AddressSanitizer.cpp revision 59cca13a80ed78ccbcdaa9bf20381bd48243f2f1 
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/Transforms/Instrumentation.h"
19#include "BlackList.h"
20#include "llvm/ADT/ArrayRef.h"
21#include "llvm/ADT/DepthFirstIterator.h"
22#include "llvm/ADT/OwningPtr.h"
23#include "llvm/ADT/SmallSet.h"
24#include "llvm/ADT/SmallString.h"
25#include "llvm/ADT/SmallVector.h"
26#include "llvm/ADT/StringExtras.h"
27#include "llvm/ADT/Triple.h"
28#include "llvm/DataLayout.h"
29#include "llvm/DIBuilder.h"
30#include "llvm/Function.h"
31#include "llvm/IRBuilder.h"
32#include "llvm/InlineAsm.h"
33#include "llvm/InstVisitor.h"
34#include "llvm/IntrinsicInst.h"
35#include "llvm/LLVMContext.h"
36#include "llvm/Module.h"
37#include "llvm/Support/CommandLine.h"
38#include "llvm/Support/DataTypes.h"
39#include "llvm/Support/Debug.h"
40#include "llvm/Support/raw_ostream.h"
41#include "llvm/Support/system_error.h"
42#include "llvm/Target/TargetMachine.h"
43#include "llvm/Transforms/Utils/BasicBlockUtils.h"
44#include "llvm/Transforms/Utils/Local.h"
45#include "llvm/Transforms/Utils/ModuleUtils.h"
46#include "llvm/Type.h"
47#include <algorithm>
48#include <string>
49
50using namespace llvm;
51
52static const uint64_t kDefaultShadowScale = 3;
53static const uint64_t kDefaultShadowOffset32 = 1ULL << 29;
54static const uint64_t kDefaultShadowOffset64 = 1ULL << 44;
55static const uint64_t kDefaultShadowOffsetAndroid = 0;
56
57static const size_t kMaxStackMallocSize = 1 << 16;  // 64K
58static const uintptr_t kCurrentStackFrameMagic = 0x41B58AB3;
59static const uintptr_t kRetiredStackFrameMagic = 0x45E0360E;
60
61static const char *kAsanModuleCtorName = "asan.module_ctor";
62static const char *kAsanModuleDtorName = "asan.module_dtor";
63static const int   kAsanCtorAndCtorPriority = 1;
64static const char *kAsanReportErrorTemplate = "__asan_report_";
65static const char *kAsanRegisterGlobalsName = "__asan_register_globals";
66static const char *kAsanUnregisterGlobalsName = "__asan_unregister_globals";
67static const char *kAsanPoisonGlobalsName = "__asan_before_dynamic_init";
68static const char *kAsanUnpoisonGlobalsName = "__asan_after_dynamic_init";
69static const char *kAsanInitName = "__asan_init";
70static const char *kAsanHandleNoReturnName = "__asan_handle_no_return";
71static const char *kAsanMappingOffsetName = "__asan_mapping_offset";
72static const char *kAsanMappingScaleName = "__asan_mapping_scale";
73static const char *kAsanStackMallocName = "__asan_stack_malloc";
74static const char *kAsanStackFreeName = "__asan_stack_free";
75static const char *kAsanGenPrefix = "__asan_gen_";
76static const char *kAsanPoisonStackMemoryName = "__asan_poison_stack_memory";
77static const char *kAsanUnpoisonStackMemoryName =
78    "__asan_unpoison_stack_memory";
79
80static const int kAsanStackLeftRedzoneMagic = 0xf1;
81static const int kAsanStackMidRedzoneMagic = 0xf2;
82static const int kAsanStackRightRedzoneMagic = 0xf3;
83static const int kAsanStackPartialRedzoneMagic = 0xf4;
84
85// Accesses sizes are powers of two: 1, 2, 4, 8, 16.
86static const size_t kNumberOfAccessSizes = 5;
87
88// Command-line flags.
89
90// This flag may need to be replaced with -f[no-]asan-reads.
91static cl::opt<bool> ClInstrumentReads("asan-instrument-reads",
92       cl::desc("instrument read instructions"), cl::Hidden, cl::init(true));
93static cl::opt<bool> ClInstrumentWrites("asan-instrument-writes",
94       cl::desc("instrument write instructions"), cl::Hidden, cl::init(true));
95static cl::opt<bool> ClInstrumentAtomics("asan-instrument-atomics",
96       cl::desc("instrument atomic instructions (rmw, cmpxchg)"),
97       cl::Hidden, cl::init(true));
98static cl::opt<bool> ClAlwaysSlowPath("asan-always-slow-path",
99       cl::desc("use instrumentation with slow path for all accesses"),
100       cl::Hidden, cl::init(false));
101// This flag limits the number of instructions to be instrumented
102// in any given BB. Normally, this should be set to unlimited (INT_MAX),
103// but due to http://llvm.org/bugs/show_bug.cgi?id=12652 we temporary
104// set it to 10000.
105static cl::opt<int> ClMaxInsnsToInstrumentPerBB("asan-max-ins-per-bb",
106       cl::init(10000),
107       cl::desc("maximal number of instructions to instrument in any given BB"),
108       cl::Hidden);
109// This flag may need to be replaced with -f[no]asan-stack.
110static cl::opt<bool> ClStack("asan-stack",
111       cl::desc("Handle stack memory"), cl::Hidden, cl::init(true));
112// This flag may need to be replaced with -f[no]asan-use-after-return.
113static cl::opt<bool> ClUseAfterReturn("asan-use-after-return",
114       cl::desc("Check return-after-free"), cl::Hidden, cl::init(false));
115// This flag may need to be replaced with -f[no]asan-globals.
116static cl::opt<bool> ClGlobals("asan-globals",
117       cl::desc("Handle global objects"), cl::Hidden, cl::init(true));
118static cl::opt<bool> ClInitializers("asan-initialization-order",
119       cl::desc("Handle C++ initializer order"), cl::Hidden, cl::init(false));
120static cl::opt<bool> ClMemIntrin("asan-memintrin",
121       cl::desc("Handle memset/memcpy/memmove"), cl::Hidden, cl::init(true));
122static cl::opt<bool> ClRealignStack("asan-realign-stack",
123       cl::desc("Realign stack to 32"), cl::Hidden, cl::init(true));
124static cl::opt<std::string> ClBlacklistFile("asan-blacklist",
125       cl::desc("File containing the list of objects to ignore "
126                "during instrumentation"), cl::Hidden);
127
128// These flags allow to change the shadow mapping.
129// The shadow mapping looks like
130//    Shadow = (Mem >> scale) + (1 << offset_log)
131static cl::opt<int> ClMappingScale("asan-mapping-scale",
132       cl::desc("scale of asan shadow mapping"), cl::Hidden, cl::init(0));
133static cl::opt<int> ClMappingOffsetLog("asan-mapping-offset-log",
134       cl::desc("offset of asan shadow mapping"), cl::Hidden, cl::init(-1));
135
136// Optimization flags. Not user visible, used mostly for testing
137// and benchmarking the tool.
138static cl::opt<bool> ClOpt("asan-opt",
139       cl::desc("Optimize instrumentation"), cl::Hidden, cl::init(true));
140static cl::opt<bool> ClOptSameTemp("asan-opt-same-temp",
141       cl::desc("Instrument the same temp just once"), cl::Hidden,
142       cl::init(true));
143static cl::opt<bool> ClOptGlobals("asan-opt-globals",
144       cl::desc("Don't instrument scalar globals"), cl::Hidden, cl::init(true));
145
146static cl::opt<bool> ClCheckLifetime("asan-check-lifetime",
147       cl::desc("Use llvm.lifetime intrinsics to insert extra checks"),
148       cl::Hidden, cl::init(false));
149
150// Debug flags.
