuilder<> FallbackIRB(FallbackBB);
919    CallInst *FallbackCall = FallbackIRB.CreateCall2(
920        DFS.DFSanUnionLoadFn, ShadowAddr, ConstantInt::get(DFS.IntptrTy, Size));
921    FallbackCall->addAttribute(AttributeSet::ReturnIndex, Attribute::ZExt);
922
923    // Compare each of the shadows stored in the loaded 64 bits to each other,
924    // by computing (WideShadow rotl ShadowWidth) == WideShadow.
925    IRBuilder<> IRB(Pos);
926    Value *WideAddr =
927        IRB.CreateBitCast(ShadowAddr, Type::getInt64PtrTy(*DFS.Ctx));
928    Value *WideShadow = IRB.CreateAlignedLoad(WideAddr, ShadowAlign);
929    Value *TruncShadow = IRB.CreateTrunc(WideShadow, DFS.ShadowTy);
930    Value *ShlShadow = IRB.CreateShl(WideShadow, DFS.ShadowWidth);
931    Value *ShrShadow = IRB.CreateLShr(WideShadow, 64 - DFS.ShadowWidth);
932    Value *RotShadow = IRB.CreateOr(ShlShadow, ShrShadow);
933    Value *ShadowsEq = IRB.CreateICmpEQ(WideShadow, RotShadow);
934
935    BasicBlock *Head = Pos->getParent();
936    BasicBlock *Tail = Head->splitBasicBlock(Pos);
937    // In the following code LastBr will refer to the previous basic block's
938    // conditional branch instruction, whose true successor is fixed up to point
939    // to the next block during the loop below or to the tail after the final
940    // iteration.
941    BranchInst *LastBr = BranchInst::Create(FallbackBB, FallbackBB, ShadowsEq);
942    ReplaceInstWithInst(Head->getTerminator(), LastBr);
943
944    for (uint64_t Ofs = 64 / DFS.ShadowWidth; Ofs != Size;
945         Ofs += 64 / DFS.ShadowWidth) {
946      BasicBlock *NextBB = BasicBlock::Create(*DFS.Ctx, "", F);
947      IRBuilder<> NextIRB(NextBB);
948      WideAddr = NextIRB.CreateGEP(WideAddr, ConstantInt::get(DFS.IntptrTy, 1));
949      Value *NextWideShadow = NextIRB.CreateAlignedLoad(WideAddr, ShadowAlign);
950      ShadowsEq = NextIRB.CreateICmpEQ(WideShadow, NextWideShadow);
951      LastBr->setSuccessor(0, NextBB);
952      LastBr = NextIRB.CreateCondBr(ShadowsEq, FallbackBB, FallbackBB);
953    }
954
955    LastBr->setSuccessor(0, Tail);
956    FallbackIRB.CreateBr(Tail);
957    PHINode *Shadow = PHINode::Create(DFS.ShadowTy, 2, "", &Tail->front());
958    Shadow->addIncoming(FallbackCall, FallbackBB);
959    Shadow->addIncoming(TruncShadow, LastBr->getParent());
960    return Shadow;
961  }
962
963  IRBuilder<> IRB(Pos);
964  CallInst *FallbackCall = IRB.CreateCall2(
965      DFS.DFSanUnionLoadFn, ShadowAddr, ConstantInt::get(DFS.IntptrTy, Size));
966  FallbackCall->addAttribute(AttributeSet::ReturnIndex, Attribute::ZExt);
967  return FallbackCall;
968}
969
970void DFSanVisitor::visitLoadInst(LoadInst &LI) {
971  uint64_t Size = DFSF.DFS.DL->getTypeStoreSize(LI.getType());
972  uint64_t Align;
973  if (ClPreserveAlignment) {
974    Align = LI.getAlignment();
975    if (Align == 0)
976      Align = DFSF.DFS.DL->getABITypeAlignment(LI.getType());
977  } else {
978    Align = 1;
979  }
980  IRBuilder<> IRB(&LI);
981  Value *LoadedShadow =
982      DFSF.loadShadow(LI.getPointerOperand(), Size, Align, &LI);
983  Value *PtrShadow = DFSF.getShadow(LI.getPointerOperand());
984  Value *CombinedShadow = DFSF.DFS.combineShadows(LoadedShadow, PtrShadow, &LI);
