AddressSanitizer.cpp revision d2703dec271d82c8c9d22afb835c07730fd25d47 
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}
438
439// This function replaces all global variables with new variables that have
440// trailing redzones. It also creates a function that poisons
441// redzones and inserts this function into llvm.global_ctors.
442bool AddressSanitizer::insertGlobalRedzones(Module &M) {
443  SmallVector<GlobalVariable *, 16> GlobalsToChange;
444
445  for (Module::GlobalListType::iterator G = M.getGlobalList().begin(),
446       E = M.getGlobalList().end(); G != E; ++G) {
447    Type *Ty = cast<PointerType>(G->getType())->getElementType();
448    DEBUG(dbgs() << "GLOBAL: " << *G);
449
450    if (!Ty->isSized()) continue;
451    if (!G->hasInitializer()) continue;
452    // Touch only those globals that will not be defined in other modules.
453    // Don't handle ODR type linkages since other modules may be built w/o asan.
454    if (G->getLinkage() != GlobalVariable::ExternalLinkage &&
455        G->getLinkage() != GlobalVariable::PrivateLinkage &&
456        G->getLinkage() != GlobalVariable::InternalLinkage)
457      continue;
458    // Two problems with thread-locals:
459    //   - The address of the main thread's copy can't be computed at link-time.
460    //   - Need to poison all copies, not just the main thread's one.
461    if (G->isThreadLocal())
462      continue;
463    // For now, just ignore this Alloca if the alignment is large.
464    if (G->getAlignment() > RedzoneSize) continue;
465
466    // Ignore all the globals with the names starting with "\01L_OBJC_".
467    // Many of those are put into the .cstring section. The linker compresses
468    // that section by removing the spare \0s after the string terminator, so
469    // our redzones get broken.
470    if ((G->getName().find("\01L_OBJC_") == 0) ||
471        (G->getName().find("\01l_OBJC_") == 0)) {
472      DEBUG(dbgs() << "Ignoring \\01L_OBJC_* global: " << *G);
473      continue;
474    }
475
476    // Ignore the globals from the __OBJC section. The ObjC runtime assumes
477    // those conform to /usr/lib/objc/runtime.h, so we can't add redzones to
478    // them.
479    if (G->hasSection()) {
480      StringRef Section(G->getSection());
481      if ((Section.find("__OBJC,") == 0) ||
482          (Section.find("__DATA, __objc_") == 0)) {
483        DEBUG(dbgs() << "Ignoring ObjC runtime global: " << *G);
484        continue;
485      }
486    }
487
488    GlobalsToChange.push_back(G);
489  }
490
491  size_t n = GlobalsToChange.size();
492  if (n == 0) return false;
493
494  // A global is described by a structure
495  //   size_t beg;
496  //   size_t size;
497  //   size_t size_with_redzone;
498  //   const char *name;
499  // We initialize an array of such structures and pass it to a run-time call.
500  StructType *GlobalStructTy = StructType::get(IntptrTy, IntptrTy,
501                                               IntptrTy, IntptrTy, NULL);
502  SmallVector<Constant *, 16> Initializers(n);
503
504  IRBuilder<> IRB(CtorInsertBefore);
505
506  for (size_t i = 0; i < n; i++) {
507    GlobalVariable *G = GlobalsToChange[i];
508    PointerType *PtrTy = cast<PointerType>(G->getType());
509    Type *Ty = PtrTy->getElementType();
510    uint64_t SizeInBytes = TD->getTypeStoreSizeInBits(Ty) / 8;
511    uint64_t RightRedzoneSize = RedzoneSize +
512        (RedzoneSize - (SizeInBytes % RedzoneSize));
513    Type *RightRedZoneTy = ArrayType::get(IRB.getInt8Ty(), RightRedzoneSize);
514
515    StructType *NewTy = StructType::get(Ty, RightRedZoneTy, NULL);
516    Constant *NewInitializer = ConstantStruct::get(
517        NewTy, G->getInitializer(),
518        Constant::getNullValue(RightRedZoneTy), NULL);
519
520    GlobalVariable *Name = createPrivateGlobalForString(M, G->getName());
521
522    // Create a new global variable with enough space for a redzone.