151static cl::opt<int> ClDebug("asan-debug", cl::desc("debug"), cl::Hidden,
152                            cl::init(0));
153static cl::opt<int> ClDebugStack("asan-debug-stack", cl::desc("debug stack"),
154                                 cl::Hidden, cl::init(0));
155static cl::opt<std::string> ClDebugFunc("asan-debug-func",
156                                        cl::Hidden, cl::desc("Debug func"));
157static cl::opt<int> ClDebugMin("asan-debug-min", cl::desc("Debug min inst"),
158                               cl::Hidden, cl::init(-1));
159static cl::opt<int> ClDebugMax("asan-debug-max", cl::desc("Debug man inst"),
160                               cl::Hidden, cl::init(-1));
161
162namespace {
163/// A set of dynamically initialized globals extracted from metadata.
164class SetOfDynamicallyInitializedGlobals {
165 public:
166  void Init(Module& M) {
167    // Clang generates metadata identifying all dynamically initialized globals.
168    NamedMDNode *DynamicGlobals =
169        M.getNamedMetadata("llvm.asan.dynamically_initialized_globals");
170    if (!DynamicGlobals)
171      return;
172    for (int i = 0, n = DynamicGlobals->getNumOperands(); i < n; ++i) {
173      MDNode *MDN = DynamicGlobals->getOperand(i);
174      assert(MDN->getNumOperands() == 1);
175      Value *VG = MDN->getOperand(0);
176      // The optimizer may optimize away a global entirely, in which case we
177      // cannot instrument access to it.
178      if (!VG)
179        continue;
180      DynInitGlobals.insert(cast<GlobalVariable>(VG));
181    }
182  }
183  bool Contains(GlobalVariable *G) { return DynInitGlobals.count(G) != 0; }
184 private:
185  SmallSet<GlobalValue*, 32> DynInitGlobals;
186};
187
188static int MappingScale() {
189  return ClMappingScale ? ClMappingScale : kDefaultShadowScale;
190}
191
192static size_t RedzoneSize() {
193  // Redzone used for stack and globals is at least 32 bytes.
194  // For scales 6 and 7, the redzone has to be 64 and 128 bytes respectively.
195  return std::max(32U, 1U << MappingScale());
196}
197
198/// AddressSanitizer: instrument the code in module to find memory bugs.
199struct AddressSanitizer : public FunctionPass {
200  AddressSanitizer(bool CheckInitOrder = false,
201                   bool CheckUseAfterReturn = false,
202                   bool CheckLifetime = false,
203                   StringRef BlacklistFile = StringRef())
204      : FunctionPass(ID),
205        CheckInitOrder(CheckInitOrder || ClInitializers),
206        CheckUseAfterReturn(CheckUseAfterReturn || ClUseAfterReturn),
207        CheckLifetime(CheckLifetime || ClCheckLifetime),
208        BlacklistFile(BlacklistFile.empty() ? ClBlacklistFile
209                                            : BlacklistFile) {}
210  virtual const char *getPassName() const {
211    return "AddressSanitizerFunctionPass";
212  }
213  void instrumentMop(Instruction *I);
214  void instrumentAddress(Instruction *OrigIns, IRBuilder<> &IRB,
215                         Value *Addr, uint32_t TypeSize, bool IsWrite);
216  Value *createSlowPathCmp(IRBuilder<> &IRB, Value *AddrLong,
217                           Value *ShadowValue, uint32_t TypeSize);
218  Instruction *generateCrashCode(Instruction *InsertBefore, Value *Addr,
219                                 bool IsWrite, size_t AccessSizeIndex);
220  bool instrumentMemIntrinsic(MemIntrinsic *MI);
221  void instrumentMemIntrinsicParam(Instruction *OrigIns, Value *Addr,
222                                   Value *Size,
223                                   Instruction *InsertBefore, bool IsWrite);
224  Value *memToShadow(Value *Shadow, IRBuilder<> &IRB);
225  bool runOnFunction(Function &F);
226  void createInitializerPoisonCalls(Module &M,
227                                    Value *FirstAddr, Value *LastAddr);
228  bool maybeInsertAsanInitAtFunctionEntry(Function &F);
229  virtual bool doInitialization(Module &M);
230  static char ID;  // Pass identification, replacement for typeid
231
232 private:
233  void initializeCallbacks(Module &M);
234
235  bool ShouldInstrumentGlobal(GlobalVariable *G);
236  bool LooksLikeCodeInBug11395(Instruction *I);
237  void FindDynamicInitializers(Module &M);
238
239  bool CheckInitOrder;
240  bool CheckUseAfterReturn;
241  bool CheckLifetime;
242  LLVMContext *C;
243  DataLayout *TD;
244  uint64_t MappingOffset;
245  int LongSize;
246  Type *IntptrTy;
247  Function *AsanCtorFunction;
248  Function *AsanInitFunction;
249  Function *AsanHandleNoReturnFunc;
250  SmallString<64> BlacklistFile;
251  OwningPtr<BlackList> BL;
252  // This array is indexed by AccessIsWrite and log2(AccessSize).
253  Function *AsanErrorCallback[2][kNumberOfAccessSizes];
254  InlineAsm *EmptyAsm;
255  SetOfDynamicallyInitializedGlobals DynamicallyInitializedGlobals;
256
257  friend struct FunctionStackPoisoner;
258};
259
260class AddressSanitizerModule : public ModulePass {
261 public:
262  AddressSanitizerModule(bool CheckInitOrder = false,
263                         StringRef BlacklistFile = StringRef())
264      : ModulePass(ID),
265        CheckInitOrder(CheckInitOrder || ClInitializers),
266        BlacklistFile(BlacklistFile.empty() ? ClBlacklistFile
267                                            : BlacklistFile) {}
268  bool runOnModule(Module &M);
269  static char ID;  // Pass identification, replacement for typeid
270  virtual const char *getPassName() const {
271    return "AddressSanitizerModule";
272  }
273
274 private:
275  bool ShouldInstrumentGlobal(GlobalVariable *G);
276  void createInitializerPoisonCalls(Module &M, Value *FirstAddr,
277                                    Value *LastAddr);
278
279  bool CheckInitOrder;
280  SmallString<64> BlacklistFile;
281  OwningPtr<BlackList> BL;
282  SetOfDynamicallyInitializedGlobals DynamicallyInitializedGlobals;
283  Type *IntptrTy;
284  LLVMContext *C;
285  DataLayout *TD;
286};
287
288// Stack poisoning does not play well with exception handling.
289// When an exception is thrown, we essentially bypass the code
290// that unpoisones the stack. This is why the run-time library has
291// to intercept __cxa_throw (as well as longjmp, etc) and unpoison the entire
292// stack in the interceptor. This however does not work inside the
293// actual function which catches the exception. Most likely because the
294// compiler hoists the load of the shadow value somewhere too high.
295// This causes asan to report a non-existing bug on 453.povray.
296// It sounds like an LLVM bug.
297struct FunctionStackPoisoner : public InstVisitor<FunctionStackPoisoner> {
298  Function &F;
299  AddressSanitizer &ASan;
300  DIBuilder DIB;
301  LLVMContext *C;
302  Type *IntptrTy;
303  Type *IntptrPtrTy;
304
305  SmallVector<AllocaInst*, 16> AllocaVec;
306  SmallVector<Instruction*, 8> RetVec;
307  uint64_t TotalStackSize;
308  unsigned StackAlignment;
309
310  Function *AsanStackMallocFunc, *AsanStackFreeFunc;
311  Function *AsanPoisonStackMemoryFunc, *AsanUnpoisonStackMemoryFunc;
312
313  FunctionStackPoisoner(Function &F, AddressSanitizer &ASan)
314      : F(F), ASan(ASan), DIB(*F.getParent()), C(ASan.C),
315        IntptrTy(ASan.IntptrTy), IntptrPtrTy(PointerType::get(IntptrTy, 0)),
316        TotalStackSize(0), StackAlignment(1 << MappingScale()) {}
317
318  bool runOnFunction() {
319    if (!ClStack) return false;
320    // Collect alloca, ret, lifetime instructions etc.