985  if (CombinedShadow != DFSF.DFS.ZeroShadow)
986    DFSF.NonZeroChecks.insert(CombinedShadow);
987
988  DFSF.setShadow(&LI, CombinedShadow);
989}
990
991void DFSanFunction::storeShadow(Value *Addr, uint64_t Size, uint64_t Align,
992                                Value *Shadow, Instruction *Pos) {
993  if (AllocaInst *AI = dyn_cast<AllocaInst>(Addr)) {
994    llvm::DenseMap<AllocaInst *, AllocaInst *>::iterator i =
995        AllocaShadowMap.find(AI);
996    if (i != AllocaShadowMap.end()) {
997      IRBuilder<> IRB(Pos);
998      IRB.CreateStore(Shadow, i->second);
999      return;
1000    }
1001  }
1002
1003  uint64_t ShadowAlign = Align * DFS.ShadowWidth / 8;
1004  IRBuilder<> IRB(Pos);
1005  Value *ShadowAddr = DFS.getShadowAddress(Addr, Pos);
1006  if (Shadow == DFS.ZeroShadow) {
1007    IntegerType *ShadowTy = IntegerType::get(*DFS.Ctx, Size * DFS.ShadowWidth);
1008    Value *ExtZeroShadow = ConstantInt::get(ShadowTy, 0);
1009    Value *ExtShadowAddr =
1010        IRB.CreateBitCast(ShadowAddr, PointerType::getUnqual(ShadowTy));
1011    IRB.CreateAlignedStore(ExtZeroShadow, ExtShadowAddr, ShadowAlign);
1012    return;
1013  }
1014
1015  const unsigned ShadowVecSize = 128 / DFS.ShadowWidth;
1016  uint64_t Offset = 0;
1017  if (Size >= ShadowVecSize) {
1018    VectorType *ShadowVecTy = VectorType::get(DFS.ShadowTy, ShadowVecSize);
1019    Value *ShadowVec = UndefValue::get(ShadowVecTy);
1020    for (unsigned i = 0; i != ShadowVecSize; ++i) {
1021      ShadowVec = IRB.CreateInsertElement(
1022          ShadowVec, Shadow, ConstantInt::get(Type::getInt32Ty(*DFS.Ctx), i));
1023    }
1024    Value *ShadowVecAddr =
1025        IRB.CreateBitCast(ShadowAddr, PointerType::getUnqual(ShadowVecTy));
1026    do {
1027      Value *CurShadowVecAddr = IRB.CreateConstGEP1_32(ShadowVecAddr, Offset);
1028      IRB.CreateAlignedStore(ShadowVec, CurShadowVecAddr, ShadowAlign);
1029      Size -= ShadowVecSize;
1030      ++Offset;
1031    } while (Size >= ShadowVecSize);
1032    Offset *= ShadowVecSize;
1033  }
1034  while (Size > 0) {
1035    Value *CurShadowAddr = IRB.CreateConstGEP1_32(ShadowAddr, Offset);
1036    IRB.CreateAlignedStore(Shadow, CurShadowAddr, ShadowAlign);
1037    --Size;
1038    ++Offset;
1039  }
1040}
1041
1042void DFSanVisitor::visitStoreInst(StoreInst &SI) {
1043  uint64_t Size =
1044      DFSF.DFS.DL->getTypeStoreSize(SI.getValueOperand()->getType());
1045  uint64_t Align;
1046  if (ClPreserveAlignment) {
1047    Align = SI.getAlignment();
1048    if (Align == 0)
1049      Align = DFSF.DFS.DL->getABITypeAlignment(SI.getValueOperand()->getType());
1050  } else {
1051    Align = 1;
1052  }
1053  DFSF.storeShadow(SI.getPointerOperand(), Size, Align,
1054                   DFSF.getShadow(SI.getValueOperand()), &SI);
1055}
1056
1057void DFSanVisitor::visitBinaryOperator(BinaryOperator &BO) {
1058  visitOperandShadowInst(BO);
1059}
1060
1061void DFSanVisitor::visitCastInst(CastInst &CI) { visitOperandShadowInst(CI); }
1062
1063void DFSanVisitor::visitCmpInst(CmpInst &CI) { visitOperandShadowInst(CI); }
1064
1065void DFSanVisitor::visitGetElementPtrInst(GetElementPtrInst &GEPI) {
1066  visitOperandShadowInst(GEPI);
