AddressSanitizer.cpp revision 56139bc493790612ee6281630678e293be6b2eb2 
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 "FunctionBlackList.h"
19#include "llvm/Function.h"
20#include "llvm/IRBuilder.h"
21#include "llvm/IntrinsicInst.h"
22#include "llvm/LLVMContext.h"
23#include "llvm/Module.h"
24#include "llvm/Type.h"
25#include "llvm/ADT/ArrayRef.h"
26#include "llvm/ADT/OwningPtr.h"
27#include "llvm/ADT/SmallSet.h"
28#include "llvm/ADT/SmallString.h"
29#include "llvm/ADT/SmallVector.h"
30#include "llvm/ADT/StringExtras.h"
31#include "llvm/ADT/Triple.h"
32#include "llvm/Support/CommandLine.h"
33#include "llvm/Support/DataTypes.h"
34#include "llvm/Support/Debug.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
43#include <string>
44#include <algorithm>
45
46using namespace llvm;
47
48static const uint64_t kDefaultShadowScale = 3;
49static const uint64_t kDefaultShadowOffset32 = 1ULL << 29;
50static const uint64_t kDefaultShadowOffset64 = 1ULL << 44;
51static const uint64_t kDefaultShadowOffsetAndroid = 0;
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 *kAsanHandleNoReturnName = "__asan_handle_no_return";
65static const char *kAsanMappingOffsetName = "__asan_mapping_offset";
66static const char *kAsanMappingScaleName = "__asan_mapping_scale";
67static const char *kAsanStackMallocName = "__asan_stack_malloc";
68static const char *kAsanStackFreeName = "__asan_stack_free";
69
70static const int kAsanStackLeftRedzoneMagic = 0xf1;
71static const int kAsanStackMidRedzoneMagic = 0xf2;
72static const int kAsanStackRightRedzoneMagic = 0xf3;
73static const int kAsanStackPartialRedzoneMagic = 0xf4;
74
75// Command-line flags.
76
77// This flag may need to be replaced with -f[no-]asan-reads.
78static cl::opt<bool> ClInstrumentReads("asan-instrument-reads",
79       cl::desc("instrument read instructions"), cl::Hidden, cl::init(true));
80static cl::opt<bool> ClInstrumentWrites("asan-instrument-writes",
81       cl::desc("instrument write instructions"), cl::Hidden, cl::init(true));
82static cl::opt<bool> ClInstrumentAtomics("asan-instrument-atomics",
83       cl::desc("instrument atomic instructions (rmw, cmpxchg)"),
84       cl::Hidden, cl::init(true));
85// This flags limits the number of instructions to be instrumented
86// in any given BB. Normally, this should be set to unlimited (INT_MAX),
87// but due to http://llvm.org/bugs/show_bug.cgi?id=12652 we temporary
88// set it to 10000.
89static cl::opt<int> ClMaxInsnsToInstrumentPerBB("asan-max-ins-per-bb",
90       cl::init(10000),
91       cl::desc("maximal number of instructions to instrument in any given BB"),
92       cl::Hidden);
93// This flag may need to be replaced with -f[no]asan-stack.
94static cl::opt<bool> ClStack("asan-stack",
95       cl::desc("Handle stack memory"), cl::Hidden, cl::init(true));
96// This flag may need to be replaced with -f[no]asan-use-after-return.
97static cl::opt<bool> ClUseAfterReturn("asan-use-after-return",
98       cl::desc("Check return-after-free"), cl::Hidden, cl::init(false));
99// This flag may need to be replaced with -f[no]asan-globals.
100static cl::opt<bool> ClGlobals("asan-globals",
101       cl::desc("Handle global objects"), cl::Hidden, cl::init(true));
102static cl::opt<bool> ClMemIntrin("asan-memintrin",
103       cl::desc("Handle memset/memcpy/memmove"), cl::Hidden, cl::init(true));
104// This flag may need to be replaced with -fasan-blacklist.
105static cl::opt<std::string>  ClBlackListFile("asan-blacklist",
106       cl::desc("File containing the list of functions to ignore "
107                "during instrumentation"), cl::Hidden);
108
109// These flags allow to change the shadow mapping.
110// The shadow mapping looks like
111//    Shadow = (Mem >> scale) + (1 << offset_log)
112static cl::opt<int> ClMappingScale("asan-mapping-scale",
113       cl::desc("scale of asan shadow mapping"), cl::Hidden, cl::init(0));
114static cl::opt<int> ClMappingOffsetLog("asan-mapping-offset-log",
115       cl::desc("offset of asan shadow mapping"), cl::Hidden, cl::init(-1));
116
117// Optimization flags. Not user visible, used mostly for testing
118// and benchmarking the tool.
119static cl::opt<bool> ClOpt("asan-opt",
120       cl::desc("Optimize instrumentation"), cl::Hidden, cl::init(true));
121static cl::opt<bool> ClOptSameTemp("asan-opt-same-temp",
122       cl::desc("Instrument the same temp just once"), cl::Hidden,
123       cl::init(true));
124static cl::opt<bool> ClOptGlobals("asan-opt-globals",
125       cl::desc("Don't instrument scalar globals"), cl::Hidden, cl::init(true));
126
127// Debug flags.
128static cl::opt<int> ClDebug("asan-debug", cl::desc("debug"), cl::Hidden,
129                            cl::init(0));
130static cl::opt<int> ClDebugStack("asan-debug-stack", cl::desc("debug stack"),
131                                 cl::Hidden, cl::init(0));
132static cl::opt<std::string> ClDebugFunc("asan-debug-func",
133                                        cl::Hidden, cl::desc("Debug func"));
134static cl::opt<int> ClDebugMin("asan-debug-min", cl::desc("Debug min inst"),
135                               cl::Hidden, cl::init(-1));
136static cl::opt<int> ClDebugMax("asan-debug-max", cl::desc("Debug man inst"),
137                               cl::Hidden, cl::init(-1));
138
139namespace {
140
141/// AddressSanitizer: instrument the code in module to find memory bugs.