523    GlobalVariable *NewGlobal = new GlobalVariable(
524        M, NewTy, G->isConstant(), G->getLinkage(),
525        NewInitializer, "", G, G->isThreadLocal());
526    NewGlobal->copyAttributesFrom(G);
527    NewGlobal->setAlignment(RedzoneSize);
528
529    Value *Indices2[2];
530    Indices2[0] = IRB.getInt32(0);
531    Indices2[1] = IRB.getInt32(0);
532
533    G->replaceAllUsesWith(
534        ConstantExpr::getGetElementPtr(NewGlobal, Indices2, 2));
535    NewGlobal->takeName(G);
536    G->eraseFromParent();
537
538    Initializers[i] = ConstantStruct::get(
539        GlobalStructTy,
540        ConstantExpr::getPointerCast(NewGlobal, IntptrTy),
541        ConstantInt::get(IntptrTy, SizeInBytes),
542        ConstantInt::get(IntptrTy, SizeInBytes + RightRedzoneSize),
543        ConstantExpr::getPointerCast(Name, IntptrTy),
544        NULL);
545    DEBUG(dbgs() << "NEW GLOBAL:\n" << *NewGlobal);
546  }
547
548  ArrayType *ArrayOfGlobalStructTy = ArrayType::get(GlobalStructTy, n);
549  GlobalVariable *AllGlobals = new GlobalVariable(
550      M, ArrayOfGlobalStructTy, false, GlobalVariable::PrivateLinkage,
551      ConstantArray::get(ArrayOfGlobalStructTy, Initializers), "");
552
553  Function *AsanRegisterGlobals = cast<Function>(M.getOrInsertFunction(
554      kAsanRegisterGlobalsName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
555  AsanRegisterGlobals->setLinkage(Function::ExternalLinkage);
556
557  IRB.CreateCall2(AsanRegisterGlobals,
558                  IRB.CreatePointerCast(AllGlobals, IntptrTy),
559                  ConstantInt::get(IntptrTy, n));
560
561  DEBUG(dbgs() << M);
562  return true;
563}
564
565// virtual
566bool AddressSanitizer::runOnModule(Module &M) {
567  // Initialize the private fields. No one has accessed them before.
568  TD = getAnalysisIfAvailable<TargetData>();
569  if (!TD)
570    return false;
571  BL.reset(new BlackList(ClBlackListFile));
572
573  CurrentModule = &M;
574  C = &(M.getContext());
575  LongSize = TD->getPointerSizeInBits();
576  IntptrTy = Type::getIntNTy(*C, LongSize);
577  IntptrPtrTy = PointerType::get(IntptrTy, 0);
578
579  AsanCtorFunction = Function::Create(
580      FunctionType::get(Type::getVoidTy(*C), false),
581      GlobalValue::InternalLinkage, kAsanModuleCtorName, &M);
582  BasicBlock *AsanCtorBB = BasicBlock::Create(*C, "", AsanCtorFunction);
583  CtorInsertBefore = ReturnInst::Create(*C, AsanCtorBB);
584
585  // call __asan_init in the module ctor.
586  IRBuilder<> IRB(CtorInsertBefore);
587  AsanInitFunction = cast<Function>(
588      M.getOrInsertFunction(kAsanInitName, IRB.getVoidTy(), NULL));
589  AsanInitFunction->setLinkage(Function::ExternalLinkage);
590  IRB.CreateCall(AsanInitFunction);
591
592  MappingOffset = LongSize == 32
593      ? kDefaultShadowOffset32 : kDefaultShadowOffset64;
594  if (ClMappingOffsetLog >= 0) {
595    if (ClMappingOffsetLog == 0) {
596      // special case
597      MappingOffset = 0;
598    } else {
599      MappingOffset = 1ULL << ClMappingOffsetLog;
600    }
601  }
602  MappingScale = kDefaultShadowScale;
603  if (ClMappingScale) {
604    MappingScale = ClMappingScale;
605  }
606  // Redzone used for stack and globals is at least 32 bytes.
607  // For scales 6 and 7, the redzone has to be 64 and 128 bytes respectively.