321    for (df_iterator<BasicBlock*> DI = df_begin(&F.getEntryBlock()),
322         DE = df_end(&F.getEntryBlock()); DI != DE; ++DI) {
323      BasicBlock *BB = *DI;
324      visit(*BB);
325    }
326    if (AllocaVec.empty()) return false;
327
328    initializeCallbacks(*F.getParent());
329
330    poisonStack();
331
332    if (ClDebugStack) {
333      DEBUG(dbgs() << F);
334    }
335    return true;
336  }
337
338  // Finds all static Alloca instructions and puts
339  // poisoned red zones around all of them.
340  // Then unpoison everything back before the function returns.
341  void poisonStack();
342
343  // ----------------------- Visitors.
344  /// \brief Collect all Ret instructions.
345  void visitReturnInst(ReturnInst &RI) {
346    RetVec.push_back(&RI);
347  }
348
349  /// \brief Collect Alloca instructions we want (and can) handle.
350  void visitAllocaInst(AllocaInst &AI) {
351    if (AI.isArrayAllocation()) return;
352    if (!AI.isStaticAlloca()) return;
353    if (!AI.getAllocatedType()->isSized()) return;
354
355    StackAlignment = std::max(StackAlignment, AI.getAlignment());
356    AllocaVec.push_back(&AI);
357    uint64_t AlignedSize =  getAlignedAllocaSize(&AI);
358    TotalStackSize += AlignedSize;
359  }
360
361  // ---------------------- Helpers.
362  void initializeCallbacks(Module &M);
363
364  uint64_t getAllocaSizeInBytes(AllocaInst *AI) {
365    Type *Ty = AI->getAllocatedType();
366    uint64_t SizeInBytes = ASan.TD->getTypeAllocSize(Ty);
367    return SizeInBytes;
368  }
369  uint64_t getAlignedSize(uint64_t SizeInBytes) {
370    size_t RZ = RedzoneSize();
371    return ((SizeInBytes + RZ - 1) / RZ) * RZ;
372  }
373  uint64_t getAlignedAllocaSize(AllocaInst *AI) {
374    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
375    return getAlignedSize(SizeInBytes);
376  }
377  void poisonRedZones(const ArrayRef<AllocaInst*> &AllocaVec, IRBuilder<> IRB,
378                      Value *ShadowBase, bool DoPoison);
379  void poisonAlloca(Value *V, uint64_t Size, IRBuilder<> IRB, bool DoPoison);
380  /// Analyze lifetime intrinsics for given alloca. Use Value* instead of
381  /// AllocaInst* here, as we call this method after we merge all allocas into a
382  /// single one. Returns true if ASan added some instrumentation.
383  bool handleAllocaLifetime(Value *Alloca);
384  /// Analyze lifetime intrinsics for a specific value, casted from alloca.
385  /// Returns true if if ASan added some instrumentation.
386  bool handleValueLifetime(Value *V);
387};
388
389}  // namespace
390
391char AddressSanitizer::ID = 0;
392INITIALIZE_PASS(AddressSanitizer, "asan",
393    "AddressSanitizer: detects use-after-free and out-of-bounds bugs.",
394    false, false)
395FunctionPass *llvm::createAddressSanitizerFunctionPass(
396    bool CheckInitOrder, bool CheckUseAfterReturn, bool CheckLifetime,
397    StringRef BlacklistFile) {
398  return new AddressSanitizer(CheckInitOrder, CheckUseAfterReturn,
399                              CheckLifetime, BlacklistFile);
400}
401
402char AddressSanitizerModule::ID = 0;
403INITIALIZE_PASS(AddressSanitizerModule, "asan-module",
404    "AddressSanitizer: detects use-after-free and out-of-bounds bugs."
405    "ModulePass", false, false)
406ModulePass *llvm::createAddressSanitizerModulePass(
407    bool CheckInitOrder, StringRef BlacklistFile) {
408  return new AddressSanitizerModule(CheckInitOrder, BlacklistFile);
409}
410
411static size_t TypeSizeToSizeIndex(uint32_t TypeSize) {
412  size_t Res = CountTrailingZeros_32(TypeSize / 8);
413  assert(Res < kNumberOfAccessSizes);
414  return Res;
415}
416
417// Create a constant for Str so that we can pass it to the run-time lib.
418static GlobalVariable *createPrivateGlobalForString(Module &M, StringRef Str) {
419  Constant *StrConst = ConstantDataArray::getString(M.getContext(), Str);
420  return new GlobalVariable(M, StrConst->getType(), true,
421                            GlobalValue::PrivateLinkage, StrConst,
422                            kAsanGenPrefix);
423}
424
425static bool GlobalWasGeneratedByAsan(GlobalVariable *G) {
426  return G->getName().find(kAsanGenPrefix) == 0;
427}
428
429Value *AddressSanitizer::memToShadow(Value *Shadow, IRBuilder<> &IRB) {
430  // Shadow >> scale
431  Shadow = IRB.CreateLShr(Shadow, MappingScale());
432  if (MappingOffset == 0)
433    return Shadow;
434  // (Shadow >> scale) | offset
435  return IRB.CreateOr(Shadow, ConstantInt::get(IntptrTy,
436                                               MappingOffset));
437}
438
439void AddressSanitizer::instrumentMemIntrinsicParam(
440    Instruction *OrigIns,
441    Value *Addr, Value *Size, Instruction *InsertBefore, bool IsWrite) {
442  // Check the first byte.
443  {
444    IRBuilder<> IRB(InsertBefore);
445    instrumentAddress(OrigIns, IRB, Addr, 8, IsWrite);
446  }
447  // Check the last byte.
448  {
449    IRBuilder<> IRB(InsertBefore);
450    Value *SizeMinusOne = IRB.CreateSub(
451        Size, ConstantInt::get(Size->getType(), 1));
452    SizeMinusOne = IRB.CreateIntCast(SizeMinusOne, IntptrTy, false);
453    Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
454    Value *AddrPlusSizeMinisOne = IRB.CreateAdd(AddrLong, SizeMinusOne);
455    instrumentAddress(OrigIns, IRB, AddrPlusSizeMinisOne, 8, IsWrite);
456  }
457}
458
459// Instrument memset/memmove/memcpy
460bool AddressSanitizer::instrumentMemIntrinsic(MemIntrinsic *MI) {
461  Value *Dst = MI->getDest();
462  MemTransferInst *MemTran = dyn_cast<MemTransferInst>(MI);
463  Value *Src = MemTran ? MemTran->getSource() : 0;
464  Value *Length = MI->getLength();
465
466  Constant *ConstLength = dyn_cast<Constant>(Length);
467  Instruction *InsertBefore = MI;
468  if (ConstLength) {
469    if (ConstLength->isNullValue()) return false;
470  } else {
471    // The size is not a constant so it could be zero -- check at run-time.
472    IRBuilder<> IRB(InsertBefore);
473
474    Value *Cmp = IRB.CreateICmpNE(Length,
475                                  Constant::getNullValue(Length->getType()));
476    InsertBefore = SplitBlockAndInsertIfThen(cast<Instruction>(Cmp), false);
477  }
478
479  instrumentMemIntrinsicParam(MI, Dst, Length, InsertBefore, true);
480  if (Src)
481    instrumentMemIntrinsicParam(MI, Src, Length, InsertBefore, false);
482  return true;
483}
484
485// If I is an interesting memory access, return the PointerOperand
486// and set IsWrite. Otherwise return NULL.
