AddressSanitizer.cpp revision cc1d856d8e8e11407b3c6c1d08768f77c3722e38 
1//===-- AddressSanitizer.cpp - memory error detector ------------*- C++ -*-===//
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 file is a part of AddressSanitizer, an address sanity checker.
11// Details of the algorithm:
12//  http://code.google.com/p/address-sanitizer/wiki/AddressSanitizerAlgorithm
13//
14//===----------------------------------------------------------------------===//
15
16#define DEBUG_TYPE "asan"
17
18#include "llvm/ADT/ArrayRef.h"
19#include "llvm/ADT/OwningPtr.h"
20#include "llvm/ADT/SmallSet.h"
21#include "llvm/ADT/SmallString.h"
22#include "llvm/ADT/SmallVector.h"
23#include "llvm/ADT/StringExtras.h"
24#include "llvm/Function.h"
25#include "llvm/InlineAsm.h"
26#include "llvm/IntrinsicInst.h"
27#include "llvm/LLVMContext.h"
28#include "llvm/Module.h"
29#include "llvm/Support/CommandLine.h"
30#include "llvm/Support/DataTypes.h"
31#include "llvm/Support/Debug.h"
32#include "llvm/Support/IRBuilder.h"
33#include "llvm/Support/MemoryBuffer.h"
34#include "llvm/Support/Regex.h"
35#include "llvm/Support/raw_ostream.h"
36#include "llvm/Support/system_error.h"
37#include "llvm/Target/TargetData.h"
38#include "llvm/Target/TargetMachine.h"
39#include "llvm/Transforms/Instrumentation.h"
40#include "llvm/Transforms/Utils/BasicBlockUtils.h"
41#include "llvm/Transforms/Utils/ModuleUtils.h"
42#include "llvm/Type.h"
43
44#include <string>
45#include <algorithm>
46
47using namespace llvm;
48
49static const uint64_t kDefaultShadowScale = 3;
50static const uint64_t kDefaultShadowOffset32 = 1ULL << 29;
51static const uint64_t kDefaultShadowOffset64 = 1ULL << 44;
52
53static const size_t kMaxStackMallocSize = 1 << 16;  // 64K
54static const uintptr_t kCurrentStackFrameMagic = 0x41B58AB3;
55static const uintptr_t kRetiredStackFrameMagic = 0x45E0360E;
56
57static const char *kAsanModuleCtorName = "asan.module_ctor";
58static const char *kAsanReportErrorTemplate = "__asan_report_";
59static const char *kAsanRegisterGlobalsName = "__asan_register_globals";
60static const char *kAsanInitName = "__asan_init";
61static const char *kAsanMappingOffsetName = "__asan_mapping_offset";
62static const char *kAsanMappingScaleName = "__asan_mapping_scale";
63static const char *kAsanStackMallocName = "__asan_stack_malloc";
64static const char *kAsanStackFreeName = "__asan_stack_free";
65
66static const int kAsanStackLeftRedzoneMagic = 0xf1;
67static const int kAsanStackMidRedzoneMagic = 0xf2;
68static const int kAsanStackRightRedzoneMagic = 0xf3;
69static const int kAsanStackPartialRedzoneMagic = 0xf4;
70
71// Command-line flags.
72
73// This flag may need to be replaced with -f[no-]asan-reads.
74static cl::opt<bool> ClInstrumentReads("asan-instrument-reads",
75       cl::desc("instrument read instructions"), cl::Hidden, cl::init(true));
76static cl::opt<bool> ClInstrumentWrites("asan-instrument-writes",
77       cl::desc("instrument write instructions"), cl::Hidden, cl::init(true));
78// This flag may need to be replaced with -f[no]asan-stack.
79static cl::opt<bool> ClStack("asan-stack",
80       cl::desc("Handle stack memory"), cl::Hidden, cl::init(true));
81// This flag may need to be replaced with -f[no]asan-use-after-return.
82static cl::opt<bool> ClUseAfterReturn("asan-use-after-return",
83       cl::desc("Check return-after-free"), cl::Hidden, cl::init(false));
84// This flag may need to be replaced with -f[no]asan-globals.
85static cl::opt<bool> ClGlobals("asan-globals",
86       cl::desc("Handle global objects"), cl::Hidden, cl::init(true));
87static cl::opt<bool> ClMemIntrin("asan-memintrin",
88       cl::desc("Handle memset/memcpy/memmove"), cl::Hidden, cl::init(true));
89// This flag may need to be replaced with -fasan-blacklist.
90static cl::opt<std::string>  ClBlackListFile("asan-blacklist",
91       cl::desc("File containing the list of functions to ignore "
92                "during instrumentation"), cl::Hidden);
93static cl::opt<bool> ClUseCall("asan-use-call",
94       cl::desc("Use function call to generate a crash"), cl::Hidden,
95       cl::init(true));
96
97// These flags allow to change the shadow mapping.
98// The shadow mapping looks like
99//    Shadow = (Mem >> scale) + (1 << offset_log)
100static cl::opt<int> ClMappingScale("asan-mapping-scale",
101       cl::desc("scale of asan shadow mapping"), cl::Hidden, cl::init(0));
102static cl::opt<int> ClMappingOffsetLog("asan-mapping-offset-log",
103       cl::desc("offset of asan shadow mapping"), cl::Hidden, cl::init(-1));
104
105// Optimization flags. Not user visible, used mostly for testing
106// and benchmarking the tool.
107static cl::opt<bool> ClOpt("asan-opt",
108       cl::desc("Optimize instrumentation"), cl::Hidden, cl::init(true));
109static cl::opt<bool> ClOptSameTemp("asan-opt-same-temp",
110       cl::desc("Instrument the same temp just once"), cl::Hidden,
111       cl::init(true));
112static cl::opt<bool> ClOptGlobals("asan-opt-globals",
113       cl::desc("Don't instrument scalar globals"), cl::Hidden, cl::init(true));
114
115// Debug flags.