1067}
1068
1069void DFSanVisitor::visitExtractElementInst(ExtractElementInst &I) {
1070  visitOperandShadowInst(I);
1071}
1072
1073void DFSanVisitor::visitInsertElementInst(InsertElementInst &I) {
1074  visitOperandShadowInst(I);
1075}
1076
1077void DFSanVisitor::visitShuffleVectorInst(ShuffleVectorInst &I) {
1078  visitOperandShadowInst(I);
1079}
1080
1081void DFSanVisitor::visitExtractValueInst(ExtractValueInst &I) {
1082  visitOperandShadowInst(I);
1083}
1084
1085void DFSanVisitor::visitInsertValueInst(InsertValueInst &I) {
1086  visitOperandShadowInst(I);
1087}
1088
1089void DFSanVisitor::visitAllocaInst(AllocaInst &I) {
1090  bool AllLoadsStores = true;
1091  for (Instruction::use_iterator i = I.use_begin(), e = I.use_end(); i != e;
1092       ++i) {
1093    if (isa<LoadInst>(*i))
1094      continue;
1095
1096    if (StoreInst *SI = dyn_cast<StoreInst>(*i)) {
1097      if (SI->getPointerOperand() == &I)
1098        continue;
1099    }
1100
1101    AllLoadsStores = false;
1102    break;
1103  }
1104  if (AllLoadsStores) {
1105    IRBuilder<> IRB(&I);
1106    DFSF.AllocaShadowMap[&I] = IRB.CreateAlloca(DFSF.DFS.ShadowTy);
1107  }
1108  DFSF.setShadow(&I, DFSF.DFS.ZeroShadow);
1109}
1110
1111void DFSanVisitor::visitSelectInst(SelectInst &I) {
1112  Value *CondShadow = DFSF.getShadow(I.getCondition());
1113  Value *TrueShadow = DFSF.getShadow(I.getTrueValue());
1114  Value *FalseShadow = DFSF.getShadow(I.getFalseValue());
1115
1116  if (isa<VectorType>(I.getCondition()->getType())) {
1117    DFSF.setShadow(
1118        &I, DFSF.DFS.combineShadows(
1119                CondShadow,
1120                DFSF.DFS.combineShadows(TrueShadow, FalseShadow, &I), &I));
1121  } else {
1122    Value *ShadowSel;
1123    if (TrueShadow == FalseShadow) {
1124      ShadowSel = TrueShadow;
1125    } else {
1126      ShadowSel =
1127          SelectInst::Create(I.getCondition(), TrueShadow, FalseShadow, "", &I);
1128    }
1129    DFSF.setShadow(&I, DFSF.DFS.combineShadows(CondShadow, ShadowSel, &I));
1130  }
1131}
1132
1133void DFSanVisitor::visitMemSetInst(MemSetInst &I) {
1134  IRBuilder<> IRB(&I);
1135  Value *ValShadow = DFSF.getShadow(I.getValue());
1136  IRB.CreateCall3(
1137      DFSF.DFS.DFSanSetLabelFn, ValShadow,
1138      IRB.CreateBitCast(I.getDest(), Type::getInt8PtrTy(*DFSF.DFS.Ctx)),
1139      IRB.CreateZExtOrTrunc(I.getLength(), DFSF.DFS.IntptrTy));
1140}
1141
1142void DFSanVisitor::visitMemTransferInst(MemTransferInst &I) {
1143  IRBuilder<> IRB(&I);
1144  Value *DestShadow = DFSF.DFS.getShadowAddress(I.getDest(), &I);
1145  Value *SrcShadow = DFSF.DFS.getShadowAddress(I.getSource(), &I);
1146  Value *LenShadow = IRB.CreateMul(
1147      I.getLength(),
1148      ConstantInt::get(I.getLength()->getType(), DFSF.DFS.ShadowWidth / 8));
1149  Value *AlignShadow;
1150  if (ClPreserveAlignment) {
1151    AlignShadow = IRB.CreateMul(I.getAlignmentCst(),
1152                                ConstantInt::get(I.getAlignmentCst()->getType(),
1153                                                 DFSF.DFS.ShadowWidth / 8));
1154  } else {
1155    AlignShadow = ConstantInt::get(I.getAlignmentCst()->getType(),