487static Value *isInterestingMemoryAccess(Instruction *I, bool *IsWrite) {
488  if (LoadInst *LI = dyn_cast<LoadInst>(I)) {
489    if (!ClInstrumentReads) return NULL;
490    *IsWrite = false;
491    return LI->getPointerOperand();
492  }
493  if (StoreInst *SI = dyn_cast<StoreInst>(I)) {
494    if (!ClInstrumentWrites) return NULL;
495    *IsWrite = true;
496    return SI->getPointerOperand();
497  }
498  if (AtomicRMWInst *RMW = dyn_cast<AtomicRMWInst>(I)) {
499    if (!ClInstrumentAtomics) return NULL;
500    *IsWrite = true;
501    return RMW->getPointerOperand();
502  }
503  if (AtomicCmpXchgInst *XCHG = dyn_cast<AtomicCmpXchgInst>(I)) {
504    if (!ClInstrumentAtomics) return NULL;
505    *IsWrite = true;
506    return XCHG->getPointerOperand();
507  }
508  return NULL;
509}
510
511void AddressSanitizer::instrumentMop(Instruction *I) {
512  bool IsWrite = false;
513  Value *Addr = isInterestingMemoryAccess(I, &IsWrite);
514  assert(Addr);
515  if (ClOpt && ClOptGlobals) {
516    if (GlobalVariable *G = dyn_cast<GlobalVariable>(Addr)) {
517      // If initialization order checking is disabled, a simple access to a
518      // dynamically initialized global is always valid.
519      if (!CheckInitOrder)
520        return;
521      // If a global variable does not have dynamic initialization we don't
522      // have to instrument it.  However, if a global does not have initailizer
523      // at all, we assume it has dynamic initializer (in other TU).
524      if (G->hasInitializer() && !DynamicallyInitializedGlobals.Contains(G))
525        return;
526    }
527  }
528
529  Type *OrigPtrTy = Addr->getType();
530  Type *OrigTy = cast<PointerType>(OrigPtrTy)->getElementType();
531
532  assert(OrigTy->isSized());
533  uint32_t TypeSize = TD->getTypeStoreSizeInBits(OrigTy);
534
535  if (TypeSize != 8  && TypeSize != 16 &&
536      TypeSize != 32 && TypeSize != 64 && TypeSize != 128) {
537    // Ignore all unusual sizes.
538    return;
539  }
540
541  IRBuilder<> IRB(I);
542  instrumentAddress(I, IRB, Addr, TypeSize, IsWrite);
543}
544
545// Validate the result of Module::getOrInsertFunction called for an interface
546// function of AddressSanitizer. If the instrumented module defines a function
547// with the same name, their prototypes must match, otherwise
548// getOrInsertFunction returns a bitcast.
549static Function *checkInterfaceFunction(Constant *FuncOrBitcast) {
550  if (isa<Function>(FuncOrBitcast)) return cast<Function>(FuncOrBitcast);
551  FuncOrBitcast->dump();
552  report_fatal_error("trying to redefine an AddressSanitizer "
553                     "interface function");
554}
555
556Instruction *AddressSanitizer::generateCrashCode(
557    Instruction *InsertBefore, Value *Addr,
558    bool IsWrite, size_t AccessSizeIndex) {
559  IRBuilder<> IRB(InsertBefore);
560  CallInst *Call = IRB.CreateCall(AsanErrorCallback[IsWrite][AccessSizeIndex],
561                                  Addr);
562  // We don't do Call->setDoesNotReturn() because the BB already has
563  // UnreachableInst at the end.
564  // This EmptyAsm is required to avoid callback merge.
565  IRB.CreateCall(EmptyAsm);
566  return Call;
567}
568
569Value *AddressSanitizer::createSlowPathCmp(IRBuilder<> &IRB, Value *AddrLong,
570                                            Value *ShadowValue,
571                                            uint32_t TypeSize) {
572  size_t Granularity = 1 << MappingScale();
573  // Addr & (Granularity - 1)
574  Value *LastAccessedByte = IRB.CreateAnd(
575      AddrLong, ConstantInt::get(IntptrTy, Granularity - 1));
576  // (Addr & (Granularity - 1)) + size - 1
577  if (TypeSize / 8 > 1)
578    LastAccessedByte = IRB.CreateAdd(
579        LastAccessedByte, ConstantInt::get(IntptrTy, TypeSize / 8 - 1));
580  // (uint8_t) ((Addr & (Granularity-1)) + size - 1)
581  LastAccessedByte = IRB.CreateIntCast(
582      LastAccessedByte, ShadowValue->getType(), false);
583  // ((uint8_t) ((Addr & (Granularity-1)) + size - 1)) >= ShadowValue
584  return IRB.CreateICmpSGE(LastAccessedByte, ShadowValue);
585}
586
587void AddressSanitizer::instrumentAddress(Instruction *OrigIns,
588                                         IRBuilder<> &IRB, Value *Addr,
589                                         uint32_t TypeSize, bool IsWrite) {
590  Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
591
592  Type *ShadowTy  = IntegerType::get(
593      *C, std::max(8U, TypeSize >> MappingScale()));
594  Type *ShadowPtrTy = PointerType::get(ShadowTy, 0);
595  Value *ShadowPtr = memToShadow(AddrLong, IRB);
596  Value *CmpVal = Constant::getNullValue(ShadowTy);
597  Value *ShadowValue = IRB.CreateLoad(
598      IRB.CreateIntToPtr(ShadowPtr, ShadowPtrTy));
599
600  Value *Cmp = IRB.CreateICmpNE(ShadowValue, CmpVal);
601  size_t AccessSizeIndex = TypeSizeToSizeIndex(TypeSize);
602  size_t Granularity = 1 << MappingScale();
603  TerminatorInst *CrashTerm = 0;
604
605  if (ClAlwaysSlowPath || (TypeSize < 8 * Granularity)) {
606    TerminatorInst *CheckTerm =
607        SplitBlockAndInsertIfThen(cast<Instruction>(Cmp), false);
608    assert(dyn_cast<BranchInst>(CheckTerm)->isUnconditional());
609    BasicBlock *NextBB = CheckTerm->getSuccessor(0);
610    IRB.SetInsertPoint(CheckTerm);
611    Value *Cmp2 = createSlowPathCmp(IRB, AddrLong, ShadowValue, TypeSize);
612    BasicBlock *CrashBlock =
613        BasicBlock::Create(*C, "", NextBB->getParent(), NextBB);
614    CrashTerm = new UnreachableInst(*C, CrashBlock);
615    BranchInst *NewTerm = BranchInst::Create(CrashBlock, NextBB, Cmp2);
616    ReplaceInstWithInst(CheckTerm, NewTerm);
617  } else {
618    CrashTerm = SplitBlockAndInsertIfThen(cast<Instruction>(Cmp), true);
619  }
620
621  Instruction *Crash =
622      generateCrashCode(CrashTerm, AddrLong, IsWrite, AccessSizeIndex);
623  Crash->setDebugLoc(OrigIns->getDebugLoc());
624}
625
626void AddressSanitizerModule::createInitializerPoisonCalls(
627    Module &M, Value *FirstAddr, Value *LastAddr) {
628  // We do all of our poisoning and unpoisoning within _GLOBAL__I_a.
629  Function *GlobalInit = M.getFunction("_GLOBAL__I_a");
630  // If that function is not present, this TU contains no globals, or they have
631  // all been optimized away
632  if (!GlobalInit)
633    return;
634
635  // Set up the arguments to our poison/unpoison functions.
636  IRBuilder<> IRB(GlobalInit->begin()->getFirstInsertionPt());
637
638  // Declare our poisoning and unpoisoning functions.
