AddressSanitizer.cpp revision 7bcfc9950bac0f411f9671e8d6ce483bd219727e 
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 *kAsanModuleDtorName = "asan.module_dtor";
59static const int   kAsanCtorAndCtorPriority = 1;
60static const char *kAsanReportErrorTemplate = "__asan_report_";
61static const char *kAsanRegisterGlobalsName = "__asan_register_globals";
62static const char *kAsanUnregisterGlobalsName = "__asan_unregister_globals";
63static const char *kAsanInitName = "__asan_init";
64static const char *kAsanMappingOffsetName = "__asan_mapping_offset";
65static const char *kAsanMappingScaleName = "__asan_mapping_scale";
66static const char *kAsanStackMallocName = "__asan_stack_malloc";
67static const char *kAsanStackFreeName = "__asan_stack_free";
68
69static const int kAsanStackLeftRedzoneMagic = 0xf1;
70static const int kAsanStackMidRedzoneMagic = 0xf2;
71static const int kAsanStackRightRedzoneMagic = 0xf3;
72static const int kAsanStackPartialRedzoneMagic = 0xf4;
73
74// Command-line flags.
75
76// This flag may need to be replaced with -f[no-]asan-reads.
77static cl::opt<bool> ClInstrumentReads("asan-instrument-reads",
78       cl::desc("instrument read instructions"), cl::Hidden, cl::init(true));
79static cl::opt<bool> ClInstrumentWrites("asan-instrument-writes",
80       cl::desc("instrument write instructions"), cl::Hidden, cl::init(true));
81// This flag may need to be replaced with -f[no]asan-stack.
82static cl::opt<bool> ClStack("asan-stack",
83       cl::desc("Handle stack memory"), cl::Hidden, cl::init(true));
84// This flag may need to be replaced with -f[no]asan-use-after-return.
85static cl::opt<bool> ClUseAfterReturn("asan-use-after-return",
86       cl::desc("Check return-after-free"), cl::Hidden, cl::init(false));
87// This flag may need to be replaced with -f[no]asan-globals.
88static cl::opt<bool> ClGlobals("asan-globals",
89       cl::desc("Handle global objects"), cl::Hidden, cl::init(true));
90static cl::opt<bool> ClMemIntrin("asan-memintrin",
91       cl::desc("Handle memset/memcpy/memmove"), cl::Hidden, cl::init(true));
92// This flag may need to be replaced with -fasan-blacklist.
93static cl::opt<std::string>  ClBlackListFile("asan-blacklist",
94       cl::desc("File containing the list of functions to ignore "
95                "during instrumentation"), cl::Hidden);
96static cl::opt<bool> ClUseCall("asan-use-call",
97       cl::desc("Use function call to generate a crash"), cl::Hidden,
98       cl::init(true));
99
100// These flags allow to change the shadow mapping.
101// The shadow mapping looks like
102//    Shadow = (Mem >> scale) + (1 << offset_log)
103static cl::opt<int> ClMappingScale("asan-mapping-scale",
104       cl::desc("scale of asan shadow mapping"), cl::Hidden, cl::init(0));
105static cl::opt<int> ClMappingOffsetLog("asan-mapping-offset-log",
106       cl::desc("offset of asan shadow mapping"), cl::Hidden, cl::init(-1));
107
108// Optimization flags. Not user visible, used mostly for testing
109// and benchmarking the tool.
110static cl::opt<bool> ClOpt("asan-opt",
111       cl::desc("Optimize instrumentation"), cl::Hidden, cl::init(true));
112static cl::opt<bool> ClOptSameTemp("asan-opt-same-temp",
113       cl::desc("Instrument the same temp just once"), cl::Hidden,
114       cl::init(true));
115static cl::opt<bool> ClOptGlobals("asan-opt-globals",
116       cl::desc("Don't instrument scalar globals"), cl::Hidden, cl::init(true));
117
118// Debug flags.
119static cl::opt<int> ClDebug("asan-debug", cl::desc("debug"), cl::Hidden,
120                            cl::init(0));
121static cl::opt<int> ClDebugStack("asan-debug-stack", cl::desc("debug stack"),
122                                 cl::Hidden, cl::init(0));
123static cl::opt<std::string> ClDebugFunc("asan-debug-func",
124                                        cl::Hidden, cl::desc("Debug func"));
125static cl::opt<int> ClDebugMin("asan-debug-min", cl::desc("Debug min inst"),
126                               cl::Hidden, cl::init(-1));
127static cl::opt<int> ClDebugMax("asan-debug-max", cl::desc("Debug man inst"),
128                               cl::Hidden, cl::init(-1));
129
130namespace {
131
132// Blacklisted functions are not instrumented.
133// The blacklist file contains one or more lines like this:
134// ---
135// fun:FunctionWildCard
136// ---
137// This is similar to the "ignore" feature of ThreadSanitizer.
138// http://code.google.com/p/data-race-test/wiki/ThreadSanitizerIgnores
139class BlackList {
140 public:
141  BlackList(const std::string &Path);
142  bool isIn(const Function &F);
143 private:
144  Regex *Functions;
145};
146
147/// AddressSanitizer: instrument the code in module to find memory bugs.