608  RedzoneSize = std::max(32, (int)(1 << MappingScale));
609
610  bool Res = false;
611
612  if (ClGlobals)
613    Res |= insertGlobalRedzones(M);
614
615  // Tell the run-time the current values of mapping offset and scale.
616  GlobalValue *asan_mapping_offset =
617      new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
618                     ConstantInt::get(IntptrTy, MappingOffset),
619                     kAsanMappingOffsetName);
620  GlobalValue *asan_mapping_scale =
621      new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
622                         ConstantInt::get(IntptrTy, MappingScale),
623                         kAsanMappingScaleName);
624  // Read these globals, otherwise they may be optimized away.
625  IRB.CreateLoad(asan_mapping_scale, true);
626  IRB.CreateLoad(asan_mapping_offset, true);
627
628
629  for (Module::iterator F = M.begin(), E = M.end(); F != E; ++F) {
630    if (F->isDeclaration()) continue;
631    Res |= handleFunction(M, *F);
632  }
633
634  appendToGlobalCtors(M, AsanCtorFunction, 1 /*high priority*/);
635
636  return Res;
637}
638
639bool AddressSanitizer::handleFunction(Module &M, Function &F) {
640  if (BL->isIn(F)) return false;
641  if (&F == AsanCtorFunction) return false;
642
643  if (!ClDebugFunc.empty() && ClDebugFunc != F.getName())
644    return false;
645  // We want to instrument every address only once per basic block
646  // (unless there are calls between uses).
647  SmallSet<Value*, 16> TempsToInstrument;
648  SmallVector<Instruction*, 16> ToInstrument;
649
650  // Fill the set of memory operations to instrument.
651  for (Function::iterator FI = F.begin(), FE = F.end();
652       FI != FE; ++FI) {
653    TempsToInstrument.clear();
654    for (BasicBlock::iterator BI = FI->begin(), BE = FI->end();
655         BI != BE; ++BI) {
656      if ((isa<LoadInst>(BI) && ClInstrumentReads) ||
657          (isa<StoreInst>(BI) && ClInstrumentWrites)) {
658        Value *Addr = getLDSTOperand(BI);
659        if (ClOpt && ClOptSameTemp) {
660          if (!TempsToInstrument.insert(Addr))
661            continue;  // We've seen this temp in the current BB.
662        }
663      } else if (isa<MemIntrinsic>(BI) && ClMemIntrin) {
664        // ok, take it.
665      } else {
666        if (isa<CallInst>(BI)) {
667          // A call inside BB.
668          TempsToInstrument.clear();
669        }
670        continue;
671      }
672      ToInstrument.push_back(BI);
673    }
674  }
675
676  // Instrument.
677  int NumInstrumented = 0;
678  for (size_t i = 0, n = ToInstrument.size(); i != n; i++) {
679    Instruction *Inst = ToInstrument[i];
680    if (ClDebugMin < 0 || ClDebugMax < 0 ||
681        (NumInstrumented >= ClDebugMin && NumInstrumented <= ClDebugMax)) {
682      if (isa<StoreInst>(Inst) || isa<LoadInst>(Inst))
683        instrumentMop(Inst);
684      else
685        instrumentMemIntrinsic(cast<MemIntrinsic>(Inst));
686    }
687    NumInstrumented++;
688  }
689
690  DEBUG(dbgs() << F);
691
692  bool ChangedStack = poisonStackInFunction(M, F);
693
694  // For each NSObject descendant having a +load method, this method is invoked
695  // by the ObjC runtime before any of the static constructors is called.
696  // Therefore we need to instrument such methods with a call to __asan_init
697  // at the beginning in order to initialize our runtime before any access to
698  // the shadow memory.
699  // We cannot just ignore these methods, because they may call other
700  // instrumented functions.