116static cl::opt<int> ClDebug("asan-debug", cl::desc("debug"), cl::Hidden,
117                            cl::init(0));
118static cl::opt<int> ClDebugStack("asan-debug-stack", cl::desc("debug stack"),
119                                 cl::Hidden, cl::init(0));
120static cl::opt<std::string> ClDebugFunc("asan-debug-func",
121                                        cl::Hidden, cl::desc("Debug func"));
122static cl::opt<int> ClDebugMin("asan-debug-min", cl::desc("Debug min inst"),
123                               cl::Hidden, cl::init(-1));
124static cl::opt<int> ClDebugMax("asan-debug-max", cl::desc("Debug man inst"),
125                               cl::Hidden, cl::init(-1));
126
127namespace {
128
129// Blacklisted functions are not instrumented.
130// The blacklist file contains one or more lines like this:
131// ---
132// fun:FunctionWildCard
133// ---
134// This is similar to the "ignore" feature of ThreadSanitizer.
135// http://code.google.com/p/data-race-test/wiki/ThreadSanitizerIgnores
136class BlackList {
137 public:
138  BlackList(const std::string &Path);
139  bool isIn(const Function &F);
140 private:
141  Regex *Functions;
142};
143
144/// AddressSanitizer: instrument the code in module to find memory bugs.
145struct AddressSanitizer : public ModulePass {
146  AddressSanitizer();
147  void instrumentMop(Instruction *I);
148  void instrumentAddress(Instruction *OrigIns, IRBuilder<> &IRB,
149                         Value *Addr, uint32_t TypeSize, bool IsWrite);
150  Instruction *generateCrashCode(IRBuilder<> &IRB, Value *Addr,
151                                 bool IsWrite, uint32_t TypeSize);
152  bool instrumentMemIntrinsic(MemIntrinsic *MI);
153  void instrumentMemIntrinsicParam(Instruction *OrigIns, Value *Addr,
154                                  Value *Size,
155                                   Instruction *InsertBefore, bool IsWrite);
156  Value *memToShadow(Value *Shadow, IRBuilder<> &IRB);
157  bool handleFunction(Module &M, Function &F);
158  bool poisonStackInFunction(Module &M, Function &F);
159  virtual bool runOnModule(Module &M);
160  bool insertGlobalRedzones(Module &M);
161  BranchInst *splitBlockAndInsertIfThen(Instruction *SplitBefore, Value *Cmp);
162  static char ID;  // Pass identification, replacement for typeid
163
164 private:
165
166  uint64_t getAllocaSizeInBytes(AllocaInst *AI) {
167    Type *Ty = AI->getAllocatedType();
168    uint64_t SizeInBytes = TD->getTypeStoreSizeInBits(Ty) / 8;
169    return SizeInBytes;
170  }
171  uint64_t getAlignedSize(uint64_t SizeInBytes) {
172    return ((SizeInBytes + RedzoneSize - 1)
173            / RedzoneSize) * RedzoneSize;
174  }
175  uint64_t getAlignedAllocaSize(AllocaInst *AI) {
176    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
177    return getAlignedSize(SizeInBytes);
178  }
179
180  void PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec, IRBuilder<> IRB,
181                   Value *ShadowBase, bool DoPoison);
182  bool LooksLikeCodeInBug11395(Instruction *I);
183
184  Module      *CurrentModule;
185  LLVMContext *C;
186  TargetData *TD;
187  uint64_t MappingOffset;
188  int MappingScale;
189  size_t RedzoneSize;
190  int LongSize;
191  Type *IntptrTy;
192  Type *IntptrPtrTy;
193  Function *AsanCtorFunction;
194  Function *AsanInitFunction;
195  Instruction *CtorInsertBefore;
196  OwningPtr<BlackList> BL;
197};
198}  // namespace
199
200char AddressSanitizer::ID = 0;
201INITIALIZE_PASS(AddressSanitizer, "asan",
202    "AddressSanitizer: detects use-after-free and out-of-bounds bugs.",
203    false, false)
204AddressSanitizer::AddressSanitizer() : ModulePass(ID) { }
205ModulePass *llvm::createAddressSanitizerPass() {
206  return new AddressSanitizer();
207}
208
209// Create a constant for Str so that we can pass it to the run-time lib.
210static GlobalVariable *createPrivateGlobalForString(Module &M, StringRef Str) {
211  Constant *StrConst = ConstantArray::get(M.getContext(), Str);
212  return new GlobalVariable(M, StrConst->getType(), true,
213                            GlobalValue::PrivateLinkage, StrConst, "");
214}
215
216// Split the basic block and insert an if-then code.
217// Before:
218//   Head
219//   SplitBefore
220//   Tail
221// After:
222//   Head
223//   if (Cmp)
224//     NewBasicBlock
225//   SplitBefore
226//   Tail
227//
228// Returns the NewBasicBlock's terminator.