148struct AddressSanitizer : public ModulePass {
149  AddressSanitizer();
150  void instrumentMop(Instruction *I);
151  void instrumentAddress(Instruction *OrigIns, IRBuilder<> &IRB,
152                         Value *Addr, uint32_t TypeSize, bool IsWrite);
153  Instruction *generateCrashCode(IRBuilder<> &IRB, Value *Addr,
154                                 bool IsWrite, uint32_t TypeSize);
155  bool instrumentMemIntrinsic(MemIntrinsic *MI);
156  void instrumentMemIntrinsicParam(Instruction *OrigIns, Value *Addr,
157                                  Value *Size,
158                                   Instruction *InsertBefore, bool IsWrite);
159  Value *memToShadow(Value *Shadow, IRBuilder<> &IRB);
160  bool handleFunction(Module &M, Function &F);
161  bool poisonStackInFunction(Module &M, Function &F);
162  virtual bool runOnModule(Module &M);
163  bool insertGlobalRedzones(Module &M);
164  BranchInst *splitBlockAndInsertIfThen(Instruction *SplitBefore, Value *Cmp);
165  static char ID;  // Pass identification, replacement for typeid
166
167 private:
168
169  uint64_t getAllocaSizeInBytes(AllocaInst *AI) {
170    Type *Ty = AI->getAllocatedType();
171    uint64_t SizeInBytes = TD->getTypeStoreSizeInBits(Ty) / 8;
172    return SizeInBytes;
173  }
174  uint64_t getAlignedSize(uint64_t SizeInBytes) {
175    return ((SizeInBytes + RedzoneSize - 1)
176            / RedzoneSize) * RedzoneSize;
177  }
178  uint64_t getAlignedAllocaSize(AllocaInst *AI) {
179    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
180    return getAlignedSize(SizeInBytes);
181  }
182
183  void PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec, IRBuilder<> IRB,
184                   Value *ShadowBase, bool DoPoison);
185  bool LooksLikeCodeInBug11395(Instruction *I);
186
187  Module      *CurrentModule;
188  LLVMContext *C;
189  TargetData *TD;
190  uint64_t MappingOffset;
191  int MappingScale;
192  size_t RedzoneSize;
193  int LongSize;
194  Type *IntptrTy;
195  Type *IntptrPtrTy;
196  Function *AsanCtorFunction;
197  Function *AsanInitFunction;
198  Instruction *CtorInsertBefore;
199  OwningPtr<BlackList> BL;
200};
201}  // namespace
202
203char AddressSanitizer::ID = 0;
204INITIALIZE_PASS(AddressSanitizer, "asan",
205    "AddressSanitizer: detects use-after-free and out-of-bounds bugs.",
206    false, false)
207AddressSanitizer::AddressSanitizer() : ModulePass(ID) { }
208ModulePass *llvm::createAddressSanitizerPass() {
209  return new AddressSanitizer();
210}
211
212// Create a constant for Str so that we can pass it to the run-time lib.
213static GlobalVariable *createPrivateGlobalForString(Module &M, StringRef Str) {
214  Constant *StrConst = ConstantArray::get(M.getContext(), Str);
215  return new GlobalVariable(M, StrConst->getType(), true,
216                            GlobalValue::PrivateLinkage, StrConst, "");
217}
218
219// Split the basic block and insert an if-then code.
220// Before:
221//   Head
222//   SplitBefore
223//   Tail
224// After:
225//   Head
226//   if (Cmp)
227//     NewBasicBlock
228//   SplitBefore
229//   Tail
230//
231// Returns the NewBasicBlock's terminator.
232BranchInst *AddressSanitizer::splitBlockAndInsertIfThen(
233    Instruction *SplitBefore, Value *Cmp) {
234  BasicBlock *Head = SplitBefore->getParent();
235  BasicBlock *Tail = Head->splitBasicBlock(SplitBefore);
236  TerminatorInst *HeadOldTerm = Head->getTerminator();
237  BasicBlock *NewBasicBlock =
238      BasicBlock::Create(*C, "", Head->getParent());
239  BranchInst *HeadNewTerm = BranchInst::Create(/*ifTrue*/NewBasicBlock,
240                                               /*ifFalse*/Tail,
241                                               Cmp);
242  ReplaceInstWithInst(HeadOldTerm, HeadNewTerm);
243
244  BranchInst *CheckTerm = BranchInst::Create(Tail, NewBasicBlock);
245  return CheckTerm;
246}
247
248Value *AddressSanitizer::memToShadow(Value *Shadow, IRBuilder<> &IRB) {
249  // Shadow >> scale
250  Shadow = IRB.CreateLShr(Shadow, MappingScale);
251  if (MappingOffset == 0)
252    return Shadow;
253  // (Shadow >> scale) | offset
254  return IRB.CreateOr(Shadow, ConstantInt::get(IntptrTy,
255                                               MappingOffset));
256}
257
258void AddressSanitizer::instrumentMemIntrinsicParam(Instruction *OrigIns,
259    Value *Addr, Value *Size, Instruction *InsertBefore, bool IsWrite) {
260  // Check the first byte.
261  {
262    IRBuilder<> IRB(InsertBefore);
263    instrumentAddress(OrigIns, IRB, Addr, 8, IsWrite);
264  }
265  // Check the last byte.
266  {
267    IRBuilder<> IRB(InsertBefore);
268    Value *SizeMinusOne = IRB.CreateSub(
269        Size, ConstantInt::get(Size->getType(), 1));
270    SizeMinusOne = IRB.CreateIntCast(SizeMinusOne, IntptrTy, false);
271    Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
272    Value *AddrPlusSizeMinisOne = IRB.CreateAdd(AddrLong, SizeMinusOne);
273    instrumentAddress(OrigIns, IRB, AddrPlusSizeMinisOne, 8, IsWrite);
274  }
275}
276
277// Instrument memset/memmove/memcpy
278bool AddressSanitizer::instrumentMemIntrinsic(MemIntrinsic *MI) {
279  Value *Dst = MI->getDest();
280  MemTransferInst *MemTran = dyn_cast<MemTransferInst>(MI);
281  Value *Src = MemTran ? MemTran->getSource() : NULL;
282  Value *Length = MI->getLength();
283
284  Constant *ConstLength = dyn_cast<Constant>(Length);
285  Instruction *InsertBefore = MI;
286  if (ConstLength) {
287    if (ConstLength->isNullValue()) return false;
288  } else {
289    // The size is not a constant so it could be zero -- check at run-time.