701  if (F.getName().find(" load]") != std::string::npos) {
702    IRBuilder<> IRB(F.begin()->begin());
703    IRB.CreateCall(AsanInitFunction);
704  }
705
706  return NumInstrumented > 0 || ChangedStack;
707}
708
709static uint64_t ValueForPoison(uint64_t PoisonByte, size_t ShadowRedzoneSize) {
710  if (ShadowRedzoneSize == 1) return PoisonByte;
711  if (ShadowRedzoneSize == 2) return (PoisonByte << 8) + PoisonByte;
712  if (ShadowRedzoneSize == 4)
713    return (PoisonByte << 24) + (PoisonByte << 16) +
714        (PoisonByte << 8) + (PoisonByte);
715  assert(0 && "ShadowRedzoneSize is either 1, 2 or 4");
716  return 0;
717}
718
719static void PoisonShadowPartialRightRedzone(uint8_t *Shadow,
720                                            size_t Size,
721                                            size_t RedzoneSize,
722                                            size_t ShadowGranularity,
723                                            uint8_t Magic) {
724  for (size_t i = 0; i < RedzoneSize;
725       i+= ShadowGranularity, Shadow++) {
726    if (i + ShadowGranularity <= Size) {
727      *Shadow = 0;  // fully addressable
728    } else if (i >= Size) {
729      *Shadow = Magic;  // unaddressable
730    } else {
731      *Shadow = Size - i;  // first Size-i bytes are addressable
732    }
733  }
734}
735
736void AddressSanitizer::PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec,
737                                   IRBuilder<> IRB,
738                                   Value *ShadowBase, bool DoPoison) {
739  size_t ShadowRZSize = RedzoneSize >> MappingScale;
740  assert(ShadowRZSize >= 1 && ShadowRZSize <= 4);
741  Type *RZTy = Type::getIntNTy(*C, ShadowRZSize * 8);
742  Type *RZPtrTy = PointerType::get(RZTy, 0);
743
744  Value *PoisonLeft  = ConstantInt::get(RZTy,
745    ValueForPoison(DoPoison ? kAsanStackLeftRedzoneMagic : 0LL, ShadowRZSize));
746  Value *PoisonMid   = ConstantInt::get(RZTy,
747    ValueForPoison(DoPoison ? kAsanStackMidRedzoneMagic : 0LL, ShadowRZSize));
748  Value *PoisonRight = ConstantInt::get(RZTy,
749    ValueForPoison(DoPoison ? kAsanStackRightRedzoneMagic : 0LL, ShadowRZSize));
750
751  // poison the first red zone.
752  IRB.CreateStore(PoisonLeft, IRB.CreateIntToPtr(ShadowBase, RZPtrTy));
753
754  // poison all other red zones.
755  uint64_t Pos = RedzoneSize;
756  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
757    AllocaInst *AI = AllocaVec[i];
758    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
759    uint64_t AlignedSize = getAlignedAllocaSize(AI);
760    assert(AlignedSize - SizeInBytes < RedzoneSize);
761    Value *Ptr = NULL;
762
763    Pos += AlignedSize;
764
765    assert(ShadowBase->getType() == IntptrTy);
766    if (SizeInBytes < AlignedSize) {
767      // Poison the partial redzone at right
768      Ptr = IRB.CreateAdd(
769          ShadowBase, ConstantInt::get(IntptrTy,
770                                       (Pos >> MappingScale) - ShadowRZSize));
771      size_t AddressableBytes = RedzoneSize - (AlignedSize - SizeInBytes);
772      uint32_t Poison = 0;
773      if (DoPoison) {
774        PoisonShadowPartialRightRedzone((uint8_t*)&Poison, AddressableBytes,
775                                        RedzoneSize,
776                                        1ULL << MappingScale,
777                                        kAsanStackPartialRedzoneMagic);
778      }
779      Value *PartialPoison = ConstantInt::get(RZTy, Poison);
780      IRB.CreateStore(PartialPoison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
781    }
782
783    // Poison the full redzone at right.
784    Ptr = IRB.CreateAdd(ShadowBase,
785                        ConstantInt::get(IntptrTy, Pos >> MappingScale));
786    Value *Poison = i == AllocaVec.size() - 1 ? PoisonRight : PoisonMid;
787    IRB.CreateStore(Poison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
788
789    Pos += RedzoneSize;
790  }
791}
792
793// Workaround for bug 11395: we don't want to instrument stack in functions
794// with large assembly blobs (32-bit only), otherwise reg alloc may crash.
795// FIXME: remove once the bug 11395 is fixed.