142struct AddressSanitizer : public ModulePass {
143  AddressSanitizer();
144  virtual const char *getPassName() const;
145  void instrumentMop(Instruction *I);
146  void instrumentAddress(Instruction *OrigIns, IRBuilder<> &IRB,
147                         Value *Addr, uint32_t TypeSize, bool IsWrite);
148  Instruction *generateCrashCode(IRBuilder<> &IRB, Value *Addr,
149                                 bool IsWrite, uint32_t TypeSize);
150  bool instrumentMemIntrinsic(MemIntrinsic *MI);
151  void instrumentMemIntrinsicParam(Instruction *OrigIns, Value *Addr,
152                                  Value *Size,
153                                   Instruction *InsertBefore, bool IsWrite);
154  Value *memToShadow(Value *Shadow, IRBuilder<> &IRB);
155  bool handleFunction(Module &M, Function &F);
156  bool maybeInsertAsanInitAtFunctionEntry(Function &F);
157  bool poisonStackInFunction(Module &M, Function &F);
158  virtual bool runOnModule(Module &M);
159  bool insertGlobalRedzones(Module &M);
160  static char ID;  // Pass identification, replacement for typeid
161
162 private:
163
164  uint64_t getAllocaSizeInBytes(AllocaInst *AI) {
165    Type *Ty = AI->getAllocatedType();
166    uint64_t SizeInBytes = TD->getTypeAllocSize(Ty);
167    return SizeInBytes;
168  }
169  uint64_t getAlignedSize(uint64_t SizeInBytes) {
170    return ((SizeInBytes + RedzoneSize - 1)
171            / RedzoneSize) * RedzoneSize;
172  }
173  uint64_t getAlignedAllocaSize(AllocaInst *AI) {
174    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
175    return getAlignedSize(SizeInBytes);
176  }
177
178  Function *checkInterfaceFunction(Constant *FuncOrBitcast);
179  void PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec, IRBuilder<> IRB,
180                   Value *ShadowBase, bool DoPoison);
181  bool LooksLikeCodeInBug11395(Instruction *I);
182
183  Module      *CurrentModule;
184  LLVMContext *C;
185  TargetData *TD;
186  uint64_t MappingOffset;
187  int MappingScale;
188  size_t RedzoneSize;
189  int LongSize;
190  Type *IntptrTy;
191  Type *IntptrPtrTy;
192  Function *AsanCtorFunction;
193  Function *AsanInitFunction;
194  Instruction *CtorInsertBefore;
195  OwningPtr<FunctionBlackList> BL;
196};
197}  // namespace
198
199char AddressSanitizer::ID = 0;
200INITIALIZE_PASS(AddressSanitizer, "asan",
201    "AddressSanitizer: detects use-after-free and out-of-bounds bugs.",
202    false, false)
203AddressSanitizer::AddressSanitizer() : ModulePass(ID) { }
204ModulePass *llvm::createAddressSanitizerPass() {
205  return new AddressSanitizer();
206}
207
208const char *AddressSanitizer::getPassName() const {
209  return "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 = ConstantDataArray::getString(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//   Cmp
223//   Tail
224// After:
225//   Head
226//   if (Cmp)
227//     ThenBlock
228//   Tail
229//
230// Returns the ThenBlock's terminator.
231static BranchInst *splitBlockAndInsertIfThen(Value *Cmp) {
232  Instruction *SplitBefore = cast<Instruction>(Cmp)->getNextNode();
233  BasicBlock *Head = SplitBefore->getParent();
234  BasicBlock *Tail = Head->splitBasicBlock(SplitBefore);
235  TerminatorInst *HeadOldTerm = Head->getTerminator();
236  LLVMContext &C = Head->getParent()->getParent()->getContext();
237  BasicBlock *ThenBlock = BasicBlock::Create(C, "", Head->getParent());
238  BranchInst *HeadNewTerm =
239    BranchInst::Create(/*ifTrue*/ThenBlock, /*ifFalse*/Tail, Cmp);
240  ReplaceInstWithInst(HeadOldTerm, HeadNewTerm);
241
242  BranchInst *CheckTerm = BranchInst::Create(Tail, ThenBlock);
243  return CheckTerm;
244}
245
246Value *AddressSanitizer::memToShadow(Value *Shadow, IRBuilder<> &IRB) {
247  // Shadow >> scale
248  Shadow = IRB.CreateLShr(Shadow, MappingScale);
249  if (MappingOffset == 0)
250    return Shadow;
251  // (Shadow >> scale) | offset
252  return IRB.CreateOr(Shadow, ConstantInt::get(IntptrTy,
253                                               MappingOffset));
254}
255
256void AddressSanitizer::instrumentMemIntrinsicParam(Instruction *OrigIns,
257    Value *Addr, Value *Size, Instruction *InsertBefore, bool IsWrite) {
258  // Check the first byte.
259  {
260    IRBuilder<> IRB(InsertBefore);
261    instrumentAddress(OrigIns, IRB, Addr, 8, IsWrite);
262  }
263  // Check the last byte.