229BranchInst *AddressSanitizer::splitBlockAndInsertIfThen(
230    Instruction *SplitBefore, Value *Cmp) {
231  BasicBlock *Head = SplitBefore->getParent();
232  BasicBlock *Tail = Head->splitBasicBlock(SplitBefore);
233  TerminatorInst *HeadOldTerm = Head->getTerminator();
234  BasicBlock *NewBasicBlock =
235      BasicBlock::Create(*C, "", Head->getParent());
236  BranchInst *HeadNewTerm = BranchInst::Create(/*ifTrue*/NewBasicBlock,
237                                               /*ifFalse*/Tail,
238                                               Cmp);
239  ReplaceInstWithInst(HeadOldTerm, HeadNewTerm);
240
241  BranchInst *CheckTerm = BranchInst::Create(Tail, NewBasicBlock);
242  return CheckTerm;
243}
244
245Value *AddressSanitizer::memToShadow(Value *Shadow, IRBuilder<> &IRB) {
246  // Shadow >> scale
247  Shadow = IRB.CreateLShr(Shadow, MappingScale);
248  if (MappingOffset == 0)
249    return Shadow;
250  // (Shadow >> scale) | offset
251  return IRB.CreateOr(Shadow, ConstantInt::get(IntptrTy,
252                                               MappingOffset));
253}
254
255void AddressSanitizer::instrumentMemIntrinsicParam(Instruction *OrigIns,
256    Value *Addr, Value *Size, Instruction *InsertBefore, bool IsWrite) {
257  // Check the first byte.
258  {
259    IRBuilder<> IRB(InsertBefore);
260    instrumentAddress(OrigIns, IRB, Addr, 8, IsWrite);
261  }
262  // Check the last byte.
263  {
264    IRBuilder<> IRB(InsertBefore);
265    Value *SizeMinusOne = IRB.CreateSub(
266        Size, ConstantInt::get(Size->getType(), 1));
267    SizeMinusOne = IRB.CreateIntCast(SizeMinusOne, IntptrTy, false);
268    Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
269    Value *AddrPlusSizeMinisOne = IRB.CreateAdd(AddrLong, SizeMinusOne);
270    instrumentAddress(OrigIns, IRB, AddrPlusSizeMinisOne, 8, IsWrite);
271  }
272}
273
274// Instrument memset/memmove/memcpy
275bool AddressSanitizer::instrumentMemIntrinsic(MemIntrinsic *MI) {
276  Value *Dst = MI->getDest();
277  MemTransferInst *MemTran = dyn_cast<MemTransferInst>(MI);
278  Value *Src = MemTran ? MemTran->getSource() : NULL;
279  Value *Length = MI->getLength();
280
281  Constant *ConstLength = dyn_cast<Constant>(Length);
282  Instruction *InsertBefore = MI;
283  if (ConstLength) {
284    if (ConstLength->isNullValue()) return false;
285  } else {
286    // The size is not a constant so it could be zero -- check at run-time.
287    IRBuilder<> IRB(InsertBefore);
288
289    Value *Cmp = IRB.CreateICmpNE(Length,
290                                   Constant::getNullValue(Length->getType()));
291    InsertBefore = splitBlockAndInsertIfThen(InsertBefore, Cmp);
292  }
293
294  instrumentMemIntrinsicParam(MI, Dst, Length, InsertBefore, true);
295  if (Src)
296    instrumentMemIntrinsicParam(MI, Src, Length, InsertBefore, false);
297  return true;
298}
299
300static Value *getLDSTOperand(Instruction *I) {
301  if (LoadInst *LI = dyn_cast<LoadInst>(I)) {
302    return LI->getPointerOperand();
303  }
304  return cast<StoreInst>(*I).getPointerOperand();
305}
306
307void AddressSanitizer::instrumentMop(Instruction *I) {
308  int IsWrite = isa<StoreInst>(*I);
309  Value *Addr = getLDSTOperand(I);
310  if (ClOpt && ClOptGlobals && isa<GlobalVariable>(Addr)) {
311    // We are accessing a global scalar variable. Nothing to catch here.
312    return;
313  }
314  Type *OrigPtrTy = Addr->getType();
315  Type *OrigTy = cast<PointerType>(OrigPtrTy)->getElementType();
316
317  assert(OrigTy->isSized());
318  uint32_t TypeSize = TD->getTypeStoreSizeInBits(OrigTy);
319
320  if (TypeSize != 8  && TypeSize != 16 &&
321      TypeSize != 32 && TypeSize != 64 && TypeSize != 128) {
322    // Ignore all unusual sizes.
323    return;
324  }
325
326  IRBuilder<> IRB(I);
327  instrumentAddress(I, IRB, Addr, TypeSize, IsWrite);
328}
329
330Instruction *AddressSanitizer::generateCrashCode(
331    IRBuilder<> &IRB, Value *Addr, bool IsWrite, uint32_t TypeSize) {
332
333  if (ClUseCall) {
334    // Here we use a call instead of arch-specific asm to report an error.
335    // This is almost always slower (because the codegen needs to generate
336    // prologue/epilogue for otherwise leaf functions) and generates more code.
337    // This mode could be useful if we can not use SIGILL for some reason.
338    //
339    // IsWrite and TypeSize are encoded in the function name.
340    std::string FunctionName = std::string(kAsanReportErrorTemplate) +
341        (IsWrite ? "store" : "load") + itostr(TypeSize / 8);
342    Value *ReportWarningFunc = CurrentModule->getOrInsertFunction(
343        FunctionName, IRB.getVoidTy(), IntptrTy, NULL);
344    CallInst *Call = IRB.CreateCall(ReportWarningFunc, Addr);
345    Call->setDoesNotReturn();
346    return Call;
347  }
348
349  uint32_t LogOfSizeInBytes = CountTrailingZeros_32(TypeSize / 8);
350  assert(8U * (1 << LogOfSizeInBytes) == TypeSize);
351  uint8_t TelltaleValue = IsWrite * 8 + LogOfSizeInBytes;
352  assert(TelltaleValue < 16);
353
354  // Move the failing address to %rax/%eax
355  FunctionType *Fn1Ty = FunctionType::get(
356      IRB.getVoidTy(), ArrayRef<Type*>(IntptrTy), false);
357  const char *MovStr = LongSize == 32
358      ? "mov $0, %eax" : "mov $0, %rax";
359  Value *AsmMov = InlineAsm::get(
360      Fn1Ty, StringRef(MovStr), StringRef("r"), true);
361  IRB.CreateCall(AsmMov, Addr);
362
363  // crash with ud2; could use int3, but it is less friendly to gdb.