796bool AddressSanitizer::LooksLikeCodeInBug11395(Instruction *I) {
797  if (LongSize != 32) return false;
798  CallInst *CI = dyn_cast<CallInst>(I);
799  if (!CI || !CI->isInlineAsm()) return false;
800  if (CI->getNumArgOperands() <= 5) return false;
801  // We have inline assembly with quite a few arguments.
802  return true;
803}
804
805// Find all static Alloca instructions and put
806// poisoned red zones around all of them.
807// Then unpoison everything back before the function returns.
808//
809// Stack poisoning does not play well with exception handling.
810// When an exception is thrown, we essentially bypass the code
811// that unpoisones the stack. This is why the run-time library has
812// to intercept __cxa_throw (as well as longjmp, etc) and unpoison the entire
813// stack in the interceptor. This however does not work inside the
814// actual function which catches the exception. Most likely because the
815// compiler hoists the load of the shadow value somewhere too high.
816// This causes asan to report a non-existing bug on 453.povray.
817// It sounds like an LLVM bug.
818bool AddressSanitizer::poisonStackInFunction(Module &M, Function &F) {
819  if (!ClStack) return false;
820  SmallVector<AllocaInst*, 16> AllocaVec;
821  SmallVector<Instruction*, 8> RetVec;
822  uint64_t TotalSize = 0;
823
824  // Filter out Alloca instructions we want (and can) handle.
825  // Collect Ret instructions.
826  for (Function::iterator FI = F.begin(), FE = F.end();
827       FI != FE; ++FI) {
828    BasicBlock &BB = *FI;
829    for (BasicBlock::iterator BI = BB.begin(), BE = BB.end();
830         BI != BE; ++BI) {
831      if (LooksLikeCodeInBug11395(BI)) return false;
832      if (isa<ReturnInst>(BI)) {
833          RetVec.push_back(BI);
834          continue;
835      }
836
837      AllocaInst *AI = dyn_cast<AllocaInst>(BI);
838      if (!AI) continue;
839      if (AI->isArrayAllocation()) continue;
840      if (!AI->isStaticAlloca()) continue;
841      if (!AI->getAllocatedType()->isSized()) continue;
842      if (AI->getAlignment() > RedzoneSize) continue;
843      AllocaVec.push_back(AI);
844      uint64_t AlignedSize =  getAlignedAllocaSize(AI);
845      TotalSize += AlignedSize;
846    }
847  }
848
849  if (AllocaVec.empty()) return false;
850
851  uint64_t LocalStackSize = TotalSize + (AllocaVec.size() + 1) * RedzoneSize;
852
853  bool DoStackMalloc = ClUseAfterReturn
854      && LocalStackSize <= kMaxStackMallocSize;
855
856  Instruction *InsBefore = AllocaVec[0];
857  IRBuilder<> IRB(InsBefore);
858
859
860  Type *ByteArrayTy = ArrayType::get(IRB.getInt8Ty(), LocalStackSize);
861  AllocaInst *MyAlloca =
862      new AllocaInst(ByteArrayTy, "MyAlloca", InsBefore);
863  MyAlloca->setAlignment(RedzoneSize);
864  assert(MyAlloca->isStaticAlloca());
865  Value *OrigStackBase = IRB.CreatePointerCast(MyAlloca, IntptrTy);
866  Value *LocalStackBase = OrigStackBase;
867
868  if (DoStackMalloc) {
869    Value *AsanStackMallocFunc = M.getOrInsertFunction(
870        kAsanStackMallocName, IntptrTy, IntptrTy, IntptrTy, NULL);
871    LocalStackBase = IRB.CreateCall2(AsanStackMallocFunc,
872        ConstantInt::get(IntptrTy, LocalStackSize), OrigStackBase);
873  }
874
875  // This string will be parsed by the run-time (DescribeStackAddress).
876  SmallString<2048> StackDescriptionStorage;
877  raw_svector_ostream StackDescription(StackDescriptionStorage);
878  StackDescription << F.getName() << " " << AllocaVec.size() << " ";
879
880  uint64_t Pos = RedzoneSize;
881  // Replace Alloca instructions with base+offset.