264  {
265    IRBuilder<> IRB(InsertBefore);
266    Value *SizeMinusOne = IRB.CreateSub(
267        Size, ConstantInt::get(Size->getType(), 1));
268    SizeMinusOne = IRB.CreateIntCast(SizeMinusOne, IntptrTy, false);
269    Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
270    Value *AddrPlusSizeMinisOne = IRB.CreateAdd(AddrLong, SizeMinusOne);
271    instrumentAddress(OrigIns, IRB, AddrPlusSizeMinisOne, 8, IsWrite);
272  }
273}
274
275// Instrument memset/memmove/memcpy
276bool AddressSanitizer::instrumentMemIntrinsic(MemIntrinsic *MI) {
277  Value *Dst = MI->getDest();
278  MemTransferInst *MemTran = dyn_cast<MemTransferInst>(MI);
279  Value *Src = MemTran ? MemTran->getSource() : NULL;
280  Value *Length = MI->getLength();
281
282  Constant *ConstLength = dyn_cast<Constant>(Length);
283  Instruction *InsertBefore = MI;
284  if (ConstLength) {
285    if (ConstLength->isNullValue()) return false;
286  } else {
287    // The size is not a constant so it could be zero -- check at run-time.
288    IRBuilder<> IRB(InsertBefore);
289
290    Value *Cmp = IRB.CreateICmpNE(Length,
291                                  Constant::getNullValue(Length->getType()));
292    InsertBefore = splitBlockAndInsertIfThen(Cmp);
293  }
294
295  instrumentMemIntrinsicParam(MI, Dst, Length, InsertBefore, true);
296  if (Src)
297    instrumentMemIntrinsicParam(MI, Src, Length, InsertBefore, false);
298  return true;
299}
300
301// If I is an interesting memory access, return the PointerOperand
302// and set IsWrite. Otherwise return NULL.
303static Value *isInterestingMemoryAccess(Instruction *I, bool *IsWrite) {
304  if (LoadInst *LI = dyn_cast<LoadInst>(I)) {
305    if (!ClInstrumentReads) return NULL;
306    *IsWrite = false;
307    return LI->getPointerOperand();
308  }
309  if (StoreInst *SI = dyn_cast<StoreInst>(I)) {
310    if (!ClInstrumentWrites) return NULL;
311    *IsWrite = true;
312    return SI->getPointerOperand();
313  }
314  if (AtomicRMWInst *RMW = dyn_cast<AtomicRMWInst>(I)) {
315    if (!ClInstrumentAtomics) return NULL;
316    *IsWrite = true;
317    return RMW->getPointerOperand();
318  }
319  if (AtomicCmpXchgInst *XCHG = dyn_cast<AtomicCmpXchgInst>(I)) {
320    if (!ClInstrumentAtomics) return NULL;
321    *IsWrite = true;
322    return XCHG->getPointerOperand();
323  }
324  return NULL;
325}
326
327void AddressSanitizer::instrumentMop(Instruction *I) {
328  bool IsWrite;
329  Value *Addr = isInterestingMemoryAccess(I, &IsWrite);
330  assert(Addr);
331  if (ClOpt && ClOptGlobals && isa<GlobalVariable>(Addr)) {
332    // We are accessing a global scalar variable. Nothing to catch here.
333    return;
334  }
335  Type *OrigPtrTy = Addr->getType();
336  Type *OrigTy = cast<PointerType>(OrigPtrTy)->getElementType();
337
338  assert(OrigTy->isSized());
339  uint32_t TypeSize = TD->getTypeStoreSizeInBits(OrigTy);
340
341  if (TypeSize != 8  && TypeSize != 16 &&
342      TypeSize != 32 && TypeSize != 64 && TypeSize != 128) {
343    // Ignore all unusual sizes.
344    return;
345  }
346
347  IRBuilder<> IRB(I);
348  instrumentAddress(I, IRB, Addr, TypeSize, IsWrite);
349}
350
351// Validate the result of Module::getOrInsertFunction called for an interface
352// function of AddressSanitizer. If the instrumented module defines a function
353// with the same name, their prototypes must match, otherwise
354// getOrInsertFunction returns a bitcast.
355Function *AddressSanitizer::checkInterfaceFunction(Constant *FuncOrBitcast) {
356  if (isa<Function>(FuncOrBitcast)) return cast<Function>(FuncOrBitcast);
357  FuncOrBitcast->dump();
358  report_fatal_error("trying to redefine an AddressSanitizer "
359                     "interface function");
360}
361
362Instruction *AddressSanitizer::generateCrashCode(
363    IRBuilder<> &IRB, Value *Addr, bool IsWrite, uint32_t TypeSize) {
364  // IsWrite and TypeSize are encoded in the function name.