290    IRBuilder<> IRB(InsertBefore);
291
292    Value *Cmp = IRB.CreateICmpNE(Length,
293                                   Constant::getNullValue(Length->getType()));
294    InsertBefore = splitBlockAndInsertIfThen(InsertBefore, Cmp);
295  }
296
297  instrumentMemIntrinsicParam(MI, Dst, Length, InsertBefore, true);
298  if (Src)
299    instrumentMemIntrinsicParam(MI, Src, Length, InsertBefore, false);
300  return true;
301}
302
303static Value *getLDSTOperand(Instruction *I) {
304  if (LoadInst *LI = dyn_cast<LoadInst>(I)) {
305    return LI->getPointerOperand();
306  }
307  return cast<StoreInst>(*I).getPointerOperand();
308}
309
310void AddressSanitizer::instrumentMop(Instruction *I) {
311  int IsWrite = isa<StoreInst>(*I);
312  Value *Addr = getLDSTOperand(I);
313  if (ClOpt && ClOptGlobals && isa<GlobalVariable>(Addr)) {
314    // We are accessing a global scalar variable. Nothing to catch here.
315    return;
316  }
317  Type *OrigPtrTy = Addr->getType();
318  Type *OrigTy = cast<PointerType>(OrigPtrTy)->getElementType();
319
320  assert(OrigTy->isSized());
321  uint32_t TypeSize = TD->getTypeStoreSizeInBits(OrigTy);
322
323  if (TypeSize != 8  && TypeSize != 16 &&
324      TypeSize != 32 && TypeSize != 64 && TypeSize != 128) {
325    // Ignore all unusual sizes.
326    return;
327  }
328
329  IRBuilder<> IRB(I);
330  instrumentAddress(I, IRB, Addr, TypeSize, IsWrite);
331}
332
333Instruction *AddressSanitizer::generateCrashCode(
334    IRBuilder<> &IRB, Value *Addr, bool IsWrite, uint32_t TypeSize) {
335
336  if (ClUseCall) {
337    // Here we use a call instead of arch-specific asm to report an error.
338    // This is almost always slower (because the codegen needs to generate
339    // prologue/epilogue for otherwise leaf functions) and generates more code.
340    // This mode could be useful if we can not use SIGILL for some reason.
341    //
342    // IsWrite and TypeSize are encoded in the function name.
343    std::string FunctionName = std::string(kAsanReportErrorTemplate) +
344        (IsWrite ? "store" : "load") + itostr(TypeSize / 8);
345    Value *ReportWarningFunc = CurrentModule->getOrInsertFunction(
346        FunctionName, IRB.getVoidTy(), IntptrTy, NULL);
347    CallInst *Call = IRB.CreateCall(ReportWarningFunc, Addr);
348    Call->setDoesNotReturn();
349    return Call;
350  }
351
352  uint32_t LogOfSizeInBytes = CountTrailingZeros_32(TypeSize / 8);
353  assert(8U * (1 << LogOfSizeInBytes) == TypeSize);
354  uint8_t TelltaleValue = IsWrite * 8 + LogOfSizeInBytes;
355  assert(TelltaleValue < 16);
356
357  // Move the failing address to %rax/%eax
358  FunctionType *Fn1Ty = FunctionType::get(
359      IRB.getVoidTy(), ArrayRef<Type*>(IntptrTy), false);
360  const char *MovStr = LongSize == 32
361      ? "mov $0, %eax" : "mov $0, %rax";
362  Value *AsmMov = InlineAsm::get(
363      Fn1Ty, StringRef(MovStr), StringRef("r"), true);
364  IRB.CreateCall(AsmMov, Addr);
365
366  // crash with ud2; could use int3, but it is less friendly to gdb.
367  // after ud2 put a 1-byte instruction that encodes the access type and size.