882  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
883    AllocaInst *AI = AllocaVec[i];
884    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
885    StringRef Name = AI->getName();
886    StackDescription << Pos << " " << SizeInBytes << " "
887                     << Name.size() << " " << Name << " ";
888    uint64_t AlignedSize = getAlignedAllocaSize(AI);
889    assert((AlignedSize % RedzoneSize) == 0);
890    AI->replaceAllUsesWith(
891        IRB.CreateIntToPtr(
892            IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy, Pos)),
893            AI->getType()));
894    Pos += AlignedSize + RedzoneSize;
895  }
896  assert(Pos == LocalStackSize);
897
898  // Write the Magic value and the frame description constant to the redzone.
899  Value *BasePlus0 = IRB.CreateIntToPtr(LocalStackBase, IntptrPtrTy);
900  IRB.CreateStore(ConstantInt::get(IntptrTy, kCurrentStackFrameMagic),
901                  BasePlus0);
902  Value *BasePlus1 = IRB.CreateAdd(LocalStackBase,
903                                   ConstantInt::get(IntptrTy, LongSize/8));
904  BasePlus1 = IRB.CreateIntToPtr(BasePlus1, IntptrPtrTy);
905  Value *Description = IRB.CreatePointerCast(
906      createPrivateGlobalForString(M, StackDescription.str()),
907      IntptrTy);
908  IRB.CreateStore(Description, BasePlus1);
909
910  // Poison the stack redzones at the entry.
911  Value *ShadowBase = memToShadow(LocalStackBase, IRB);
912  PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRB, ShadowBase, true);
913
914  Value *AsanStackFreeFunc = NULL;
915  if (DoStackMalloc) {
916    AsanStackFreeFunc = M.getOrInsertFunction(
917        kAsanStackFreeName, IRB.getVoidTy(),
918        IntptrTy, IntptrTy, IntptrTy, NULL);
919  }
920
921  // Unpoison the stack before all ret instructions.
922  for (size_t i = 0, n = RetVec.size(); i < n; i++) {
923    Instruction *Ret = RetVec[i];
924    IRBuilder<> IRBRet(Ret);
925
926    // Mark the current frame as retired.
927    IRBRet.CreateStore(ConstantInt::get(IntptrTy, kRetiredStackFrameMagic),
928                       BasePlus0);
929    // Unpoison the stack.
930    PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRBRet, ShadowBase, false);
931
932    if (DoStackMalloc) {
933      IRBRet.CreateCall3(AsanStackFreeFunc, LocalStackBase,
934                         ConstantInt::get(IntptrTy, LocalStackSize),
935                         OrigStackBase);
936    }
937  }
938
939  if (ClDebugStack) {
940    DEBUG(dbgs() << F);
941  }
942
943  return true;
944}
945
946BlackList::BlackList(const std::string &Path) {
947  Functions = NULL;
948  const char *kFunPrefix = "fun:";
949  if (!ClBlackListFile.size()) return;
950  std::string Fun;
951
952  OwningPtr<MemoryBuffer> File;
953  if (error_code EC = MemoryBuffer::getFile(ClBlackListFile.c_str(), File)) {
954    errs() << EC.message();
955    exit(1);
956  }
957  MemoryBuffer *Buff = File.take();
958  const char *Data = Buff->getBufferStart();
959  size_t DataLen = Buff->getBufferSize();
960  SmallVector<StringRef, 16> Lines;
961  SplitString(StringRef(Data, DataLen), Lines, "\n\r");
962  for (size_t i = 0, numLines = Lines.size(); i < numLines; i++) {
963    if (Lines[i].startswith(kFunPrefix)) {
964      std::string ThisFunc = Lines[i].substr(strlen(kFunPrefix));
965      if (Fun.size()) {
966        Fun += "|";
967      }
968      // add ThisFunc replacing * with .*
969      for (size_t j = 0, n = ThisFunc.size(); j < n; j++) {
970        if (ThisFunc[j] == '*')
971          Fun += '.';
972        Fun += ThisFunc[j];
973      }
974    }
975  }
976  if (Fun.size()) {
977    Functions = new Regex(Fun);
978  }
979}
980
981bool BlackList::isIn(const Function &F) {
982  if (Functions) {
983    bool Res = Functions->match(F.getName());
984    return Res;
985  }
986  return false;
987}
988