365  std::string FunctionName = std::string(kAsanReportErrorTemplate) +
366      (IsWrite ? "store" : "load") + itostr(TypeSize / 8);
367  Value *ReportWarningFunc = CurrentModule->getOrInsertFunction(
368      FunctionName, IRB.getVoidTy(), IntptrTy, NULL);
369  CallInst *Call = IRB.CreateCall(ReportWarningFunc, Addr);
370  Call->setDoesNotReturn();
371  return Call;
372}
373
374void AddressSanitizer::instrumentAddress(Instruction *OrigIns,
375                                         IRBuilder<> &IRB, Value *Addr,
376                                         uint32_t TypeSize, bool IsWrite) {
377  Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
378
379  Type *ShadowTy  = IntegerType::get(
380      *C, std::max(8U, TypeSize >> MappingScale));
381  Type *ShadowPtrTy = PointerType::get(ShadowTy, 0);
382  Value *ShadowPtr = memToShadow(AddrLong, IRB);
383  Value *CmpVal = Constant::getNullValue(ShadowTy);
384  Value *ShadowValue = IRB.CreateLoad(
385      IRB.CreateIntToPtr(ShadowPtr, ShadowPtrTy));
386
387  Value *Cmp = IRB.CreateICmpNE(ShadowValue, CmpVal);
388
389  Instruction *CheckTerm = splitBlockAndInsertIfThen(Cmp);
390  IRBuilder<> IRB2(CheckTerm);
391
392  size_t Granularity = 1 << MappingScale;
393  if (TypeSize < 8 * Granularity) {
394    // Addr & (Granularity - 1)
395    Value *LastAccessedByte = IRB2.CreateAnd(
396        AddrLong, ConstantInt::get(IntptrTy, Granularity - 1));
397    // (Addr & (Granularity - 1)) + size - 1
398    if (TypeSize / 8 > 1)
399      LastAccessedByte = IRB2.CreateAdd(
400          LastAccessedByte, ConstantInt::get(IntptrTy, TypeSize / 8 - 1));
401    // (uint8_t) ((Addr & (Granularity-1)) + size - 1)
402    LastAccessedByte = IRB2.CreateIntCast(
403        LastAccessedByte, IRB.getInt8Ty(), false);
404    // ((uint8_t) ((Addr & (Granularity-1)) + size - 1)) >= ShadowValue
405    Value *Cmp2 = IRB2.CreateICmpSGE(LastAccessedByte, ShadowValue);
406
407    CheckTerm = splitBlockAndInsertIfThen(Cmp2);
408  }
409
410  IRBuilder<> IRB1(CheckTerm);
411  Instruction *Crash = generateCrashCode(IRB1, AddrLong, IsWrite, TypeSize);
412  Crash->setDebugLoc(OrigIns->getDebugLoc());
413  ReplaceInstWithInst(CheckTerm, new UnreachableInst(*C));
414}
415
416// This function replaces all global variables with new variables that have
417// trailing redzones. It also creates a function that poisons
418// redzones and inserts this function into llvm.global_ctors.
419bool AddressSanitizer::insertGlobalRedzones(Module &M) {
420  SmallVector<GlobalVariable *, 16> GlobalsToChange;
421
422  for (Module::GlobalListType::iterator G = M.getGlobalList().begin(),
423       E = M.getGlobalList().end(); G != E; ++G) {
424    Type *Ty = cast<PointerType>(G->getType())->getElementType();
425    DEBUG(dbgs() << "GLOBAL: " << *G);
426
427    if (!Ty->isSized()) continue;
428    if (!G->hasInitializer()) continue;
429    // Touch only those globals that will not be defined in other modules.
430    // Don't handle ODR type linkages since other modules may be built w/o asan.
431    if (G->getLinkage() != GlobalVariable::ExternalLinkage &&
432        G->getLinkage() != GlobalVariable::PrivateLinkage &&
433        G->getLinkage() != GlobalVariable::InternalLinkage)
434      continue;
435    // Two problems with thread-locals:
436    //   - The address of the main thread's copy can't be computed at link-time.
437    //   - Need to poison all copies, not just the main thread's one.
438    if (G->isThreadLocal())
439      continue;
440    // For now, just ignore this Alloca if the alignment is large.
441    if (G->getAlignment() > RedzoneSize) continue;
442
443    // Ignore all the globals with the names starting with "\01L_OBJC_".
444    // Many of those are put into the .cstring section. The linker compresses
445    // that section by removing the spare \0s after the string terminator, so
446    // our redzones get broken.
447    if ((G->getName().find("\01L_OBJC_") == 0) ||
448        (G->getName().find("\01l_OBJC_") == 0)) {
449      DEBUG(dbgs() << "Ignoring \\01L_OBJC_* global: " << *G);
450      continue;
451    }
452
453    if (G->hasSection()) {
454      StringRef Section(G->getSection());
455      // Ignore the globals from the __OBJC section. The ObjC runtime assumes
456      // those conform to /usr/lib/objc/runtime.h, so we can't add redzones to
457      // them.
458      if ((Section.find("__OBJC,") == 0) ||
459          (Section.find("__DATA, __objc_") == 0)) {
460        DEBUG(dbgs() << "Ignoring ObjC runtime global: " << *G);
461        continue;
462      }
463      // See http://code.google.com/p/address-sanitizer/issues/detail?id=32
464      // Constant CFString instances are compiled in the following way:
465      //  -- the string buffer is emitted into
466      //     __TEXT,__cstring,cstring_literals
467      //  -- the constant NSConstantString structure referencing that buffer
468      //     is placed into __DATA,__cfstring
469      // Therefore there's no point in placing redzones into __DATA,__cfstring.
470      // Moreover, it causes the linker to crash on OS X 10.7
471      if (Section.find("__DATA,__cfstring") == 0) {
472        DEBUG(dbgs() << "Ignoring CFString: " << *G);
473        continue;
474      }
475    }
476
477    GlobalsToChange.push_back(G);
478  }
479
480  size_t n = GlobalsToChange.size();
481  if (n == 0) return false;
482
483  // A global is described by a structure
484  //   size_t beg;
485  //   size_t size;
486  //   size_t size_with_redzone;
487  //   const char *name;
488  // We initialize an array of such structures and pass it to a run-time call.