364  // after ud2 put a 1-byte instruction that encodes the access type and size.
365
366  const char *TelltaleInsns[16] = {
367    "push   %eax",  // 0x50
368    "push   %ecx",  // 0x51
369    "push   %edx",  // 0x52
370    "push   %ebx",  // 0x53
371    "push   %esp",  // 0x54
372    "push   %ebp",  // 0x55
373    "push   %esi",  // 0x56
374    "push   %edi",  // 0x57
375    "pop    %eax",  // 0x58
376    "pop    %ecx",  // 0x59
377    "pop    %edx",  // 0x5a
378    "pop    %ebx",  // 0x5b
379    "pop    %esp",  // 0x5c
380    "pop    %ebp",  // 0x5d
381    "pop    %esi",  // 0x5e
382    "pop    %edi"   // 0x5f
383  };
384
385  std::string AsmStr = "ud2;";
386  AsmStr += TelltaleInsns[TelltaleValue];
387  Value *MyAsm = InlineAsm::get(FunctionType::get(Type::getVoidTy(*C), false),
388                                StringRef(AsmStr), StringRef(""), true);
389  CallInst *AsmCall = IRB.CreateCall(MyAsm);
390
391  // This saves us one jump, but triggers a bug in RA (or somewhere else):
392  // while building 483.xalancbmk the compiler goes into infinite loop in
393  // llvm::SpillPlacement::iterate() / RAGreedy::growRegion
394  // AsmCall->setDoesNotReturn();
395  return AsmCall;
396}
397
398void AddressSanitizer::instrumentAddress(Instruction *OrigIns,
399                                         IRBuilder<> &IRB, Value *Addr,
400                                         uint32_t TypeSize, bool IsWrite) {
401  Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
402
403  Type *ShadowTy  = IntegerType::get(
404      *C, std::max(8U, TypeSize >> MappingScale));
405  Type *ShadowPtrTy = PointerType::get(ShadowTy, 0);
406  Value *ShadowPtr = memToShadow(AddrLong, IRB);
407  Value *CmpVal = Constant::getNullValue(ShadowTy);
408  Value *ShadowValue = IRB.CreateLoad(
409      IRB.CreateIntToPtr(ShadowPtr, ShadowPtrTy));
410
411  Value *Cmp = IRB.CreateICmpNE(ShadowValue, CmpVal);
412
413  Instruction *CheckTerm = splitBlockAndInsertIfThen(
414      cast<Instruction>(Cmp)->getNextNode(), Cmp);
415  IRBuilder<> IRB2(CheckTerm);
416
417  size_t Granularity = 1 << MappingScale;
418  if (TypeSize < 8 * Granularity) {
419    // Addr & (Granularity - 1)
420    Value *Lower3Bits = IRB2.CreateAnd(
421        AddrLong, ConstantInt::get(IntptrTy, Granularity - 1));
422    // (Addr & (Granularity - 1)) + size - 1
423    Value *LastAccessedByte = IRB2.CreateAdd(
424        Lower3Bits, ConstantInt::get(IntptrTy, TypeSize / 8 - 1));
425    // (uint8_t) ((Addr & (Granularity-1)) + size - 1)
426    LastAccessedByte = IRB2.CreateIntCast(
427        LastAccessedByte, IRB.getInt8Ty(), false);
428    // ((uint8_t) ((Addr & (Granularity-1)) + size - 1)) >= ShadowValue
429    Value *Cmp2 = IRB2.CreateICmpSGE(LastAccessedByte, ShadowValue);
430
431    CheckTerm = splitBlockAndInsertIfThen(CheckTerm, Cmp2);
432  }
433
434  IRBuilder<> IRB1(CheckTerm);
435  Instruction *Crash = generateCrashCode(IRB1, AddrLong, IsWrite, TypeSize);
436  Crash->setDebugLoc(OrigIns->getDebugLoc());
437  ReplaceInstWithInst(CheckTerm, new UnreachableInst(*C));
438}
439
440// This function replaces all global variables with new variables that have
441// trailing redzones. It also creates a function that poisons
442// redzones and inserts this function into llvm.global_ctors.
443bool AddressSanitizer::insertGlobalRedzones(Module &M) {
444  SmallVector<GlobalVariable *, 16> GlobalsToChange;
445
446  for (Module::GlobalListType::iterator G = M.getGlobalList().begin(),
447       E = M.getGlobalList().end(); G != E; ++G) {
448    Type *Ty = cast<PointerType>(G->getType())->getElementType();
449    DEBUG(dbgs() << "GLOBAL: " << *G);
450
451    if (!Ty->isSized()) continue;
452    if (!G->hasInitializer()) continue;
453    // Touch only those globals that will not be defined in other modules.
454    // Don't handle ODR type linkages since other modules may be built w/o asan.
455    if (G->getLinkage() != GlobalVariable::ExternalLinkage &&
456        G->getLinkage() != GlobalVariable::PrivateLinkage &&
457        G->getLinkage() != GlobalVariable::InternalLinkage)
458      continue;
459    // Two problems with thread-locals:
460    //   - The address of the main thread's copy can't be computed at link-time.
461    //   - Need to poison all copies, not just the main thread's one.
462    if (G->isThreadLocal())
463      continue;
464    // For now, just ignore this Alloca if the alignment is large.
465    if (G->getAlignment() > RedzoneSize) continue;
466
467    // Ignore all the globals with the names starting with "\01L_OBJC_".
468    // Many of those are put into the .cstring section. The linker compresses
469    // that section by removing the spare \0s after the string terminator, so
470    // our redzones get broken.