368
369  const char *TelltaleInsns[16] = {
370    "push   %eax",  // 0x50
371    "push   %ecx",  // 0x51
372    "push   %edx",  // 0x52
373    "push   %ebx",  // 0x53
374    "push   %esp",  // 0x54
375    "push   %ebp",  // 0x55
376    "push   %esi",  // 0x56
377    "push   %edi",  // 0x57
378    "pop    %eax",  // 0x58
379    "pop    %ecx",  // 0x59
380    "pop    %edx",  // 0x5a
381    "pop    %ebx",  // 0x5b
382    "pop    %esp",  // 0x5c
383    "pop    %ebp",  // 0x5d
384    "pop    %esi",  // 0x5e
385    "pop    %edi"   // 0x5f
386  };
387
388  std::string AsmStr = "ud2;";
389  AsmStr += TelltaleInsns[TelltaleValue];
390  Value *MyAsm = InlineAsm::get(FunctionType::get(Type::getVoidTy(*C), false),
391                                StringRef(AsmStr), StringRef(""), true);
392  CallInst *AsmCall = IRB.CreateCall(MyAsm);
393
394  // This saves us one jump, but triggers a bug in RA (or somewhere else):
395  // while building 483.xalancbmk the compiler goes into infinite loop in
396  // llvm::SpillPlacement::iterate() / RAGreedy::growRegion
397  // AsmCall->setDoesNotReturn();
398  return AsmCall;
399}
400
401void AddressSanitizer::instrumentAddress(Instruction *OrigIns,
402                                         IRBuilder<> &IRB, Value *Addr,
403                                         uint32_t TypeSize, bool IsWrite) {
404  Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
405
406  Type *ShadowTy  = IntegerType::get(
407      *C, std::max(8U, TypeSize >> MappingScale));
408  Type *ShadowPtrTy = PointerType::get(ShadowTy, 0);
409  Value *ShadowPtr = memToShadow(AddrLong, IRB);
410  Value *CmpVal = Constant::getNullValue(ShadowTy);
411  Value *ShadowValue = IRB.CreateLoad(
412      IRB.CreateIntToPtr(ShadowPtr, ShadowPtrTy));
413
414  Value *Cmp = IRB.CreateICmpNE(ShadowValue, CmpVal);
415
416  Instruction *CheckTerm = splitBlockAndInsertIfThen(
417      cast<Instruction>(Cmp)->getNextNode(), Cmp);
418  IRBuilder<> IRB2(CheckTerm);
419
420  size_t Granularity = 1 << MappingScale;
421  if (TypeSize < 8 * Granularity) {
422    // Addr & (Granularity - 1)
423    Value *Lower3Bits = IRB2.CreateAnd(
424        AddrLong, ConstantInt::get(IntptrTy, Granularity - 1));
425    // (Addr & (Granularity - 1)) + size - 1
426    Value *LastAccessedByte = IRB2.CreateAdd(
427        Lower3Bits, ConstantInt::get(IntptrTy, TypeSize / 8 - 1));
428    // (uint8_t) ((Addr & (Granularity-1)) + size - 1)
429    LastAccessedByte = IRB2.CreateIntCast(
430        LastAccessedByte, IRB.getInt8Ty(), false);
431    // ((uint8_t) ((Addr & (Granularity-1)) + size - 1)) >= ShadowValue
432    Value *Cmp2 = IRB2.CreateICmpSGE(LastAccessedByte, ShadowValue);
433
434    CheckTerm = splitBlockAndInsertIfThen(CheckTerm, Cmp2);
435  }
436
437  IRBuilder<> IRB1(CheckTerm);
438  Instruction *Crash = generateCrashCode(IRB1, AddrLong, IsWrite, TypeSize);
439  Crash->setDebugLoc(OrigIns->getDebugLoc());
440  ReplaceInstWithInst(CheckTerm, new UnreachableInst(*C));
441}
442
443// This function replaces all global variables with new variables that have
444// trailing redzones. It also creates a function that poisons
445// redzones and inserts this function into llvm.global_ctors.
446bool AddressSanitizer::insertGlobalRedzones(Module &M) {
447  SmallVector<GlobalVariable *, 16> GlobalsToChange;
448
449  for (Module::GlobalListType::iterator G = M.getGlobalList().begin(),
450       E = M.getGlobalList().end(); G != E; ++G) {
451    Type *Ty = cast<PointerType>(G->getType())->getElementType();
452    DEBUG(dbgs() << "GLOBAL: " << *G);
453
454    if (!Ty->isSized()) continue;
455    if (!G->hasInitializer()) continue;
456    // Touch only those globals that will not be defined in other modules.
457    // Don't handle ODR type linkages since other modules may be built w/o asan.
458    if (G->getLinkage() != GlobalVariable::ExternalLinkage &&
459        G->getLinkage() != GlobalVariable::PrivateLinkage &&
460        G->getLinkage() != GlobalVariable::InternalLinkage)
461      continue;
462    // Two problems with thread-locals:
463    //   - The address of the main thread's copy can't be computed at link-time.
464    //   - Need to poison all copies, not just the main thread's one.
465    if (G->isThreadLocal())
466      continue;
467    // For now, just ignore this Alloca if the alignment is large.
468    if (G->getAlignment() > RedzoneSize) continue;
469
470    // Ignore all the globals with the names starting with "\01L_OBJC_".
471    // Many of those are put into the .cstring section. The linker compresses
472    // that section by removing the spare \0s after the string terminator, so
473    // our redzones get broken.
474    if ((G->getName().find("\01L_OBJC_") == 0) ||
475        (G->getName().find("\01l_OBJC_") == 0)) {
476      DEBUG(dbgs() << "Ignoring \\01L_OBJC_* global: " << *G);
477      continue;
478    }
479
480    // Ignore the globals from the __OBJC section. The ObjC runtime assumes
481    // those conform to /usr/lib/objc/runtime.h, so we can't add redzones to
482    // them.
483    if (G->hasSection()) {
484      StringRef Section(G->getSection());
485      if ((Section.find("__OBJC,") == 0) ||
486          (Section.find("__DATA, __objc_") == 0)) {
487        DEBUG(dbgs() << "Ignoring ObjC runtime global: " << *G);
488        continue;
489      }
490    }
491
492    GlobalsToChange.push_back(G);
493  }
494
495  size_t n = GlobalsToChange.size();
496  if (n == 0) return false;
497
498  // A global is described by a structure
499  //   size_t beg;
500  //   size_t size;
501  //   size_t size_with_redzone;
502  //   const char *name;
503  // We initialize an array of such structures and pass it to a run-time call.
504  StructType *GlobalStructTy = StructType::get(IntptrTy, IntptrTy,
505                                               IntptrTy, IntptrTy, NULL);
506  SmallVector<Constant *, 16> Initializers(n);
507
508  IRBuilder<> IRB(CtorInsertBefore);
509
510  for (size_t i = 0; i < n; i++) {
511    GlobalVariable *G = GlobalsToChange[i];
512    PointerType *PtrTy = cast<PointerType>(G->getType());
513    Type *Ty = PtrTy->getElementType();
514    uint64_t SizeInBytes = TD->getTypeStoreSizeInBits(Ty) / 8;
515    uint64_t RightRedzoneSize = RedzoneSize +
516        (RedzoneSize - (SizeInBytes % RedzoneSize));
517    Type *RightRedZoneTy = ArrayType::get(IRB.getInt8Ty(), RightRedzoneSize);
518
519    StructType *NewTy = StructType::get(Ty, RightRedZoneTy, NULL);
520    Constant *NewInitializer = ConstantStruct::get(
521        NewTy, G->getInitializer(),
522        Constant::getNullValue(RightRedZoneTy), NULL);
523
524    GlobalVariable *Name = createPrivateGlobalForString(M, G->getName());
525
526    // Create a new global variable with enough space for a redzone.