489  StructType *GlobalStructTy = StructType::get(IntptrTy, IntptrTy,
490                                               IntptrTy, IntptrTy, NULL);
491  SmallVector<Constant *, 16> Initializers(n);
492
493  IRBuilder<> IRB(CtorInsertBefore);
494
495  for (size_t i = 0; i < n; i++) {
496    GlobalVariable *G = GlobalsToChange[i];
497    PointerType *PtrTy = cast<PointerType>(G->getType());
498    Type *Ty = PtrTy->getElementType();
499    uint64_t SizeInBytes = TD->getTypeAllocSize(Ty);
500    uint64_t RightRedzoneSize = RedzoneSize +
501        (RedzoneSize - (SizeInBytes % RedzoneSize));
502    Type *RightRedZoneTy = ArrayType::get(IRB.getInt8Ty(), RightRedzoneSize);
503
504    StructType *NewTy = StructType::get(Ty, RightRedZoneTy, NULL);
505    Constant *NewInitializer = ConstantStruct::get(
506        NewTy, G->getInitializer(),
507        Constant::getNullValue(RightRedZoneTy), NULL);
508
509    SmallString<2048> DescriptionOfGlobal = G->getName();
510    DescriptionOfGlobal += " (";
511    DescriptionOfGlobal += M.getModuleIdentifier();
512    DescriptionOfGlobal += ")";
513    GlobalVariable *Name = createPrivateGlobalForString(M, DescriptionOfGlobal);
514
515    // Create a new global variable with enough space for a redzone.
516    GlobalVariable *NewGlobal = new GlobalVariable(
517        M, NewTy, G->isConstant(), G->getLinkage(),
518        NewInitializer, "", G, G->getThreadLocalMode());
519    NewGlobal->copyAttributesFrom(G);
520    NewGlobal->setAlignment(RedzoneSize);
521
522    Value *Indices2[2];
523    Indices2[0] = IRB.getInt32(0);
524    Indices2[1] = IRB.getInt32(0);
525
526    G->replaceAllUsesWith(
527        ConstantExpr::getGetElementPtr(NewGlobal, Indices2, true));
528    NewGlobal->takeName(G);
529    G->eraseFromParent();
530
531    Initializers[i] = ConstantStruct::get(
532        GlobalStructTy,
533        ConstantExpr::getPointerCast(NewGlobal, IntptrTy),
534        ConstantInt::get(IntptrTy, SizeInBytes),
535        ConstantInt::get(IntptrTy, SizeInBytes + RightRedzoneSize),
536        ConstantExpr::getPointerCast(Name, IntptrTy),
537        NULL);
538    DEBUG(dbgs() << "NEW GLOBAL:\n" << *NewGlobal);
539  }
540
541  ArrayType *ArrayOfGlobalStructTy = ArrayType::get(GlobalStructTy, n);
542  GlobalVariable *AllGlobals = new GlobalVariable(
543      M, ArrayOfGlobalStructTy, false, GlobalVariable::PrivateLinkage,
544      ConstantArray::get(ArrayOfGlobalStructTy, Initializers), "");
545
546  Function *AsanRegisterGlobals = checkInterfaceFunction(M.getOrInsertFunction(
547      kAsanRegisterGlobalsName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
548  AsanRegisterGlobals->setLinkage(Function::ExternalLinkage);
549
550  IRB.CreateCall2(AsanRegisterGlobals,
551                  IRB.CreatePointerCast(AllGlobals, IntptrTy),
552                  ConstantInt::get(IntptrTy, n));
553
554  // We also need to unregister globals at the end, e.g. when a shared library
555  // gets closed.
556  Function *AsanDtorFunction = Function::Create(
557      FunctionType::get(Type::getVoidTy(*C), false),
558      GlobalValue::InternalLinkage, kAsanModuleDtorName, &M);
559  BasicBlock *AsanDtorBB = BasicBlock::Create(*C, "", AsanDtorFunction);
560  IRBuilder<> IRB_Dtor(ReturnInst::Create(*C, AsanDtorBB));
561  Function *AsanUnregisterGlobals =
562      checkInterfaceFunction(M.getOrInsertFunction(
563          kAsanUnregisterGlobalsName,
564          IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
565  AsanUnregisterGlobals->setLinkage(Function::ExternalLinkage);
566
567  IRB_Dtor.CreateCall2(AsanUnregisterGlobals,
568                       IRB.CreatePointerCast(AllGlobals, IntptrTy),
569                       ConstantInt::get(IntptrTy, n));
570  appendToGlobalDtors(M, AsanDtorFunction, kAsanCtorAndCtorPriority);
571
572  DEBUG(dbgs() << M);
573  return true;
574}
575
576// virtual
577bool AddressSanitizer::runOnModule(Module &M) {
578  // Initialize the private fields. No one has accessed them before.
579  TD = getAnalysisIfAvailable<TargetData>();
580  if (!TD)
581    return false;
582  BL.reset(new FunctionBlackList(ClBlackListFile));
583
584  CurrentModule = &M;
585  C = &(M.getContext());
586  LongSize = TD->getPointerSizeInBits();
587  IntptrTy = Type::getIntNTy(*C, LongSize);
588  IntptrPtrTy = PointerType::get(IntptrTy, 0);
589
590  AsanCtorFunction = Function::Create(
591      FunctionType::get(Type::getVoidTy(*C), false),
592      GlobalValue::InternalLinkage, kAsanModuleCtorName, &M);
593  BasicBlock *AsanCtorBB = BasicBlock::Create(*C, "", AsanCtorFunction);
594  CtorInsertBefore = ReturnInst::Create(*C, AsanCtorBB);
595
596  // call __asan_init in the module ctor.
597  IRBuilder<> IRB(CtorInsertBefore);
598  AsanInitFunction = checkInterfaceFunction(
599      M.getOrInsertFunction(kAsanInitName, IRB.getVoidTy(), NULL));
600  AsanInitFunction->setLinkage(Function::ExternalLinkage);
601  IRB.CreateCall(AsanInitFunction);
602
603  llvm::Triple targetTriple(M.getTargetTriple());
604  bool isAndroid = targetTriple.getEnvironment() == llvm::Triple::ANDROIDEABI;
605
606  MappingOffset = isAndroid ? kDefaultShadowOffsetAndroid :
607    (LongSize == 32 ? kDefaultShadowOffset32 : kDefaultShadowOffset64);
608  if (ClMappingOffsetLog >= 0) {
609    if (ClMappingOffsetLog == 0) {
610      // special case
611      MappingOffset = 0;
612    } else {
613      MappingOffset = 1ULL << ClMappingOffsetLog;
614    }
615  }
616  MappingScale = kDefaultShadowScale;
617  if (ClMappingScale) {
618    MappingScale = ClMappingScale;
619  }
620  // Redzone used for stack and globals is at least 32 bytes.