639  Function *AsanPoisonGlobals = checkInterfaceFunction(M.getOrInsertFunction(
640      kAsanPoisonGlobalsName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
641  AsanPoisonGlobals->setLinkage(Function::ExternalLinkage);
642  Function *AsanUnpoisonGlobals = checkInterfaceFunction(M.getOrInsertFunction(
643      kAsanUnpoisonGlobalsName, IRB.getVoidTy(), NULL));
644  AsanUnpoisonGlobals->setLinkage(Function::ExternalLinkage);
645
646  // Add a call to poison all external globals before the given function starts.
647  IRB.CreateCall2(AsanPoisonGlobals, FirstAddr, LastAddr);
648
649  // Add calls to unpoison all globals before each return instruction.
650  for (Function::iterator I = GlobalInit->begin(), E = GlobalInit->end();
651      I != E; ++I) {
652    if (ReturnInst *RI = dyn_cast<ReturnInst>(I->getTerminator())) {
653      CallInst::Create(AsanUnpoisonGlobals, "", RI);
654    }
655  }
656}
657
658bool AddressSanitizerModule::ShouldInstrumentGlobal(GlobalVariable *G) {
659  Type *Ty = cast<PointerType>(G->getType())->getElementType();
660  DEBUG(dbgs() << "GLOBAL: " << *G << "\n");
661
662  if (BL->isIn(*G)) return false;
663  if (!Ty->isSized()) return false;
664  if (!G->hasInitializer()) return false;
665  if (GlobalWasGeneratedByAsan(G)) return false;  // Our own global.
666  // Touch only those globals that will not be defined in other modules.
667  // Don't handle ODR type linkages since other modules may be built w/o asan.
668  if (G->getLinkage() != GlobalVariable::ExternalLinkage &&
669      G->getLinkage() != GlobalVariable::PrivateLinkage &&
670      G->getLinkage() != GlobalVariable::InternalLinkage)
671    return false;
672  // Two problems with thread-locals:
673  //   - The address of the main thread's copy can't be computed at link-time.
674  //   - Need to poison all copies, not just the main thread's one.
675  if (G->isThreadLocal())
676    return false;
677  // For now, just ignore this Alloca if the alignment is large.
678  if (G->getAlignment() > RedzoneSize()) return false;
679
680  // Ignore all the globals with the names starting with "\01L_OBJC_".
681  // Many of those are put into the .cstring section. The linker compresses
682  // that section by removing the spare \0s after the string terminator, so
683  // our redzones get broken.
684  if ((G->getName().find("\01L_OBJC_") == 0) ||
685      (G->getName().find("\01l_OBJC_") == 0)) {
686    DEBUG(dbgs() << "Ignoring \\01L_OBJC_* global: " << *G);
687    return false;
688  }
689
690  if (G->hasSection()) {
691    StringRef Section(G->getSection());
692    // Ignore the globals from the __OBJC section. The ObjC runtime assumes
693    // those conform to /usr/lib/objc/runtime.h, so we can't add redzones to
694    // them.
695    if ((Section.find("__OBJC,") == 0) ||
696        (Section.find("__DATA, __objc_") == 0)) {
697      DEBUG(dbgs() << "Ignoring ObjC runtime global: " << *G);
698      return false;
699    }
700    // See http://code.google.com/p/address-sanitizer/issues/detail?id=32
701    // Constant CFString instances are compiled in the following way:
702    //  -- the string buffer is emitted into
703    //     __TEXT,__cstring,cstring_literals
704    //  -- the constant NSConstantString structure referencing that buffer
705    //     is placed into __DATA,__cfstring
706    // Therefore there's no point in placing redzones into __DATA,__cfstring.
707    // Moreover, it causes the linker to crash on OS X 10.7
708    if (Section.find("__DATA,__cfstring") == 0) {
709      DEBUG(dbgs() << "Ignoring CFString: " << *G);
710      return false;
711    }
712  }
713
714  return true;
715}
716
717// This function replaces all global variables with new variables that have
718// trailing redzones. It also creates a function that poisons
719// redzones and inserts this function into llvm.global_ctors.
720bool AddressSanitizerModule::runOnModule(Module &M) {
721  if (!ClGlobals) return false;
722  TD = getAnalysisIfAvailable<DataLayout>();
723  if (!TD)
724    return false;
725  BL.reset(new BlackList(BlacklistFile));
726  if (BL->isIn(M)) return false;
727  DynamicallyInitializedGlobals.Init(M);
728  C = &(M.getContext());
729  IntptrTy = Type::getIntNTy(*C, TD->getPointerSizeInBits());
730
731  SmallVector<GlobalVariable *, 16> GlobalsToChange;
732
733  for (Module::GlobalListType::iterator G = M.global_begin(),
734       E = M.global_end(); G != E; ++G) {
735    if (ShouldInstrumentGlobal(G))
736      GlobalsToChange.push_back(G);
737  }
738
739  size_t n = GlobalsToChange.size();
740  if (n == 0) return false;
741
742  // A global is described by a structure
743  //   size_t beg;
744  //   size_t size;
745  //   size_t size_with_redzone;
746  //   const char *name;
747  //   size_t has_dynamic_init;
748  // We initialize an array of such structures and pass it to a run-time call.
749  StructType *GlobalStructTy = StructType::get(IntptrTy, IntptrTy,
750                                               IntptrTy, IntptrTy,
751                                               IntptrTy, NULL);
752  SmallVector<Constant *, 16> Initializers(n), DynamicInit;
753
754
755  Function *CtorFunc = M.getFunction(kAsanModuleCtorName);
756  assert(CtorFunc);
757  IRBuilder<> IRB(CtorFunc->getEntryBlock().getTerminator());
758
759  // The addresses of the first and last dynamically initialized globals in
760  // this TU.  Used in initialization order checking.
761  Value *FirstDynamic = 0, *LastDynamic = 0;
762
763  for (size_t i = 0; i < n; i++) {
764    GlobalVariable *G = GlobalsToChange[i];
765    PointerType *PtrTy = cast<PointerType>(G->getType());
766    Type *Ty = PtrTy->getElementType();
767    uint64_t SizeInBytes = TD->getTypeAllocSize(Ty);
768    size_t RZ = RedzoneSize();
769    uint64_t RightRedzoneSize = RZ + (RZ - (SizeInBytes % RZ));
770    Type *RightRedZoneTy = ArrayType::get(IRB.getInt8Ty(), RightRedzoneSize);
771    // Determine whether this global should be poisoned in initialization.
772    bool GlobalHasDynamicInitializer =
773        DynamicallyInitializedGlobals.Contains(G);
774    // Don't check initialization order if this global is blacklisted.
775    GlobalHasDynamicInitializer &= !BL->isInInit(*G);
776
777    StructType *NewTy = StructType::get(Ty, RightRedZoneTy, NULL);
778    Constant *NewInitializer = ConstantStruct::get(
779        NewTy, G->getInitializer(),
780        Constant::getNullValue(RightRedZoneTy), NULL);
781
782    SmallString<2048> DescriptionOfGlobal = G->getName();
783    DescriptionOfGlobal += " (";
784    DescriptionOfGlobal += M.getModuleIdentifier();
785    DescriptionOfGlobal += ")";
786    GlobalVariable *Name = createPrivateGlobalForString(M, DescriptionOfGlobal);
787
788    // Create a new global variable with enough space for a redzone.