1156                                   DFSF.DFS.ShadowWidth / 8);
1157  }
1158  Type *Int8Ptr = Type::getInt8PtrTy(*DFSF.DFS.Ctx);
1159  DestShadow = IRB.CreateBitCast(DestShadow, Int8Ptr);
1160  SrcShadow = IRB.CreateBitCast(SrcShadow, Int8Ptr);
1161  IRB.CreateCall5(I.getCalledValue(), DestShadow, SrcShadow, LenShadow,
1162                  AlignShadow, I.getVolatileCst());
1163}
1164
1165void DFSanVisitor::visitReturnInst(ReturnInst &RI) {
1166  if (!DFSF.IsNativeABI && RI.getReturnValue()) {
1167    switch (DFSF.IA) {
1168    case DataFlowSanitizer::IA_TLS: {
1169      Value *S = DFSF.getShadow(RI.getReturnValue());
1170      IRBuilder<> IRB(&RI);
1171      IRB.CreateStore(S, DFSF.getRetvalTLS());
1172      break;
1173    }
1174    case DataFlowSanitizer::IA_Args: {
1175      IRBuilder<> IRB(&RI);
1176      Type *RT = DFSF.F->getFunctionType()->getReturnType();
1177      Value *InsVal =
1178          IRB.CreateInsertValue(UndefValue::get(RT), RI.getReturnValue(), 0);
1179      Value *InsShadow =
1180          IRB.CreateInsertValue(InsVal, DFSF.getShadow(RI.getReturnValue()), 1);
1181      RI.setOperand(0, InsShadow);
1182      break;
1183    }
1184    }
1185  }
1186}
1187
1188void DFSanVisitor::visitCallSite(CallSite CS) {
1189  Function *F = CS.getCalledFunction();
1190  if ((F && F->isIntrinsic()) || isa<InlineAsm>(CS.getCalledValue())) {
1191    visitOperandShadowInst(*CS.getInstruction());
1192    return;
1193  }
1194
1195  IRBuilder<> IRB(CS.getInstruction());
1196
1197  DenseMap<Value *, Function *>::iterator i =
1198      DFSF.DFS.UnwrappedFnMap.find(CS.getCalledValue());
1199  if (i != DFSF.DFS.UnwrappedFnMap.end()) {
1200    Function *F = i->second;
1201    switch (DFSF.DFS.getWrapperKind(F)) {
1202    case DataFlowSanitizer::WK_Warning: {
1203      CS.setCalledFunction(F);
1204      IRB.CreateCall(DFSF.DFS.DFSanUnimplementedFn,
1205                     IRB.CreateGlobalStringPtr(F->getName()));
1206      DFSF.setShadow(CS.getInstruction(), DFSF.DFS.ZeroShadow);
1207      return;
1208    }
1209    case DataFlowSanitizer::WK_Discard: {
1210      CS.setCalledFunction(F);
1211      DFSF.setShadow(CS.getInstruction(), DFSF.DFS.ZeroShadow);
1212      return;
1213    }
1214    case DataFlowSanitizer::WK_Functional: {
1215      CS.setCalledFunction(F);
1216      visitOperandShadowInst(*CS.getInstruction());
1217      return;
1218    }
1219    case DataFlowSanitizer::WK_Custom: {
1220      // Don't try to handle invokes of custom functions, it's too complicated.
1221      // Instead, invoke the dfsw$ wrapper, which will in turn call the __dfsw_
1222      // wrapper.
1223      if (CallInst *CI = dyn_cast<CallInst>(CS.getInstruction())) {
1224        FunctionType *FT = F->getFunctionType();
1225        FunctionType *CustomFT = DFSF.DFS.getCustomFunctionType(FT);
1226        std::string CustomFName = "__dfsw_";
1227        CustomFName += F->getName();
1228        Constant *CustomF =
1229            DFSF.DFS.Mod->getOrInsertFunction(CustomFName, CustomFT);
1230        if (Function *CustomFn = dyn_cast<Function>(CustomF)) {
1231          CustomFn->copyAttributesFrom(F);
1232
1233          // Custom functions returning non-void will write to the return label.