621  // For scales 6 and 7, the redzone has to be 64 and 128 bytes respectively.
622  RedzoneSize = std::max(32, (int)(1 << MappingScale));
623
624  bool Res = false;
625
626  if (ClGlobals)
627    Res |= insertGlobalRedzones(M);
628
629  if (ClMappingOffsetLog >= 0) {
630    // Tell the run-time the current values of mapping offset and scale.
631    GlobalValue *asan_mapping_offset =
632        new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
633                       ConstantInt::get(IntptrTy, MappingOffset),
634                       kAsanMappingOffsetName);
635    // Read the global, otherwise it may be optimized away.
636    IRB.CreateLoad(asan_mapping_offset, true);
637  }
638  if (ClMappingScale) {
639    GlobalValue *asan_mapping_scale =
640        new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
641                           ConstantInt::get(IntptrTy, MappingScale),
642                           kAsanMappingScaleName);
643    // Read the global, otherwise it may be optimized away.
644    IRB.CreateLoad(asan_mapping_scale, true);
645  }
646
647
648  for (Module::iterator F = M.begin(), E = M.end(); F != E; ++F) {
649    if (F->isDeclaration()) continue;
650    Res |= handleFunction(M, *F);
651  }
652
653  appendToGlobalCtors(M, AsanCtorFunction, kAsanCtorAndCtorPriority);
654
655  return Res;
656}
657
658bool AddressSanitizer::maybeInsertAsanInitAtFunctionEntry(Function &F) {
659  // For each NSObject descendant having a +load method, this method is invoked
660  // by the ObjC runtime before any of the static constructors is called.
661  // Therefore we need to instrument such methods with a call to __asan_init
662  // at the beginning in order to initialize our runtime before any access to
663  // the shadow memory.
664  // We cannot just ignore these methods, because they may call other
665  // instrumented functions.
666  if (F.getName().find(" load]") != std::string::npos) {
667    IRBuilder<> IRB(F.begin()->begin());
668    IRB.CreateCall(AsanInitFunction);
669    return true;
670  }
671  return false;
672}
673
674bool AddressSanitizer::handleFunction(Module &M, Function &F) {
675  if (BL->isIn(F)) return false;
676  if (&F == AsanCtorFunction) return false;
677
678  // If needed, insert __asan_init before checking for AddressSafety attr.
679  maybeInsertAsanInitAtFunctionEntry(F);
680
681  if (!F.hasFnAttr(Attribute::AddressSafety)) return false;
682
683  if (!ClDebugFunc.empty() && ClDebugFunc != F.getName())
684    return false;
685  // We want to instrument every address only once per basic block
686  // (unless there are calls between uses).
687  SmallSet<Value*, 16> TempsToInstrument;
688  SmallVector<Instruction*, 16> ToInstrument;
689  SmallVector<Instruction*, 8> NoReturnCalls;
690  bool IsWrite;
691
692  // Fill the set of memory operations to instrument.
693  for (Function::iterator FI = F.begin(), FE = F.end();
694       FI != FE; ++FI) {
695    TempsToInstrument.clear();
696    int NumInsnsPerBB = 0;
697    for (BasicBlock::iterator BI = FI->begin(), BE = FI->end();
698         BI != BE; ++BI) {
699      if (LooksLikeCodeInBug11395(BI)) return false;
700      if (Value *Addr = isInterestingMemoryAccess(BI, &IsWrite)) {
701        if (ClOpt && ClOptSameTemp) {
702          if (!TempsToInstrument.insert(Addr))
703            continue;  // We've seen this temp in the current BB.
704        }
705      } else if (isa<MemIntrinsic>(BI) && ClMemIntrin) {
706        // ok, take it.
707      } else {
708        if (CallInst *CI = dyn_cast<CallInst>(BI)) {
709          // A call inside BB.
710          TempsToInstrument.clear();
711          if (CI->doesNotReturn()) {
712            NoReturnCalls.push_back(CI);
713          }
714        }
715        continue;
716      }
717      ToInstrument.push_back(BI);
718      NumInsnsPerBB++;
719      if (NumInsnsPerBB >= ClMaxInsnsToInstrumentPerBB)
720        break;
721    }
722  }
723
724  // Instrument.
725  int NumInstrumented = 0;
726  for (size_t i = 0, n = ToInstrument.size(); i != n; i++) {
727    Instruction *Inst = ToInstrument[i];
728    if (ClDebugMin < 0 || ClDebugMax < 0 ||
729        (NumInstrumented >= ClDebugMin && NumInstrumented <= ClDebugMax)) {
730      if (isInterestingMemoryAccess(Inst, &IsWrite))
731        instrumentMop(Inst);
732      else
733        instrumentMemIntrinsic(cast<MemIntrinsic>(Inst));
734    }
735    NumInstrumented++;
736  }
737
738  DEBUG(dbgs() << F);
739
740  bool ChangedStack = poisonStackInFunction(M, F);
741
742  // We must unpoison the stack before every NoReturn call (throw, _exit, etc).