789    GlobalVariable *NewGlobal = new GlobalVariable(
790        M, NewTy, G->isConstant(), G->getLinkage(),
791        NewInitializer, "", G, G->getThreadLocalMode());
792    NewGlobal->copyAttributesFrom(G);
793    NewGlobal->setAlignment(RZ);
794
795    Value *Indices2[2];
796    Indices2[0] = IRB.getInt32(0);
797    Indices2[1] = IRB.getInt32(0);
798
799    G->replaceAllUsesWith(
800        ConstantExpr::getGetElementPtr(NewGlobal, Indices2, true));
801    NewGlobal->takeName(G);
802    G->eraseFromParent();
803
804    Initializers[i] = ConstantStruct::get(
805        GlobalStructTy,
806        ConstantExpr::getPointerCast(NewGlobal, IntptrTy),
807        ConstantInt::get(IntptrTy, SizeInBytes),
808        ConstantInt::get(IntptrTy, SizeInBytes + RightRedzoneSize),
809        ConstantExpr::getPointerCast(Name, IntptrTy),
810        ConstantInt::get(IntptrTy, GlobalHasDynamicInitializer),
811        NULL);
812
813    // Populate the first and last globals declared in this TU.
814    if (CheckInitOrder && GlobalHasDynamicInitializer) {
815      LastDynamic = ConstantExpr::getPointerCast(NewGlobal, IntptrTy);
816      if (FirstDynamic == 0)
817        FirstDynamic = LastDynamic;
818    }
819
820    DEBUG(dbgs() << "NEW GLOBAL: " << *NewGlobal << "\n");
821  }
822
823  ArrayType *ArrayOfGlobalStructTy = ArrayType::get(GlobalStructTy, n);
824  GlobalVariable *AllGlobals = new GlobalVariable(
825      M, ArrayOfGlobalStructTy, false, GlobalVariable::PrivateLinkage,
826      ConstantArray::get(ArrayOfGlobalStructTy, Initializers), "");
827
828  // Create calls for poisoning before initializers run and unpoisoning after.
829  if (CheckInitOrder && FirstDynamic && LastDynamic)
830    createInitializerPoisonCalls(M, FirstDynamic, LastDynamic);
831
832  Function *AsanRegisterGlobals = checkInterfaceFunction(M.getOrInsertFunction(
833      kAsanRegisterGlobalsName, IRB.getVoidTy(),
834      IntptrTy, IntptrTy, NULL));
835  AsanRegisterGlobals->setLinkage(Function::ExternalLinkage);
836
837  IRB.CreateCall2(AsanRegisterGlobals,
838                  IRB.CreatePointerCast(AllGlobals, IntptrTy),
839                  ConstantInt::get(IntptrTy, n));
840
841  // We also need to unregister globals at the end, e.g. when a shared library
842  // gets closed.
843  Function *AsanDtorFunction = Function::Create(
844      FunctionType::get(Type::getVoidTy(*C), false),
845      GlobalValue::InternalLinkage, kAsanModuleDtorName, &M);
846  BasicBlock *AsanDtorBB = BasicBlock::Create(*C, "", AsanDtorFunction);
847  IRBuilder<> IRB_Dtor(ReturnInst::Create(*C, AsanDtorBB));
848  Function *AsanUnregisterGlobals =
849      checkInterfaceFunction(M.getOrInsertFunction(
850          kAsanUnregisterGlobalsName,
851          IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
852  AsanUnregisterGlobals->setLinkage(Function::ExternalLinkage);
853
854  IRB_Dtor.CreateCall2(AsanUnregisterGlobals,
855                       IRB.CreatePointerCast(AllGlobals, IntptrTy),
856                       ConstantInt::get(IntptrTy, n));
857  appendToGlobalDtors(M, AsanDtorFunction, kAsanCtorAndCtorPriority);
858
859  DEBUG(dbgs() << M);
860  return true;
861}
862
863void AddressSanitizer::initializeCallbacks(Module &M) {
864  IRBuilder<> IRB(*C);
865  // Create __asan_report* callbacks.
866  for (size_t AccessIsWrite = 0; AccessIsWrite <= 1; AccessIsWrite++) {
867    for (size_t AccessSizeIndex = 0; AccessSizeIndex < kNumberOfAccessSizes;
868         AccessSizeIndex++) {
869      // IsWrite and TypeSize are encoded in the function name.
870      std::string FunctionName = std::string(kAsanReportErrorTemplate) +
871          (AccessIsWrite ? "store" : "load") + itostr(1 << AccessSizeIndex);
872      // If we are merging crash callbacks, they have two parameters.
873      AsanErrorCallback[AccessIsWrite][AccessSizeIndex] =
874          checkInterfaceFunction(M.getOrInsertFunction(
875              FunctionName, IRB.getVoidTy(), IntptrTy, NULL));
876    }
877  }
878
879  AsanHandleNoReturnFunc = checkInterfaceFunction(M.getOrInsertFunction(
880      kAsanHandleNoReturnName, IRB.getVoidTy(), NULL));
881  // We insert an empty inline asm after __asan_report* to avoid callback merge.
882  EmptyAsm = InlineAsm::get(FunctionType::get(IRB.getVoidTy(), false),
883                            StringRef(""), StringRef(""),
884                            /*hasSideEffects=*/true);
885}
886
887// virtual
888bool AddressSanitizer::doInitialization(Module &M) {
889  // Initialize the private fields. No one has accessed them before.
890  TD = getAnalysisIfAvailable<DataLayout>();
891
892  if (!TD)
893    return false;
894  BL.reset(new BlackList(BlacklistFile));
895  DynamicallyInitializedGlobals.Init(M);
896
897  C = &(M.getContext());
898  LongSize = TD->getPointerSizeInBits();
899  IntptrTy = Type::getIntNTy(*C, LongSize);
900
901  AsanCtorFunction = Function::Create(
902      FunctionType::get(Type::getVoidTy(*C), false),
903      GlobalValue::InternalLinkage, kAsanModuleCtorName, &M);
904  BasicBlock *AsanCtorBB = BasicBlock::Create(*C, "", AsanCtorFunction);
905  // call __asan_init in the module ctor.
906  IRBuilder<> IRB(ReturnInst::Create(*C, AsanCtorBB));
907  AsanInitFunction = checkInterfaceFunction(
908      M.getOrInsertFunction(kAsanInitName, IRB.getVoidTy(), NULL));
909  AsanInitFunction->setLinkage(Function::ExternalLinkage);
910  IRB.CreateCall(AsanInitFunction);
911
912  llvm::Triple targetTriple(M.getTargetTriple());
913  bool isAndroid = targetTriple.getEnvironment() == llvm::Triple::Android;
914
915  MappingOffset = isAndroid ? kDefaultShadowOffsetAndroid :
916    (LongSize == 32 ? kDefaultShadowOffset32 : kDefaultShadowOffset64);
917  if (ClMappingOffsetLog >= 0) {
918    if (ClMappingOffsetLog == 0) {
919      // special case
920      MappingOffset = 0;
921    } else {
922      MappingOffset = 1ULL << ClMappingOffsetLog;
923    }
924  }
925
926
927  if (ClMappingOffsetLog >= 0) {
928    // Tell the run-time the current values of mapping offset and scale.
929    GlobalValue *asan_mapping_offset =
930        new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
931                       ConstantInt::get(IntptrTy, MappingOffset),
932                       kAsanMappingOffsetName);
933    // Read the global, otherwise it may be optimized away.
934    IRB.CreateLoad(asan_mapping_offset, true);
935  }
936  if (ClMappingScale) {
937    GlobalValue *asan_mapping_scale =
938        new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
939                           ConstantInt::get(IntptrTy, MappingScale()),
940                           kAsanMappingScaleName);
941    // Read the global, otherwise it may be optimized away.
942    IRB.CreateLoad(asan_mapping_scale, true);
943  }
944
945  appendToGlobalCtors(M, AsanCtorFunction, kAsanCtorAndCtorPriority);
946
947  return true;
948}
949
950bool AddressSanitizer::maybeInsertAsanInitAtFunctionEntry(Function &F) {
951  // For each NSObject descendant having a +load method, this method is invoked
952  // by the ObjC runtime before any of the static constructors is called.
953  // Therefore we need to instrument such methods with a call to __asan_init
954  // at the beginning in order to initialize our runtime before any access to
955  // the shadow memory.