471    if ((G->getName().find("\01L_OBJC_") == 0) ||
472        (G->getName().find("\01l_OBJC_") == 0)) {
473      DEBUG(dbgs() << "Ignoring \\01L_OBJC_* global: " << *G);
474      continue;
475    }
476
477    // Ignore the globals from the __OBJC section. The ObjC runtime assumes
478    // those conform to /usr/lib/objc/runtime.h, so we can't add redzones to
479    // them.
480    if (G->hasSection()) {
481      StringRef Section(G->getSection());
482      if ((Section.find("__OBJC,") == 0) ||
483          (Section.find("__DATA, __objc_") == 0)) {
484        DEBUG(dbgs() << "Ignoring ObjC runtime global: " << *G);
485        continue;
486      }
487    }
488
489    GlobalsToChange.push_back(G);
490  }
491
492  size_t n = GlobalsToChange.size();
493  if (n == 0) return false;
494
495  // A global is described by a structure
496  //   size_t beg;
497  //   size_t size;
498  //   size_t size_with_redzone;
499  //   const char *name;
500  // We initialize an array of such structures and pass it to a run-time call.
501  StructType *GlobalStructTy = StructType::get(IntptrTy, IntptrTy,
502                                               IntptrTy, IntptrTy, NULL);
503  SmallVector<Constant *, 16> Initializers(n);
504
505  IRBuilder<> IRB(CtorInsertBefore);
506
507  for (size_t i = 0; i < n; i++) {
508    GlobalVariable *G = GlobalsToChange[i];
509    PointerType *PtrTy = cast<PointerType>(G->getType());
510    Type *Ty = PtrTy->getElementType();
511    uint64_t SizeInBytes = TD->getTypeStoreSizeInBits(Ty) / 8;
512    uint64_t RightRedzoneSize = RedzoneSize +
513        (RedzoneSize - (SizeInBytes % RedzoneSize));
514    Type *RightRedZoneTy = ArrayType::get(IRB.getInt8Ty(), RightRedzoneSize);
515
516    StructType *NewTy = StructType::get(Ty, RightRedZoneTy, NULL);
517    Constant *NewInitializer = ConstantStruct::get(
518        NewTy, G->getInitializer(),
519        Constant::getNullValue(RightRedZoneTy), NULL);
520
521    GlobalVariable *Name = createPrivateGlobalForString(M, G->getName());
522
523    // Create a new global variable with enough space for a redzone.
524    GlobalVariable *NewGlobal = new GlobalVariable(
525        M, NewTy, G->isConstant(), G->getLinkage(),
526        NewInitializer, "", G, G->isThreadLocal());
527    NewGlobal->copyAttributesFrom(G);
528    NewGlobal->setAlignment(RedzoneSize);
529
530    Value *Indices2[2];
531    Indices2[0] = IRB.getInt32(0);
532    Indices2[1] = IRB.getInt32(0);
533
534    G->replaceAllUsesWith(
535        ConstantExpr::getGetElementPtr(NewGlobal, Indices2, 2));
536    NewGlobal->takeName(G);
537    G->eraseFromParent();
538
539    Initializers[i] = ConstantStruct::get(
540        GlobalStructTy,
541        ConstantExpr::getPointerCast(NewGlobal, IntptrTy),
542        ConstantInt::get(IntptrTy, SizeInBytes),
543        ConstantInt::get(IntptrTy, SizeInBytes + RightRedzoneSize),
544        ConstantExpr::getPointerCast(Name, IntptrTy),
545        NULL);
546    DEBUG(dbgs() << "NEW GLOBAL:\n" << *NewGlobal);
547  }
548
549  ArrayType *ArrayOfGlobalStructTy = ArrayType::get(GlobalStructTy, n);
550  GlobalVariable *AllGlobals = new GlobalVariable(
551      M, ArrayOfGlobalStructTy, false, GlobalVariable::PrivateLinkage,
552      ConstantArray::get(ArrayOfGlobalStructTy, Initializers), "");
553
554  Function *AsanRegisterGlobals = cast<Function>(M.getOrInsertFunction(
555      kAsanRegisterGlobalsName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
556  AsanRegisterGlobals->setLinkage(Function::ExternalLinkage);
557
558  IRB.CreateCall2(AsanRegisterGlobals,
559                  IRB.CreatePointerCast(AllGlobals, IntptrTy),
560                  ConstantInt::get(IntptrTy, n));
561
562  DEBUG(dbgs() << M);
563  return true;
564}
565
566// virtual
567bool AddressSanitizer::runOnModule(Module &M) {
568  // Initialize the private fields. No one has accessed them before.
569  TD = getAnalysisIfAvailable<TargetData>();
570  if (!TD)
571    return false;
572  BL.reset(new BlackList(ClBlackListFile));
573
574  CurrentModule = &M;
575  C = &(M.getContext());
576  LongSize = TD->getPointerSizeInBits();
577  IntptrTy = Type::getIntNTy(*C, LongSize);
578  IntptrPtrTy = PointerType::get(IntptrTy, 0);
579
580  AsanCtorFunction = Function::Create(
581      FunctionType::get(Type::getVoidTy(*C), false),
582      GlobalValue::InternalLinkage, kAsanModuleCtorName, &M);
583  BasicBlock *AsanCtorBB = BasicBlock::Create(*C, "", AsanCtorFunction);
584  CtorInsertBefore = ReturnInst::Create(*C, AsanCtorBB);
585
586  // call __asan_init in the module ctor.