743  // See e.g. http://code.google.com/p/address-sanitizer/issues/detail?id=37
744  for (size_t i = 0, n = NoReturnCalls.size(); i != n; i++) {
745    Instruction *CI = NoReturnCalls[i];
746    IRBuilder<> IRB(CI);
747    IRB.CreateCall(M.getOrInsertFunction(kAsanHandleNoReturnName,
748                                         IRB.getVoidTy(), NULL));
749  }
750
751  return NumInstrumented > 0 || ChangedStack || !NoReturnCalls.empty();
752}
753
754static uint64_t ValueForPoison(uint64_t PoisonByte, size_t ShadowRedzoneSize) {
755  if (ShadowRedzoneSize == 1) return PoisonByte;
756  if (ShadowRedzoneSize == 2) return (PoisonByte << 8) + PoisonByte;
757  if (ShadowRedzoneSize == 4)
758    return (PoisonByte << 24) + (PoisonByte << 16) +
759        (PoisonByte << 8) + (PoisonByte);
760  llvm_unreachable("ShadowRedzoneSize is either 1, 2 or 4");
761}
762
763static void PoisonShadowPartialRightRedzone(uint8_t *Shadow,
764                                            size_t Size,
765                                            size_t RedzoneSize,
766                                            size_t ShadowGranularity,
767                                            uint8_t Magic) {
768  for (size_t i = 0; i < RedzoneSize;
769       i+= ShadowGranularity, Shadow++) {
770    if (i + ShadowGranularity <= Size) {
771      *Shadow = 0;  // fully addressable
772    } else if (i >= Size) {
773      *Shadow = Magic;  // unaddressable
774    } else {
775      *Shadow = Size - i;  // first Size-i bytes are addressable
776    }
777  }
778}
779
780void AddressSanitizer::PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec,
781                                   IRBuilder<> IRB,
782                                   Value *ShadowBase, bool DoPoison) {
783  size_t ShadowRZSize = RedzoneSize >> MappingScale;
784  assert(ShadowRZSize >= 1 && ShadowRZSize <= 4);
785  Type *RZTy = Type::getIntNTy(*C, ShadowRZSize * 8);
786  Type *RZPtrTy = PointerType::get(RZTy, 0);
787
788  Value *PoisonLeft  = ConstantInt::get(RZTy,
789    ValueForPoison(DoPoison ? kAsanStackLeftRedzoneMagic : 0LL, ShadowRZSize));
790  Value *PoisonMid   = ConstantInt::get(RZTy,
791    ValueForPoison(DoPoison ? kAsanStackMidRedzoneMagic : 0LL, ShadowRZSize));
792  Value *PoisonRight = ConstantInt::get(RZTy,
793    ValueForPoison(DoPoison ? kAsanStackRightRedzoneMagic : 0LL, ShadowRZSize));
794
795  // poison the first red zone.
796  IRB.CreateStore(PoisonLeft, IRB.CreateIntToPtr(ShadowBase, RZPtrTy));
797
798  // poison all other red zones.
799  uint64_t Pos = RedzoneSize;
800  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
801    AllocaInst *AI = AllocaVec[i];
802    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
803    uint64_t AlignedSize = getAlignedAllocaSize(AI);
804    assert(AlignedSize - SizeInBytes < RedzoneSize);
805    Value *Ptr = NULL;
806
807    Pos += AlignedSize;
808
809    assert(ShadowBase->getType() == IntptrTy);
810    if (SizeInBytes < AlignedSize) {
811      // Poison the partial redzone at right
812      Ptr = IRB.CreateAdd(
813          ShadowBase, ConstantInt::get(IntptrTy,
814                                       (Pos >> MappingScale) - ShadowRZSize));
815      size_t AddressableBytes = RedzoneSize - (AlignedSize - SizeInBytes);
816      uint32_t Poison = 0;
817      if (DoPoison) {
818        PoisonShadowPartialRightRedzone((uint8_t*)&Poison, AddressableBytes,
819                                        RedzoneSize,
820                                        1ULL << MappingScale,
821                                        kAsanStackPartialRedzoneMagic);
822      }
823      Value *PartialPoison = ConstantInt::get(RZTy, Poison);
824      IRB.CreateStore(PartialPoison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
825    }
826
827    // Poison the full redzone at right.
828    Ptr = IRB.CreateAdd(ShadowBase,
829                        ConstantInt::get(IntptrTy, Pos >> MappingScale));
830    Value *Poison = i == AllocaVec.size() - 1 ? PoisonRight : PoisonMid;
831    IRB.CreateStore(Poison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
832
833    Pos += RedzoneSize;
834  }
835}
836
837// Workaround for bug 11395: we don't want to instrument stack in functions
838// with large assembly blobs (32-bit only), otherwise reg alloc may crash.
839// FIXME: remove once the bug 11395 is fixed.
840bool AddressSanitizer::LooksLikeCodeInBug11395(Instruction *I) {
841  if (LongSize != 32) return false;
842  CallInst *CI = dyn_cast<CallInst>(I);
843  if (!CI || !CI->isInlineAsm()) return false;
844  if (CI->getNumArgOperands() <= 5) return false;
845  // We have inline assembly with quite a few arguments.
846  return true;
847}
848
849// Find all static Alloca instructions and put
850// poisoned red zones around all of them.
851// Then unpoison everything back before the function returns.
852//
853// Stack poisoning does not play well with exception handling.
854// When an exception is thrown, we essentially bypass the code
855// that unpoisones the stack. This is why the run-time library has
856// to intercept __cxa_throw (as well as longjmp, etc) and unpoison the entire
857// stack in the interceptor. This however does not work inside the
858// actual function which catches the exception. Most likely because the
859// compiler hoists the load of the shadow value somewhere too high.
860// This causes asan to report a non-existing bug on 453.povray.
861// It sounds like an LLVM bug.