527    GlobalVariable *NewGlobal = new GlobalVariable(
528        M, NewTy, G->isConstant(), G->getLinkage(),
529        NewInitializer, "", G, G->isThreadLocal());
530    NewGlobal->copyAttributesFrom(G);
531    NewGlobal->setAlignment(RedzoneSize);
532
533    Value *Indices2[2];
534    Indices2[0] = IRB.getInt32(0);
535    Indices2[1] = IRB.getInt32(0);
536
537    G->replaceAllUsesWith(
538        ConstantExpr::getGetElementPtr(NewGlobal, Indices2, 2));
539    NewGlobal->takeName(G);
540    G->eraseFromParent();
541
542    Initializers[i] = ConstantStruct::get(
543        GlobalStructTy,
544        ConstantExpr::getPointerCast(NewGlobal, IntptrTy),
545        ConstantInt::get(IntptrTy, SizeInBytes),
546        ConstantInt::get(IntptrTy, SizeInBytes + RightRedzoneSize),
547        ConstantExpr::getPointerCast(Name, IntptrTy),
548        NULL);
549    DEBUG(dbgs() << "NEW GLOBAL:\n" << *NewGlobal);
550  }
551
552  ArrayType *ArrayOfGlobalStructTy = ArrayType::get(GlobalStructTy, n);
553  GlobalVariable *AllGlobals = new GlobalVariable(
554      M, ArrayOfGlobalStructTy, false, GlobalVariable::PrivateLinkage,
555      ConstantArray::get(ArrayOfGlobalStructTy, Initializers), "");
556
557  Function *AsanRegisterGlobals = cast<Function>(M.getOrInsertFunction(
558      kAsanRegisterGlobalsName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
559  AsanRegisterGlobals->setLinkage(Function::ExternalLinkage);
560
561  IRB.CreateCall2(AsanRegisterGlobals,
562                  IRB.CreatePointerCast(AllGlobals, IntptrTy),
563                  ConstantInt::get(IntptrTy, n));
564
565  // We also need to unregister globals at the end, e.g. when a shared library
566  // gets closed.
567  Function *AsanDtorFunction = Function::Create(
568      FunctionType::get(Type::getVoidTy(*C), false),
569      GlobalValue::InternalLinkage, kAsanModuleDtorName, &M);
570  BasicBlock *AsanDtorBB = BasicBlock::Create(*C, "", AsanDtorFunction);
571  IRBuilder<> IRB_Dtor(ReturnInst::Create(*C, AsanDtorBB));
572  Function *AsanUnregisterGlobals = cast<Function>(M.getOrInsertFunction(
573      kAsanUnregisterGlobalsName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
574  AsanUnregisterGlobals->setLinkage(Function::ExternalLinkage);
575
576  IRB_Dtor.CreateCall2(AsanUnregisterGlobals,
577                       IRB.CreatePointerCast(AllGlobals, IntptrTy),
578                       ConstantInt::get(IntptrTy, n));
579  appendToGlobalDtors(M, AsanDtorFunction, kAsanCtorAndCtorPriority);
580
581  DEBUG(dbgs() << M);
582  return true;
583}
584
585// virtual
586bool AddressSanitizer::runOnModule(Module &M) {
587  // Initialize the private fields. No one has accessed them before.
588  TD = getAnalysisIfAvailable<TargetData>();
589  if (!TD)
590    return false;
591  BL.reset(new BlackList(ClBlackListFile));
592
593  CurrentModule = &M;
594  C = &(M.getContext());
595  LongSize = TD->getPointerSizeInBits();
596  IntptrTy = Type::getIntNTy(*C, LongSize);
597  IntptrPtrTy = PointerType::get(IntptrTy, 0);
598
599  AsanCtorFunction = Function::Create(
600      FunctionType::get(Type::getVoidTy(*C), false),
601      GlobalValue::InternalLinkage, kAsanModuleCtorName, &M);
602  BasicBlock *AsanCtorBB = BasicBlock::Create(*C, "", AsanCtorFunction);
603  CtorInsertBefore = ReturnInst::Create(*C, AsanCtorBB);
604
605  // call __asan_init in the module ctor.
606  IRBuilder<> IRB(CtorInsertBefore);
607  AsanInitFunction = cast<Function>(
608      M.getOrInsertFunction(kAsanInitName, IRB.getVoidTy(), NULL));
609  AsanInitFunction->setLinkage(Function::ExternalLinkage);
610  IRB.CreateCall(AsanInitFunction);
611
612  MappingOffset = LongSize == 32
613      ? kDefaultShadowOffset32 : kDefaultShadowOffset64;
614  if (ClMappingOffsetLog >= 0) {
615    if (ClMappingOffsetLog == 0) {
616      // special case
617      MappingOffset = 0;
618    } else {
619      MappingOffset = 1ULL << ClMappingOffsetLog;
620    }
621  }
622  MappingScale = kDefaultShadowScale;
623  if (ClMappingScale) {
624    MappingScale = ClMappingScale;
625  }
626  // Redzone used for stack and globals is at least 32 bytes.
627  // For scales 6 and 7, the redzone has to be 64 and 128 bytes respectively.