956  // We cannot just ignore these methods, because they may call other
957  // instrumented functions.
958  if (F.getName().find(" load]") != std::string::npos) {
959    IRBuilder<> IRB(F.begin()->begin());
960    IRB.CreateCall(AsanInitFunction);
961    return true;
962  }
963  return false;
964}
965
966bool AddressSanitizer::runOnFunction(Function &F) {
967  if (BL->isIn(F)) return false;
968  if (&F == AsanCtorFunction) return false;
969  DEBUG(dbgs() << "ASAN instrumenting:\n" << F << "\n");
970  initializeCallbacks(*F.getParent());
971
972  // If needed, insert __asan_init before checking for AddressSafety attr.
973  maybeInsertAsanInitAtFunctionEntry(F);
974
975  if (!F.getFnAttributes().hasAttribute(Attribute::AddressSafety))
976    return false;
977
978  if (!ClDebugFunc.empty() && ClDebugFunc != F.getName())
979    return false;
980
981  // We want to instrument every address only once per basic block (unless there
982  // are calls between uses).
983  SmallSet<Value*, 16> TempsToInstrument;
984  SmallVector<Instruction*, 16> ToInstrument;
985  SmallVector<Instruction*, 8> NoReturnCalls;
986  bool IsWrite;
987
988  // Fill the set of memory operations to instrument.
989  for (Function::iterator FI = F.begin(), FE = F.end();
990       FI != FE; ++FI) {
991    TempsToInstrument.clear();
992    int NumInsnsPerBB = 0;
993    for (BasicBlock::iterator BI = FI->begin(), BE = FI->end();
994         BI != BE; ++BI) {
995      if (LooksLikeCodeInBug11395(BI)) return false;
996      if (Value *Addr = isInterestingMemoryAccess(BI, &IsWrite)) {
997        if (ClOpt && ClOptSameTemp) {
998          if (!TempsToInstrument.insert(Addr))
999            continue;  // We've seen this temp in the current BB.
1000        }
1001      } else if (isa<MemIntrinsic>(BI) && ClMemIntrin) {
1002        // ok, take it.
1003      } else {
1004        if (CallInst *CI = dyn_cast<CallInst>(BI)) {
1005          // A call inside BB.
1006          TempsToInstrument.clear();
1007          if (CI->doesNotReturn()) {
1008            NoReturnCalls.push_back(CI);
1009          }
1010        }
1011        continue;
1012      }
1013      ToInstrument.push_back(BI);
1014      NumInsnsPerBB++;
1015      if (NumInsnsPerBB >= ClMaxInsnsToInstrumentPerBB)
1016        break;
1017    }
1018  }
1019
1020  // Instrument.
1021  int NumInstrumented = 0;
1022  for (size_t i = 0, n = ToInstrument.size(); i != n; i++) {
1023    Instruction *Inst = ToInstrument[i];
1024    if (ClDebugMin < 0 || ClDebugMax < 0 ||
1025        (NumInstrumented >= ClDebugMin && NumInstrumented <= ClDebugMax)) {
1026      if (isInterestingMemoryAccess(Inst, &IsWrite))
1027        instrumentMop(Inst);
1028      else
1029        instrumentMemIntrinsic(cast<MemIntrinsic>(Inst));
1030    }
1031    NumInstrumented++;
1032  }
1033
1034  FunctionStackPoisoner FSP(F, *this);
1035  bool ChangedStack = FSP.runOnFunction();
1036
1037  // We must unpoison the stack before every NoReturn call (throw, _exit, etc).
1038  // See e.g. http://code.google.com/p/address-sanitizer/issues/detail?id=37
1039  for (size_t i = 0, n = NoReturnCalls.size(); i != n; i++) {
1040    Instruction *CI = NoReturnCalls[i];
1041    IRBuilder<> IRB(CI);
1042    IRB.CreateCall(AsanHandleNoReturnFunc);
1043  }
1044  DEBUG(dbgs() << "ASAN done instrumenting:\n" << F << "\n");
1045
1046  return NumInstrumented > 0 || ChangedStack || !NoReturnCalls.empty();
1047}
1048
1049static uint64_t ValueForPoison(uint64_t PoisonByte, size_t ShadowRedzoneSize) {
1050  if (ShadowRedzoneSize == 1) return PoisonByte;
1051  if (ShadowRedzoneSize == 2) return (PoisonByte << 8) + PoisonByte;
1052  if (ShadowRedzoneSize == 4)
1053    return (PoisonByte << 24) + (PoisonByte << 16) +
1054        (PoisonByte << 8) + (PoisonByte);
1055  llvm_unreachable("ShadowRedzoneSize is either 1, 2 or 4");
1056}
1057
1058static void PoisonShadowPartialRightRedzone(uint8_t *Shadow,
1059                                            size_t Size,
1060                                            size_t RZSize,
1061                                            size_t ShadowGranularity,
1062                                            uint8_t Magic) {
1063  for (size_t i = 0; i < RZSize;
1064       i+= ShadowGranularity, Shadow++) {
1065    if (i + ShadowGranularity <= Size) {
1066      *Shadow = 0;  // fully addressable
1067    } else if (i >= Size) {
1068      *Shadow = Magic;  // unaddressable
1069    } else {
1070      *Shadow = Size - i;  // first Size-i bytes are addressable
1071    }
1072  }
1073}
1074
1075// Workaround for bug 11395: we don't want to instrument stack in functions
1076// with large assembly blobs (32-bit only), otherwise reg alloc may crash.
1077// FIXME: remove once the bug 11395 is fixed.
1078bool AddressSanitizer::LooksLikeCodeInBug11395(Instruction *I) {
1079  if (LongSize != 32) return false;
1080  CallInst *CI = dyn_cast<CallInst>(I);
1081  if (!CI || !CI->isInlineAsm()) return false;
1082  if (CI->getNumArgOperands() <= 5) return false;
1083  // We have inline assembly with quite a few arguments.
1084  return true;
1085}
1086
1087void FunctionStackPoisoner::initializeCallbacks(Module &M) {
1088  IRBuilder<> IRB(*C);
1089  AsanStackMallocFunc = checkInterfaceFunction(M.getOrInsertFunction(
1090      kAsanStackMallocName, IntptrTy, IntptrTy, IntptrTy, NULL));
1091  AsanStackFreeFunc = checkInterfaceFunction(M.getOrInsertFunction(
1092      kAsanStackFreeName, IRB.getVoidTy(),
1093      IntptrTy, IntptrTy, IntptrTy, NULL));
1094  AsanPoisonStackMemoryFunc = checkInterfaceFunction(M.getOrInsertFunction(
1095      kAsanPoisonStackMemoryName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
1096  AsanUnpoisonStackMemoryFunc = checkInterfaceFunction(M.getOrInsertFunction(
1097      kAsanUnpoisonStackMemoryName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
1098}
1099
1100void FunctionStackPoisoner::poisonRedZones(
1101  const ArrayRef<AllocaInst*> &AllocaVec, IRBuilder<> IRB, Value *ShadowBase,
1102  bool DoPoison) {
1103  size_t ShadowRZSize = RedzoneSize() >> MappingScale();
1104  assert(ShadowRZSize >= 1 && ShadowRZSize <= 4);
1105  Type *RZTy = Type::getIntNTy(*C, ShadowRZSize * 8);
1106  Type *RZPtrTy = PointerType::get(RZTy, 0);
1107
1108  Value *PoisonLeft  = ConstantInt::get(RZTy,
1109    ValueForPoison(DoPoison ? kAsanStackLeftRedzoneMagic : 0LL, ShadowRZSize));
1110  Value *PoisonMid   = ConstantInt::get(RZTy,
1111    ValueForPoison(DoPoison ? kAsanStackMidRedzoneMagic : 0LL, ShadowRZSize));
1112  Value *PoisonRight = ConstantInt::get(RZTy,
1113    ValueForPoison(DoPoison ? kAsanStackRightRedzoneMagic : 0LL, ShadowRZSize));
1114
1115  // poison the first red zone.