862bool AddressSanitizer::poisonStackInFunction(Module &M, Function &F) {
863  if (!ClStack) return false;
864  SmallVector<AllocaInst*, 16> AllocaVec;
865  SmallVector<Instruction*, 8> RetVec;
866  uint64_t TotalSize = 0;
867
868  // Filter out Alloca instructions we want (and can) handle.
869  // Collect Ret instructions.
870  for (Function::iterator FI = F.begin(), FE = F.end();
871       FI != FE; ++FI) {
872    BasicBlock &BB = *FI;
873    for (BasicBlock::iterator BI = BB.begin(), BE = BB.end();
874         BI != BE; ++BI) {
875      if (isa<ReturnInst>(BI)) {
876          RetVec.push_back(BI);
877          continue;
878      }
879
880      AllocaInst *AI = dyn_cast<AllocaInst>(BI);
881      if (!AI) continue;
882      if (AI->isArrayAllocation()) continue;
883      if (!AI->isStaticAlloca()) continue;
884      if (!AI->getAllocatedType()->isSized()) continue;
885      if (AI->getAlignment() > RedzoneSize) continue;
886      AllocaVec.push_back(AI);
887      uint64_t AlignedSize =  getAlignedAllocaSize(AI);
888      TotalSize += AlignedSize;
889    }
890  }
891
892  if (AllocaVec.empty()) return false;
893
894  uint64_t LocalStackSize = TotalSize + (AllocaVec.size() + 1) * RedzoneSize;
895
896  bool DoStackMalloc = ClUseAfterReturn
897      && LocalStackSize <= kMaxStackMallocSize;
898
899  Instruction *InsBefore = AllocaVec[0];
900  IRBuilder<> IRB(InsBefore);
901
902
903  Type *ByteArrayTy = ArrayType::get(IRB.getInt8Ty(), LocalStackSize);
904  AllocaInst *MyAlloca =
905      new AllocaInst(ByteArrayTy, "MyAlloca", InsBefore);
906  MyAlloca->setAlignment(RedzoneSize);
907  assert(MyAlloca->isStaticAlloca());
908  Value *OrigStackBase = IRB.CreatePointerCast(MyAlloca, IntptrTy);
909  Value *LocalStackBase = OrigStackBase;
910
911  if (DoStackMalloc) {
912    Value *AsanStackMallocFunc = M.getOrInsertFunction(
913        kAsanStackMallocName, IntptrTy, IntptrTy, IntptrTy, NULL);
914    LocalStackBase = IRB.CreateCall2(AsanStackMallocFunc,
915        ConstantInt::get(IntptrTy, LocalStackSize), OrigStackBase);
916  }
917
918  // This string will be parsed by the run-time (DescribeStackAddress).
919  SmallString<2048> StackDescriptionStorage;
920  raw_svector_ostream StackDescription(StackDescriptionStorage);
921  StackDescription << F.getName() << " " << AllocaVec.size() << " ";
922
923  uint64_t Pos = RedzoneSize;
924  // Replace Alloca instructions with base+offset.
925  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
926    AllocaInst *AI = AllocaVec[i];
927    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
928    StringRef Name = AI->getName();
929    StackDescription << Pos << " " << SizeInBytes << " "
930                     << Name.size() << " " << Name << " ";
931    uint64_t AlignedSize = getAlignedAllocaSize(AI);
932    assert((AlignedSize % RedzoneSize) == 0);
933    AI->replaceAllUsesWith(
934        IRB.CreateIntToPtr(
935            IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy, Pos)),
936            AI->getType()));
937    Pos += AlignedSize + RedzoneSize;
938  }
939  assert(Pos == LocalStackSize);
940
941  // Write the Magic value and the frame description constant to the redzone.
942  Value *BasePlus0 = IRB.CreateIntToPtr(LocalStackBase, IntptrPtrTy);
943  IRB.CreateStore(ConstantInt::get(IntptrTy, kCurrentStackFrameMagic),
944                  BasePlus0);
945  Value *BasePlus1 = IRB.CreateAdd(LocalStackBase,
946                                   ConstantInt::get(IntptrTy, LongSize/8));
947  BasePlus1 = IRB.CreateIntToPtr(BasePlus1, IntptrPtrTy);
948  Value *Description = IRB.CreatePointerCast(
949      createPrivateGlobalForString(M, StackDescription.str()),
950      IntptrTy);
951  IRB.CreateStore(Description, BasePlus1);
952
953  // Poison the stack redzones at the entry.
954  Value *ShadowBase = memToShadow(LocalStackBase, IRB);
955  PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRB, ShadowBase, true);
956
957  Value *AsanStackFreeFunc = NULL;
958  if (DoStackMalloc) {
959    AsanStackFreeFunc = M.getOrInsertFunction(
960        kAsanStackFreeName, IRB.getVoidTy(),
961        IntptrTy, IntptrTy, IntptrTy, NULL);
962  }
963
964  // Unpoison the stack before all ret instructions.
965  for (size_t i = 0, n = RetVec.size(); i < n; i++) {
966    Instruction *Ret = RetVec[i];
967    IRBuilder<> IRBRet(Ret);
968
969    // Mark the current frame as retired.
970    IRBRet.CreateStore(ConstantInt::get(IntptrTy, kRetiredStackFrameMagic),
971                       BasePlus0);
972    // Unpoison the stack.
973    PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRBRet, ShadowBase, false);
974
975    if (DoStackMalloc) {
976      IRBRet.CreateCall3(AsanStackFreeFunc, LocalStackBase,
977                         ConstantInt::get(IntptrTy, LocalStackSize),
978                         OrigStackBase);
979    }
980  }
981
982  if (ClDebugStack) {
983    DEBUG(dbgs() << F);
984  }
985
986  return true;
987}
988