628  RedzoneSize = std::max(32, (int)(1 << MappingScale));
629
630  bool Res = false;
631
632  if (ClGlobals)
633    Res |= insertGlobalRedzones(M);
634
635  // Tell the run-time the current values of mapping offset and scale.
636  GlobalValue *asan_mapping_offset =
637      new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
638                     ConstantInt::get(IntptrTy, MappingOffset),
639                     kAsanMappingOffsetName);
640  GlobalValue *asan_mapping_scale =
641      new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
642                         ConstantInt::get(IntptrTy, MappingScale),
643                         kAsanMappingScaleName);
644  // Read these globals, otherwise they may be optimized away.
645  IRB.CreateLoad(asan_mapping_scale, true);
646  IRB.CreateLoad(asan_mapping_offset, true);
647
648
649  for (Module::iterator F = M.begin(), E = M.end(); F != E; ++F) {
650    if (F->isDeclaration()) continue;
651    Res |= handleFunction(M, *F);
652  }
653
654  appendToGlobalCtors(M, AsanCtorFunction, kAsanCtorAndCtorPriority);
655
656  return Res;
657}
658
659bool AddressSanitizer::handleFunction(Module &M, Function &F) {
660  if (BL->isIn(F)) return false;
661  if (&F == AsanCtorFunction) return false;
662
663  if (!ClDebugFunc.empty() && ClDebugFunc != F.getName())
664    return false;
665  // We want to instrument every address only once per basic block
666  // (unless there are calls between uses).
667  SmallSet<Value*, 16> TempsToInstrument;
668  SmallVector<Instruction*, 16> ToInstrument;
669
670  // Fill the set of memory operations to instrument.
671  for (Function::iterator FI = F.begin(), FE = F.end();
672       FI != FE; ++FI) {
673    TempsToInstrument.clear();
674    for (BasicBlock::iterator BI = FI->begin(), BE = FI->end();
675         BI != BE; ++BI) {
676      if ((isa<LoadInst>(BI) && ClInstrumentReads) ||
677          (isa<StoreInst>(BI) && ClInstrumentWrites)) {
678        Value *Addr = getLDSTOperand(BI);
679        if (ClOpt && ClOptSameTemp) {
680          if (!TempsToInstrument.insert(Addr))
681            continue;  // We've seen this temp in the current BB.
682        }
683      } else if (isa<MemIntrinsic>(BI) && ClMemIntrin) {
684        // ok, take it.
685      } else {
686        if (isa<CallInst>(BI)) {
687          // A call inside BB.
688          TempsToInstrument.clear();
689        }
690        continue;
691      }
692      ToInstrument.push_back(BI);
693    }
694  }
695
696  // Instrument.
697  int NumInstrumented = 0;
698  for (size_t i = 0, n = ToInstrument.size(); i != n; i++) {
699    Instruction *Inst = ToInstrument[i];
700    if (ClDebugMin < 0 || ClDebugMax < 0 ||
701        (NumInstrumented >= ClDebugMin && NumInstrumented <= ClDebugMax)) {
702      if (isa<StoreInst>(Inst) || isa<LoadInst>(Inst))
703        instrumentMop(Inst);
704      else
705        instrumentMemIntrinsic(cast<MemIntrinsic>(Inst));
706    }
707    NumInstrumented++;
708  }
709
710  DEBUG(dbgs() << F);
711
712  bool ChangedStack = poisonStackInFunction(M, F);
713
714  // For each NSObject descendant having a +load method, this method is invoked
715  // by the ObjC runtime before any of the static constructors is called.
716  // Therefore we need to instrument such methods with a call to __asan_init
717  // at the beginning in order to initialize our runtime before any access to
718  // the shadow memory.
719  // We cannot just ignore these methods, because they may call other
720  // instrumented functions.
721  if (F.getName().find(" load]") != std::string::npos) {
722    IRBuilder<> IRB(F.begin()->begin());
723    IRB.CreateCall(AsanInitFunction);
724  }
725
726  return NumInstrumented > 0 || ChangedStack;
727}
728
729static uint64_t ValueForPoison(uint64_t PoisonByte, size_t ShadowRedzoneSize) {
730  if (ShadowRedzoneSize == 1) return PoisonByte;
731  if (ShadowRedzoneSize == 2) return (PoisonByte << 8) + PoisonByte;
732  if (ShadowRedzoneSize == 4)
733    return (PoisonByte << 24) + (PoisonByte << 16) +
734        (PoisonByte << 8) + (PoisonByte);
735  assert(0 && "ShadowRedzoneSize is either 1, 2 or 4");
736  return 0;
737}
738
739static void PoisonShadowPartialRightRedzone(uint8_t *Shadow,
740                                            size_t Size,
741                                            size_t RedzoneSize,
742                                            size_t ShadowGranularity,
743                                            uint8_t Magic) {
744  for (size_t i = 0; i < RedzoneSize;
745       i+= ShadowGranularity, Shadow++) {
746    if (i + ShadowGranularity <= Size) {
747      *Shadow = 0;  // fully addressable
748    } else if (i >= Size) {
749      *Shadow = Magic;  // unaddressable
750    } else {
751      *Shadow = Size - i;  // first Size-i bytes are addressable
752    }
753  }
754}
755
756void AddressSanitizer::PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec,
757                                   IRBuilder<> IRB,
758                                   Value *ShadowBase, bool DoPoison) {
759  size_t ShadowRZSize = RedzoneSize >> MappingScale;
760  assert(ShadowRZSize >= 1 && ShadowRZSize <= 4);
761  Type *RZTy = Type::getIntNTy(*C, ShadowRZSize * 8);
762  Type *RZPtrTy = PointerType::get(RZTy, 0);
763
764  Value *PoisonLeft  = ConstantInt::get(RZTy,
765    ValueForPoison(DoPoison ? kAsanStackLeftRedzoneMagic : 0LL, ShadowRZSize));
766  Value *PoisonMid   = ConstantInt::get(RZTy,
767    ValueForPoison(DoPoison ? kAsanStackMidRedzoneMagic : 0LL, ShadowRZSize));
768  Value *PoisonRight = ConstantInt::get(RZTy,
769    ValueForPoison(DoPoison ? kAsanStackRightRedzoneMagic : 0LL, ShadowRZSize));
770
771  // poison the first red zone.