587  IRBuilder<> IRB(CtorInsertBefore);
588  AsanInitFunction = cast<Function>(
589      M.getOrInsertFunction(kAsanInitName, IRB.getVoidTy(), NULL));
590  AsanInitFunction->setLinkage(Function::ExternalLinkage);
591  IRB.CreateCall(AsanInitFunction);
592
593  MappingOffset = LongSize == 32
594      ? kDefaultShadowOffset32 : kDefaultShadowOffset64;
595  if (ClMappingOffsetLog >= 0) {
596    if (ClMappingOffsetLog == 0) {
597      // special case
598      MappingOffset = 0;
599    } else {
600      MappingOffset = 1ULL << ClMappingOffsetLog;
601    }
602  }
603  MappingScale = kDefaultShadowScale;
604  if (ClMappingScale) {
605    MappingScale = ClMappingScale;
606  }
607  // Redzone used for stack and globals is at least 32 bytes.
608  // For scales 6 and 7, the redzone has to be 64 and 128 bytes respectively.
609  RedzoneSize = std::max(32, (int)(1 << MappingScale));
610
611  bool Res = false;
612
613  if (ClGlobals)
614    Res |= insertGlobalRedzones(M);
615
616  // Tell the run-time the current values of mapping offset and scale.
617  GlobalValue *asan_mapping_offset =
618      new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
619                     ConstantInt::get(IntptrTy, MappingOffset),
620                     kAsanMappingOffsetName);
621  GlobalValue *asan_mapping_scale =
622      new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
623                         ConstantInt::get(IntptrTy, MappingScale),
624                         kAsanMappingScaleName);
625  // Read these globals, otherwise they may be optimized away.
626  IRB.CreateLoad(asan_mapping_scale, true);
627  IRB.CreateLoad(asan_mapping_offset, true);
628
629
630  for (Module::iterator F = M.begin(), E = M.end(); F != E; ++F) {
631    if (F->isDeclaration()) continue;
632    Res |= handleFunction(M, *F);
633  }
634
635  appendToGlobalCtors(M, AsanCtorFunction, 1 /*high priority*/);
636
637  return Res;
638}
639
640bool AddressSanitizer::handleFunction(Module &M, Function &F) {
641  if (BL->isIn(F)) return false;
642  if (&F == AsanCtorFunction) return false;
643
644  if (!ClDebugFunc.empty() && ClDebugFunc != F.getName())
645    return false;
646  // We want to instrument every address only once per basic block
647  // (unless there are calls between uses).
648  SmallSet<Value*, 16> TempsToInstrument;
649  SmallVector<Instruction*, 16> ToInstrument;
650
651  // Fill the set of memory operations to instrument.
652  for (Function::iterator FI = F.begin(), FE = F.end();
653       FI != FE; ++FI) {
654    TempsToInstrument.clear();
655    for (BasicBlock::iterator BI = FI->begin(), BE = FI->end();
656         BI != BE; ++BI) {
657      if ((isa<LoadInst>(BI) && ClInstrumentReads) ||
658          (isa<StoreInst>(BI) && ClInstrumentWrites)) {
659        Value *Addr = getLDSTOperand(BI);
660        if (ClOpt && ClOptSameTemp) {
661          if (!TempsToInstrument.insert(Addr))
662            continue;  // We've seen this temp in the current BB.
663        }
664      } else if (isa<MemIntrinsic>(BI) && ClMemIntrin) {
665        // ok, take it.
666      } else {
667        if (isa<CallInst>(BI)) {
668          // A call inside BB.
669          TempsToInstrument.clear();
670        }
671        continue;
672      }
673      ToInstrument.push_back(BI);
674    }
675  }
676
677  // Instrument.
678  int NumInstrumented = 0;
679  for (size_t i = 0, n = ToInstrument.size(); i != n; i++) {
680    Instruction *Inst = ToInstrument[i];
681    if (ClDebugMin < 0 || ClDebugMax < 0 ||
682        (NumInstrumented >= ClDebugMin && NumInstrumented <= ClDebugMax)) {
683      if (isa<StoreInst>(Inst) || isa<LoadInst>(Inst))
684        instrumentMop(Inst);
685      else
686        instrumentMemIntrinsic(cast<MemIntrinsic>(Inst));
687    }
688    NumInstrumented++;
689  }
690
691  DEBUG(dbgs() << F);
692
693  bool ChangedStack = poisonStackInFunction(M, F);
694
695  // For each NSObject descendant having a +load method, this method is invoked
696  // by the ObjC runtime before any of the static constructors is called.
697  // Therefore we need to instrument such methods with a call to __asan_init
698  // at the beginning in order to initialize our runtime before any access to
699  // the shadow memory.
700  // We cannot just ignore these methods, because they may call other
701  // instrumented functions.
702  if (F.getName().find(" load]") != std::string::npos) {
703    IRBuilder<> IRB(F.begin()->begin());
704    IRB.CreateCall(AsanInitFunction);
705  }
706
707  return NumInstrumented > 0 || ChangedStack;
708}
709
710static uint64_t ValueForPoison(uint64_t PoisonByte, size_t ShadowRedzoneSize) {
711  if (ShadowRedzoneSize == 1) return PoisonByte;
712  if (ShadowRedzoneSize == 2) return (PoisonByte << 8) + PoisonByte;
713  if (ShadowRedzoneSize == 4)
714    return (PoisonByte << 24) + (PoisonByte << 16) +
715        (PoisonByte << 8) + (PoisonByte);
716  assert(0 && "ShadowRedzoneSize is either 1, 2 or 4");
717  return 0;
718}
719
720static void PoisonShadowPartialRightRedzone(uint8_t *Shadow,
721                                            size_t Size,
722                                            size_t RedzoneSize,
723                                            size_t ShadowGranularity,
724                                            uint8_t Magic) {
725  for (size_t i = 0; i < RedzoneSize;
726       i+= ShadowGranularity, Shadow++) {
727    if (i + ShadowGranularity <= Size) {
728      *Shadow = 0;  // fully addressable
729    } else if (i >= Size) {
730      *Shadow = Magic;  // unaddressable
731    } else {
732      *Shadow = Size - i;  // first Size-i bytes are addressable
733    }
734  }
735}
736
737void AddressSanitizer::PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec,
738                                   IRBuilder<> IRB,
739                                   Value *ShadowBase, bool DoPoison) {
740  size_t ShadowRZSize = RedzoneSize >> MappingScale;
741  assert(ShadowRZSize >= 1 && ShadowRZSize <= 4);
742  Type *RZTy = Type::getIntNTy(*C, ShadowRZSize * 8);
743  Type *RZPtrTy = PointerType::get(RZTy, 0);
744
745  Value *PoisonLeft  = ConstantInt::get(RZTy,
746    ValueForPoison(DoPoison ? kAsanStackLeftRedzoneMagic : 0LL, ShadowRZSize));
747  Value *PoisonMid   = ConstantInt::get(RZTy,
748    ValueForPoison(DoPoison ? kAsanStackMidRedzoneMagic : 0LL, ShadowRZSize));
749  Value *PoisonRight = ConstantInt::get(RZTy,
750    ValueForPoison(DoPoison ? kAsanStackRightRedzoneMagic : 0LL, ShadowRZSize));
751
752  // poison the first red zone.