1234          if (!FT->getReturnType()->isVoidTy()) {
1235            CustomFn->removeAttributes(AttributeSet::FunctionIndex,
1236                                       DFSF.DFS.ReadOnlyNoneAttrs);
1237          }
1238        }
1239
1240        std::vector<Value *> Args;
1241
1242        CallSite::arg_iterator i = CS.arg_begin();
1243        for (unsigned n = FT->getNumParams(); n != 0; ++i, --n) {
1244          Type *T = (*i)->getType();
1245          FunctionType *ParamFT;
1246          if (isa<PointerType>(T) &&
1247              (ParamFT = dyn_cast<FunctionType>(
1248                   cast<PointerType>(T)->getElementType()))) {
1249            std::string TName = "dfst";
1250            TName += utostr(FT->getNumParams() - n);
1251            TName += "$";
1252            TName += F->getName();
1253            Constant *T = DFSF.DFS.getOrBuildTrampolineFunction(ParamFT, TName);
1254            Args.push_back(T);
1255            Args.push_back(
1256                IRB.CreateBitCast(*i, Type::getInt8PtrTy(*DFSF.DFS.Ctx)));
1257          } else {
1258            Args.push_back(*i);
1259          }
1260        }
1261
1262        i = CS.arg_begin();
1263        for (unsigned n = FT->getNumParams(); n != 0; ++i, --n)
1264          Args.push_back(DFSF.getShadow(*i));
1265
1266        if (!FT->getReturnType()->isVoidTy()) {
1267          if (!DFSF.LabelReturnAlloca) {
1268            DFSF.LabelReturnAlloca =
1269                new AllocaInst(DFSF.DFS.ShadowTy, "labelreturn",
1270                               DFSF.F->getEntryBlock().begin());
1271          }
1272          Args.push_back(DFSF.LabelReturnAlloca);
1273        }
1274
1275        CallInst *CustomCI = IRB.CreateCall(CustomF, Args);
1276        CustomCI->setCallingConv(CI->getCallingConv());
1277        CustomCI->setAttributes(CI->getAttributes());
1278
1279        if (!FT->getReturnType()->isVoidTy()) {
1280          LoadInst *LabelLoad = IRB.CreateLoad(DFSF.LabelReturnAlloca);
1281          DFSF.setShadow(CustomCI, LabelLoad);
1282        }
1283
1284        CI->replaceAllUsesWith(CustomCI);
1285        CI->eraseFromParent();
1286        return;
1287      }
1288      break;
1289    }
1290    }
1291  }
1292
1293  FunctionType *FT = cast<FunctionType>(
1294      CS.getCalledValue()->getType()->getPointerElementType());
1295  if (DFSF.DFS.getInstrumentedABI() == DataFlowSanitizer::IA_TLS) {
1296    for (unsigned i = 0, n = FT->getNumParams(); i != n; ++i) {
1297      IRB.CreateStore(DFSF.getShadow(CS.getArgument(i)),
1298                      DFSF.getArgTLS(i, CS.getInstruction()));
1299    }
1300  }
1301
1302  Instruction *Next = 0;
1303  if (!CS.getType()->isVoidTy()) {
1304    if (InvokeInst *II = dyn_cast<InvokeInst>(CS.getInstruction())) {
1305      if (II->getNormalDest()->getSinglePredecessor()) {
1306        Next = II->getNormalDest()->begin();
1307      } else {
1308        BasicBlock *NewBB =
1309            SplitEdge(II->getParent(), II->getNormalDest(), &DFSF.DFS);
1310        Next = NewBB->begin();
1311      }
1312    } else {
1313      Next = CS->getNextNode();
1314    }
1315
1316    if (DFSF.DFS.getInstrumentedABI() == DataFlowSanitizer::IA_TLS) {
1317      IRBuilder<> NextIRB(Next);
1318      LoadInst *LI = NextIRB.CreateLoad(DFSF.getRetvalTLS());
1319      DFSF.SkipInsts.insert(LI);
1320      DFSF.setShadow(CS.getInstruction(), LI);
1321      DFSF.NonZeroChecks.insert(LI);
1322    }
1323  }
1324
1325  // Do all instrumentation for IA_Args down here to defer tampering with the
1326  // CFG in a way that SplitEdge may be able to detect.