772  IRB.CreateStore(PoisonLeft, IRB.CreateIntToPtr(ShadowBase, RZPtrTy));
773
774  // poison all other red zones.
775  uint64_t Pos = RedzoneSize;
776  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
777    AllocaInst *AI = AllocaVec[i];
778    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
779    uint64_t AlignedSize = getAlignedAllocaSize(AI);
780    assert(AlignedSize - SizeInBytes < RedzoneSize);
781    Value *Ptr = NULL;
782
783    Pos += AlignedSize;
784
785    assert(ShadowBase->getType() == IntptrTy);
786    if (SizeInBytes < AlignedSize) {
787      // Poison the partial redzone at right
788      Ptr = IRB.CreateAdd(
789          ShadowBase, ConstantInt::get(IntptrTy,
790                                       (Pos >> MappingScale) - ShadowRZSize));
791      size_t AddressableBytes = RedzoneSize - (AlignedSize - SizeInBytes);
792      uint32_t Poison = 0;
793      if (DoPoison) {
794        PoisonShadowPartialRightRedzone((uint8_t*)&Poison, AddressableBytes,
795                                        RedzoneSize,
796                                        1ULL << MappingScale,
797                                        kAsanStackPartialRedzoneMagic);
798      }
799      Value *PartialPoison = ConstantInt::get(RZTy, Poison);
800      IRB.CreateStore(PartialPoison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
801    }
802
803    // Poison the full redzone at right.
804    Ptr = IRB.CreateAdd(ShadowBase,
805                        ConstantInt::get(IntptrTy, Pos >> MappingScale));
806    Value *Poison = i == AllocaVec.size() - 1 ? PoisonRight : PoisonMid;
807    IRB.CreateStore(Poison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
808
809    Pos += RedzoneSize;
810  }
811}
812
813// Workaround for bug 11395: we don't want to instrument stack in functions
814// with large assembly blobs (32-bit only), otherwise reg alloc may crash.
815// FIXME: remove once the bug 11395 is fixed.
816bool AddressSanitizer::LooksLikeCodeInBug11395(Instruction *I) {
817  if (LongSize != 32) return false;
818  CallInst *CI = dyn_cast<CallInst>(I);
819  if (!CI || !CI->isInlineAsm()) return false;
820  if (CI->getNumArgOperands() <= 5) return false;
821  // We have inline assembly with quite a few arguments.
822  return true;
823}
824
825// Find all static Alloca instructions and put
826// poisoned red zones around all of them.
827// Then unpoison everything back before the function returns.
828//
829// Stack poisoning does not play well with exception handling.
830// When an exception is thrown, we essentially bypass the code
831// that unpoisones the stack. This is why the run-time library has
832// to intercept __cxa_throw (as well as longjmp, etc) and unpoison the entire
833// stack in the interceptor. This however does not work inside the
834// actual function which catches the exception. Most likely because the
835// compiler hoists the load of the shadow value somewhere too high.
836// This causes asan to report a non-existing bug on 453.povray.
837// It sounds like an LLVM bug.
838bool AddressSanitizer::poisonStackInFunction(Module &M, Function &F) {
839  if (!ClStack) return false;
840  SmallVector<AllocaInst*, 16> AllocaVec;
841  SmallVector<Instruction*, 8> RetVec;
842  uint64_t TotalSize = 0;
843
844  // Filter out Alloca instructions we want (and can) handle.
845  // Collect Ret instructions.
846  for (Function::iterator FI = F.begin(), FE = F.end();
847       FI != FE; ++FI) {
848    BasicBlock &BB = *FI;
849    for (BasicBlock::iterator BI = BB.begin(), BE = BB.end();
850         BI != BE; ++BI) {
851      if (LooksLikeCodeInBug11395(BI)) return false;
852      if (isa<ReturnInst>(BI)) {
853          RetVec.push_back(BI);
854          continue;
855      }
856
857      AllocaInst *AI = dyn_cast<AllocaInst>(BI);
858      if (!AI) continue;
859      if (AI->isArrayAllocation()) continue;
860      if (!AI->isStaticAlloca()) continue;
861      if (!AI->getAllocatedType()->isSized()) continue;
862      if (AI->getAlignment() > RedzoneSize) continue;
863      AllocaVec.push_back(AI);
864      uint64_t AlignedSize =  getAlignedAllocaSize(AI);
865      TotalSize += AlignedSize;
866    }
867  }
868
869  if (AllocaVec.empty()) return false;
870
871  uint64_t LocalStackSize = TotalSize + (AllocaVec.size() + 1) * RedzoneSize;
872
873  bool DoStackMalloc = ClUseAfterReturn
874      && LocalStackSize <= kMaxStackMallocSize;
875
876  Instruction *InsBefore = AllocaVec[0];
877  IRBuilder<> IRB(InsBefore);
878
879
880  Type *ByteArrayTy = ArrayType::get(IRB.getInt8Ty(), LocalStackSize);
881  AllocaInst *MyAlloca =
882      new AllocaInst(ByteArrayTy, "MyAlloca", InsBefore);
883  MyAlloca->setAlignment(RedzoneSize);
884  assert(MyAlloca->isStaticAlloca());
885  Value *OrigStackBase = IRB.CreatePointerCast(MyAlloca, IntptrTy);
886  Value *LocalStackBase = OrigStackBase;
887
888  if (DoStackMalloc) {
889    Value *AsanStackMallocFunc = M.getOrInsertFunction(
890        kAsanStackMallocName, IntptrTy, IntptrTy, IntptrTy, NULL);
891    LocalStackBase = IRB.CreateCall2(AsanStackMallocFunc,
892        ConstantInt::get(IntptrTy, LocalStackSize), OrigStackBase);
893  }
894
895  // This string will be parsed by the run-time (DescribeStackAddress).