753  IRB.CreateStore(PoisonLeft, IRB.CreateIntToPtr(ShadowBase, RZPtrTy));
754
755  // poison all other red zones.
756  uint64_t Pos = RedzoneSize;
757  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
758    AllocaInst *AI = AllocaVec[i];
759    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
760    uint64_t AlignedSize = getAlignedAllocaSize(AI);
761    assert(AlignedSize - SizeInBytes < RedzoneSize);
762    Value *Ptr = NULL;
763
764    Pos += AlignedSize;
765
766    assert(ShadowBase->getType() == IntptrTy);
767    if (SizeInBytes < AlignedSize) {
768      // Poison the partial redzone at right
769      Ptr = IRB.CreateAdd(
770          ShadowBase, ConstantInt::get(IntptrTy,
771                                       (Pos >> MappingScale) - ShadowRZSize));
772      size_t AddressableBytes = RedzoneSize - (AlignedSize - SizeInBytes);
773      uint32_t Poison = 0;
774      if (DoPoison) {
775        PoisonShadowPartialRightRedzone((uint8_t*)&Poison, AddressableBytes,
776                                        RedzoneSize,
777                                        1ULL << MappingScale,
778                                        kAsanStackPartialRedzoneMagic);
779      }
780      Value *PartialPoison = ConstantInt::get(RZTy, Poison);
781      IRB.CreateStore(PartialPoison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
782    }
783
784    // Poison the full redzone at right.
785    Ptr = IRB.CreateAdd(ShadowBase,
786                        ConstantInt::get(IntptrTy, Pos >> MappingScale));
787    Value *Poison = i == AllocaVec.size() - 1 ? PoisonRight : PoisonMid;
788    IRB.CreateStore(Poison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
789
790    Pos += RedzoneSize;
791  }
792}
793
794// Workaround for bug 11395: we don't want to instrument stack in functions
795// with large assembly blobs (32-bit only), otherwise reg alloc may crash.
796// FIXME: remove once the bug 11395 is fixed.
797bool AddressSanitizer::LooksLikeCodeInBug11395(Instruction *I) {
798  if (LongSize != 32) return false;
799  CallInst *CI = dyn_cast<CallInst>(I);
800  if (!CI || !CI->isInlineAsm()) return false;
801  if (CI->getNumArgOperands() <= 5) return false;
802  // We have inline assembly with quite a few arguments.
803  return true;
804}
805
806// Find all static Alloca instructions and put
807// poisoned red zones around all of them.
808// Then unpoison everything back before the function returns.
809//
810// Stack poisoning does not play well with exception handling.
811// When an exception is thrown, we essentially bypass the code
812// that unpoisones the stack. This is why the run-time library has
813// to intercept __cxa_throw (as well as longjmp, etc) and unpoison the entire
814// stack in the interceptor. This however does not work inside the
815// actual function which catches the exception. Most likely because the
816// compiler hoists the load of the shadow value somewhere too high.
817// This causes asan to report a non-existing bug on 453.povray.
818// It sounds like an LLVM bug.
819bool AddressSanitizer::poisonStackInFunction(Module &M, Function &F) {
820  if (!ClStack) return false;
821  SmallVector<AllocaInst*, 16> AllocaVec;
822  SmallVector<Instruction*, 8> RetVec;
823  uint64_t TotalSize = 0;
824
825  // Filter out Alloca instructions we want (and can) handle.
826  // Collect Ret instructions.
827  for (Function::iterator FI = F.begin(), FE = F.end();
828       FI != FE; ++FI) {
829    BasicBlock &BB = *FI;
830    for (BasicBlock::iterator BI = BB.begin(), BE = BB.end();
831         BI != BE; ++BI) {
832      if (LooksLikeCodeInBug11395(BI)) return false;
833      if (isa<ReturnInst>(BI)) {
834          RetVec.push_back(BI);
835          continue;
836      }
837
838      AllocaInst *AI = dyn_cast<AllocaInst>(BI);
839      if (!AI) continue;
840      if (AI->isArrayAllocation()) continue;
841      if (!AI->isStaticAlloca()) continue;
842      if (!AI->getAllocatedType()->isSized()) continue;
843      if (AI->getAlignment() > RedzoneSize) continue;
844      AllocaVec.push_back(AI);
845      uint64_t AlignedSize =  getAlignedAllocaSize(AI);
846      TotalSize += AlignedSize;
847    }
848  }
849
850  if (AllocaVec.empty()) return false;
851
852  uint64_t LocalStackSize = TotalSize + (AllocaVec.size() + 1) * RedzoneSize;
853
854  bool DoStackMalloc = ClUseAfterReturn
855      && LocalStackSize <= kMaxStackMallocSize;
856
857  Instruction *InsBefore = AllocaVec[0];
858  IRBuilder<> IRB(InsBefore);
859
860
861  Type *ByteArrayTy = ArrayType::get(IRB.getInt8Ty(), LocalStackSize);
862  AllocaInst *MyAlloca =
863      new AllocaInst(ByteArrayTy, "MyAlloca", InsBefore);
864  MyAlloca->setAlignment(RedzoneSize);
865  assert(MyAlloca->isStaticAlloca());
866  Value *OrigStackBase = IRB.CreatePointerCast(MyAlloca, IntptrTy);
867  Value *LocalStackBase = OrigStackBase;
868
869  if (DoStackMalloc) {
870    Value *AsanStackMallocFunc = M.getOrInsertFunction(
871        kAsanStackMallocName, IntptrTy, IntptrTy, IntptrTy, NULL);
872    LocalStackBase = IRB.CreateCall2(AsanStackMallocFunc,
873        ConstantInt::get(IntptrTy, LocalStackSize), OrigStackBase);
874  }
875
876  // This string will be parsed by the run-time (DescribeStackAddress).