1327  if (DFSF.DFS.getInstrumentedABI() == DataFlowSanitizer::IA_Args) {
1328    FunctionType *NewFT = DFSF.DFS.getArgsFunctionType(FT);
1329    Value *Func =
1330        IRB.CreateBitCast(CS.getCalledValue(), PointerType::getUnqual(NewFT));
1331    std::vector<Value *> Args;
1332
1333    CallSite::arg_iterator i = CS.arg_begin(), e = CS.arg_end();
1334    for (unsigned n = FT->getNumParams(); n != 0; ++i, --n)
1335      Args.push_back(*i);
1336
1337    i = CS.arg_begin();
1338    for (unsigned n = FT->getNumParams(); n != 0; ++i, --n)
1339      Args.push_back(DFSF.getShadow(*i));
1340
1341    if (FT->isVarArg()) {
1342      unsigned VarArgSize = CS.arg_size() - FT->getNumParams();
1343      ArrayType *VarArgArrayTy = ArrayType::get(DFSF.DFS.ShadowTy, VarArgSize);
1344      AllocaInst *VarArgShadow =
1345          new AllocaInst(VarArgArrayTy, "", DFSF.F->getEntryBlock().begin());
1346      Args.push_back(IRB.CreateConstGEP2_32(VarArgShadow, 0, 0));
1347      for (unsigned n = 0; i != e; ++i, ++n) {
1348        IRB.CreateStore(DFSF.getShadow(*i),
1349                        IRB.CreateConstGEP2_32(VarArgShadow, 0, n));
1350        Args.push_back(*i);
1351      }
1352    }
1353
1354    CallSite NewCS;
1355    if (InvokeInst *II = dyn_cast<InvokeInst>(CS.getInstruction())) {
1356      NewCS = IRB.CreateInvoke(Func, II->getNormalDest(), II->getUnwindDest(),
1357                               Args);
1358    } else {
1359      NewCS = IRB.CreateCall(Func, Args);
1360    }
1361    NewCS.setCallingConv(CS.getCallingConv());
1362    NewCS.setAttributes(CS.getAttributes().removeAttributes(
1363        *DFSF.DFS.Ctx, AttributeSet::ReturnIndex,
1364        AttributeFuncs::typeIncompatible(NewCS.getInstruction()->getType(),
1365                                         AttributeSet::ReturnIndex)));
1366
1367    if (Next) {
1368      ExtractValueInst *ExVal =
1369          ExtractValueInst::Create(NewCS.getInstruction(), 0, "", Next);
1370      DFSF.SkipInsts.insert(ExVal);
1371      ExtractValueInst *ExShadow =
1372          ExtractValueInst::Create(NewCS.getInstruction(), 1, "", Next);
1373      DFSF.SkipInsts.insert(ExShadow);
1374      DFSF.setShadow(ExVal, ExShadow);
1375      DFSF.NonZeroChecks.insert(ExShadow);
1376
1377      CS.getInstruction()->replaceAllUsesWith(ExVal);
1378    }
1379
1380    CS.getInstruction()->eraseFromParent();
1381  }
1382}
1383
1384void DFSanVisitor::visitPHINode(PHINode &PN) {
1385  PHINode *ShadowPN =
1386      PHINode::Create(DFSF.DFS.ShadowTy, PN.getNumIncomingValues(), "", &PN);
1387
1388  // Give the shadow phi node valid predecessors to fool SplitEdge into working.
1389  Value *UndefShadow = UndefValue::get(DFSF.DFS.ShadowTy);
1390  for (PHINode::block_iterator i = PN.block_begin(), e = PN.block_end(); i != e;
1391       ++i) {
1392    ShadowPN->addIncoming(UndefShadow, *i);
1393  }
1394
1395  DFSF.PHIFixups.push_back(std::make_pair(&PN, ShadowPN));
1396  DFSF.setShadow(&PN, ShadowPN);
1397}
1398