896  SmallString<2048> StackDescriptionStorage;
897  raw_svector_ostream StackDescription(StackDescriptionStorage);
898  StackDescription << F.getName() << " " << AllocaVec.size() << " ";
899
900  uint64_t Pos = RedzoneSize;
901  // Replace Alloca instructions with base+offset.
902  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
903    AllocaInst *AI = AllocaVec[i];
904    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
905    StringRef Name = AI->getName();
906    StackDescription << Pos << " " << SizeInBytes << " "
907                     << Name.size() << " " << Name << " ";
908    uint64_t AlignedSize = getAlignedAllocaSize(AI);
909    assert((AlignedSize % RedzoneSize) == 0);
910    AI->replaceAllUsesWith(
911        IRB.CreateIntToPtr(
912            IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy, Pos)),
913            AI->getType()));
914    Pos += AlignedSize + RedzoneSize;
915  }
916  assert(Pos == LocalStackSize);
917
918  // Write the Magic value and the frame description constant to the redzone.
919  Value *BasePlus0 = IRB.CreateIntToPtr(LocalStackBase, IntptrPtrTy);
920  IRB.CreateStore(ConstantInt::get(IntptrTy, kCurrentStackFrameMagic),
921                  BasePlus0);
922  Value *BasePlus1 = IRB.CreateAdd(LocalStackBase,
923                                   ConstantInt::get(IntptrTy, LongSize/8));
924  BasePlus1 = IRB.CreateIntToPtr(BasePlus1, IntptrPtrTy);
925  Value *Description = IRB.CreatePointerCast(
926      createPrivateGlobalForString(M, StackDescription.str()),
927      IntptrTy);
928  IRB.CreateStore(Description, BasePlus1);
929
930  // Poison the stack redzones at the entry.
931  Value *ShadowBase = memToShadow(LocalStackBase, IRB);
932  PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRB, ShadowBase, true);
933
934  Value *AsanStackFreeFunc = NULL;
935  if (DoStackMalloc) {
936    AsanStackFreeFunc = M.getOrInsertFunction(
937        kAsanStackFreeName, IRB.getVoidTy(),
938        IntptrTy, IntptrTy, IntptrTy, NULL);
939  }
940
941  // Unpoison the stack before all ret instructions.
942  for (size_t i = 0, n = RetVec.size(); i < n; i++) {
943    Instruction *Ret = RetVec[i];
944    IRBuilder<> IRBRet(Ret);
945
946    // Mark the current frame as retired.
947    IRBRet.CreateStore(ConstantInt::get(IntptrTy, kRetiredStackFrameMagic),
948                       BasePlus0);
949    // Unpoison the stack.
950    PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRBRet, ShadowBase, false);
951
952    if (DoStackMalloc) {
953      IRBRet.CreateCall3(AsanStackFreeFunc, LocalStackBase,
954                         ConstantInt::get(IntptrTy, LocalStackSize),
955                         OrigStackBase);
956    }
957  }
958
959  if (ClDebugStack) {
960    DEBUG(dbgs() << F);
961  }
962
963  return true;
964}
965
966BlackList::BlackList(const std::string &Path) {
967  Functions = NULL;
968  const char *kFunPrefix = "fun:";
969  if (!ClBlackListFile.size()) return;
970  std::string Fun;
971
972  OwningPtr<MemoryBuffer> File;
973  if (error_code EC = MemoryBuffer::getFile(ClBlackListFile.c_str(), File)) {
974    report_fatal_error("Can't open blacklist file " + ClBlackListFile + ": " +
975                       EC.message());
976  }
977  MemoryBuffer *Buff = File.take();
978  const char *Data = Buff->getBufferStart();
979  size_t DataLen = Buff->getBufferSize();
980  SmallVector<StringRef, 16> Lines;
981  SplitString(StringRef(Data, DataLen), Lines, "\n\r");
982  for (size_t i = 0, numLines = Lines.size(); i < numLines; i++) {
983    if (Lines[i].startswith(kFunPrefix)) {
984      std::string ThisFunc = Lines[i].substr(strlen(kFunPrefix));
985      std::string ThisFuncRE;
986      // add ThisFunc replacing * with .*
987      for (size_t j = 0, n = ThisFunc.size(); j < n; j++) {
988        if (ThisFunc[j] == '*')
989          ThisFuncRE += '.';
990        ThisFuncRE += ThisFunc[j];
991      }
992      // Check that the regexp is valid.
993      Regex CheckRE(ThisFuncRE);
994      std::string Error;
995      if (!CheckRE.isValid(Error))
996        report_fatal_error("malformed blacklist regex: " + ThisFunc +
997                           ": " + Error);
998      // Append to the final regexp.
999      if (Fun.size())
1000        Fun += "|";
1001      Fun += ThisFuncRE;
1002    }
1003  }
1004  if (Fun.size()) {
1005    Functions = new Regex(Fun);
1006  }
1007}
1008
1009bool BlackList::isIn(const Function &F) {
1010  if (Functions) {
1011    bool Res = Functions->match(F.getName());
1012    return Res;
1013  }
1014  return false;
1015}
1016