877  SmallString<2048> StackDescriptionStorage;
878  raw_svector_ostream StackDescription(StackDescriptionStorage);
879  StackDescription << F.getName() << " " << AllocaVec.size() << " ";
880
881  uint64_t Pos = RedzoneSize;
882  // Replace Alloca instructions with base+offset.
883  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
884    AllocaInst *AI = AllocaVec[i];
885    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
886    StringRef Name = AI->getName();
887    StackDescription << Pos << " " << SizeInBytes << " "
888                     << Name.size() << " " << Name << " ";
889    uint64_t AlignedSize = getAlignedAllocaSize(AI);
890    assert((AlignedSize % RedzoneSize) == 0);
891    AI->replaceAllUsesWith(
892        IRB.CreateIntToPtr(
893            IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy, Pos)),
894            AI->getType()));
895    Pos += AlignedSize + RedzoneSize;
896  }
897  assert(Pos == LocalStackSize);
898
899  // Write the Magic value and the frame description constant to the redzone.
900  Value *BasePlus0 = IRB.CreateIntToPtr(LocalStackBase, IntptrPtrTy);
901  IRB.CreateStore(ConstantInt::get(IntptrTy, kCurrentStackFrameMagic),
902                  BasePlus0);
903  Value *BasePlus1 = IRB.CreateAdd(LocalStackBase,
904                                   ConstantInt::get(IntptrTy, LongSize/8));
905  BasePlus1 = IRB.CreateIntToPtr(BasePlus1, IntptrPtrTy);
906  Value *Description = IRB.CreatePointerCast(
907      createPrivateGlobalForString(M, StackDescription.str()),
908      IntptrTy);
909  IRB.CreateStore(Description, BasePlus1);
910
911  // Poison the stack redzones at the entry.
912  Value *ShadowBase = memToShadow(LocalStackBase, IRB);
913  PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRB, ShadowBase, true);
914
915  Value *AsanStackFreeFunc = NULL;
916  if (DoStackMalloc) {
917    AsanStackFreeFunc = M.getOrInsertFunction(
918        kAsanStackFreeName, IRB.getVoidTy(),
919        IntptrTy, IntptrTy, IntptrTy, NULL);
920  }
921
922  // Unpoison the stack before all ret instructions.
923  for (size_t i = 0, n = RetVec.size(); i < n; i++) {
924    Instruction *Ret = RetVec[i];
925    IRBuilder<> IRBRet(Ret);
926
927    // Mark the current frame as retired.
928    IRBRet.CreateStore(ConstantInt::get(IntptrTy, kRetiredStackFrameMagic),
929                       BasePlus0);
930    // Unpoison the stack.
931    PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRBRet, ShadowBase, false);
932
933    if (DoStackMalloc) {
934      IRBRet.CreateCall3(AsanStackFreeFunc, LocalStackBase,
935                         ConstantInt::get(IntptrTy, LocalStackSize),
936                         OrigStackBase);
937    }
938  }
939
940  if (ClDebugStack) {
941    DEBUG(dbgs() << F);
942  }
943
944  return true;
945}
946
947BlackList::BlackList(const std::string &Path) {
948  Functions = NULL;
949  const char *kFunPrefix = "fun:";
950  if (!ClBlackListFile.size()) return;
951  std::string Fun;
952
953  OwningPtr<MemoryBuffer> File;
954  if (error_code EC = MemoryBuffer::getFile(ClBlackListFile.c_str(), File)) {
955    report_fatal_error("Can't open blacklist file " + ClBlackListFile + ": " +
956                       EC.message());
957  }
958  MemoryBuffer *Buff = File.take();
959  const char *Data = Buff->getBufferStart();
960  size_t DataLen = Buff->getBufferSize();
961  SmallVector<StringRef, 16> Lines;
962  SplitString(StringRef(Data, DataLen), Lines, "\n\r");
963  for (size_t i = 0, numLines = Lines.size(); i < numLines; i++) {
964    if (Lines[i].startswith(kFunPrefix)) {
965      std::string ThisFunc = Lines[i].substr(strlen(kFunPrefix));
966      if (Fun.size()) {
967        Fun += "|";
968      }
969      // add ThisFunc replacing * with .*
970      for (size_t j = 0, n = ThisFunc.size(); j < n; j++) {
971        if (ThisFunc[j] == '*')
972          Fun += '.';
973        Fun += ThisFunc[j];
974      }
975    }
976  }
977  if (Fun.size()) {
978    Functions = new Regex(Fun);
979  }
980}
981
982bool BlackList::isIn(const Function &F) {
983  if (Functions) {
984    bool Res = Functions->match(F.getName());
985    return Res;
986  }
987  return false;
988}
989