1116  IRB.CreateStore(PoisonLeft, IRB.CreateIntToPtr(ShadowBase, RZPtrTy));
1117
1118  // poison all other red zones.
1119  uint64_t Pos = RedzoneSize();
1120  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
1121    AllocaInst *AI = AllocaVec[i];
1122    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
1123    uint64_t AlignedSize = getAlignedAllocaSize(AI);
1124    assert(AlignedSize - SizeInBytes < RedzoneSize());
1125    Value *Ptr = NULL;
1126
1127    Pos += AlignedSize;
1128
1129    assert(ShadowBase->getType() == IntptrTy);
1130    if (SizeInBytes < AlignedSize) {
1131      // Poison the partial redzone at right
1132      Ptr = IRB.CreateAdd(
1133          ShadowBase, ConstantInt::get(IntptrTy,
1134                                       (Pos >> MappingScale()) - ShadowRZSize));
1135      size_t AddressableBytes = RedzoneSize() - (AlignedSize - SizeInBytes);
1136      uint32_t Poison = 0;
1137      if (DoPoison) {
1138        PoisonShadowPartialRightRedzone((uint8_t*)&Poison, AddressableBytes,
1139                                        RedzoneSize(),
1140                                        1ULL << MappingScale(),
1141                                        kAsanStackPartialRedzoneMagic);
1142      }
1143      Value *PartialPoison = ConstantInt::get(RZTy, Poison);
1144      IRB.CreateStore(PartialPoison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
1145    }
1146
1147    // Poison the full redzone at right.
1148    Ptr = IRB.CreateAdd(ShadowBase,
1149                        ConstantInt::get(IntptrTy, Pos >> MappingScale()));
1150    Value *Poison = i == AllocaVec.size() - 1 ? PoisonRight : PoisonMid;
1151    IRB.CreateStore(Poison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
1152
1153    Pos += RedzoneSize();
1154  }
1155}
1156
1157void FunctionStackPoisoner::poisonStack() {
1158  bool HavePoisonedAllocas = false;
1159  uint64_t LocalStackSize = TotalStackSize +
1160                            (AllocaVec.size() + 1) * RedzoneSize();
1161
1162  bool DoStackMalloc = ASan.CheckUseAfterReturn
1163      && LocalStackSize <= kMaxStackMallocSize;
1164
1165  Instruction *InsBefore = AllocaVec[0];
1166  IRBuilder<> IRB(InsBefore);
1167
1168
1169  Type *ByteArrayTy = ArrayType::get(IRB.getInt8Ty(), LocalStackSize);
1170  AllocaInst *MyAlloca =
1171      new AllocaInst(ByteArrayTy, "MyAlloca", InsBefore);
1172  if (ClRealignStack && StackAlignment < RedzoneSize())
1173    StackAlignment = RedzoneSize();
1174  MyAlloca->setAlignment(StackAlignment);
1175  assert(MyAlloca->isStaticAlloca());
1176  Value *OrigStackBase = IRB.CreatePointerCast(MyAlloca, IntptrTy);
1177  Value *LocalStackBase = OrigStackBase;
1178
1179  if (DoStackMalloc) {
1180    LocalStackBase = IRB.CreateCall2(AsanStackMallocFunc,
1181        ConstantInt::get(IntptrTy, LocalStackSize), OrigStackBase);
1182  }
1183
1184  // This string will be parsed by the run-time (DescribeStackAddress).
1185  SmallString<2048> StackDescriptionStorage;
1186  raw_svector_ostream StackDescription(StackDescriptionStorage);
1187  StackDescription << F.getName() << " " << AllocaVec.size() << " ";
1188
1189  uint64_t Pos = RedzoneSize();
1190  // Replace Alloca instructions with base+offset.
1191  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
1192    AllocaInst *AI = AllocaVec[i];
1193    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
1194    StringRef Name = AI->getName();
1195    StackDescription << Pos << " " << SizeInBytes << " "
1196                     << Name.size() << " " << Name << " ";
1197    uint64_t AlignedSize = getAlignedAllocaSize(AI);
1198    assert((AlignedSize % RedzoneSize()) == 0);
1199    Value *NewAllocaPtr = IRB.CreateIntToPtr(
1200            IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy, Pos)),
1201            AI->getType());
1202    replaceDbgDeclareForAlloca(AI, NewAllocaPtr, DIB);
1203    AI->replaceAllUsesWith(NewAllocaPtr);
1204    // Analyze lifetime intrinsics only for static allocas we handle.
1205    if (ASan.CheckLifetime)
1206      HavePoisonedAllocas |= handleAllocaLifetime(NewAllocaPtr);
1207    Pos += AlignedSize + RedzoneSize();
1208  }
1209  assert(Pos == LocalStackSize);
1210
1211  // Write the Magic value and the frame description constant to the redzone.
1212  Value *BasePlus0 = IRB.CreateIntToPtr(LocalStackBase, IntptrPtrTy);
1213  IRB.CreateStore(ConstantInt::get(IntptrTy, kCurrentStackFrameMagic),
1214                  BasePlus0);
1215  Value *BasePlus1 = IRB.CreateAdd(LocalStackBase,
1216                                   ConstantInt::get(IntptrTy,
1217                                                    ASan.LongSize/8));
1218  BasePlus1 = IRB.CreateIntToPtr(BasePlus1, IntptrPtrTy);
1219  GlobalVariable *StackDescriptionGlobal =
1220      createPrivateGlobalForString(*F.getParent(), StackDescription.str());
1221  Value *Description = IRB.CreatePointerCast(StackDescriptionGlobal,
1222                                             IntptrTy);
1223  IRB.CreateStore(Description, BasePlus1);
1224
1225  // Poison the stack redzones at the entry.
1226  Value *ShadowBase = ASan.memToShadow(LocalStackBase, IRB);
1227  poisonRedZones(AllocaVec, IRB, ShadowBase, true);
1228
1229  // Unpoison the stack before all ret instructions.
1230  for (size_t i = 0, n = RetVec.size(); i < n; i++) {
1231    Instruction *Ret = RetVec[i];
1232    IRBuilder<> IRBRet(Ret);
1233    // Mark the current frame as retired.
1234    IRBRet.CreateStore(ConstantInt::get(IntptrTy, kRetiredStackFrameMagic),
1235                       BasePlus0);
1236    // Unpoison the stack.
1237    poisonRedZones(AllocaVec, IRBRet, ShadowBase, false);
1238    if (DoStackMalloc) {
1239      // In use-after-return mode, mark the whole stack frame unaddressable.
1240      IRBRet.CreateCall3(AsanStackFreeFunc, LocalStackBase,
1241                         ConstantInt::get(IntptrTy, LocalStackSize),
1242                         OrigStackBase);
1243    } else if (HavePoisonedAllocas) {
1244      // If we poisoned some allocas in llvm.lifetime analysis,
1245      // unpoison whole stack frame now.
1246      assert(LocalStackBase == OrigStackBase);
1247      poisonAlloca(LocalStackBase, LocalStackSize, IRBRet, false);
1248    }
1249  }
1250
1251  // We are done. Remove the old unused alloca instructions.
1252  for (size_t i = 0, n = AllocaVec.size(); i < n; i++)
1253    AllocaVec[i]->eraseFromParent();
1254}
1255
1256void FunctionStackPoisoner::poisonAlloca(Value *V, uint64_t Size,
1257                                         IRBuilder<> IRB, bool DoPoison) {
1258  // For now just insert the call to ASan runtime.
1259  Value *AddrArg = IRB.CreatePointerCast(V, IntptrTy);
1260  Value *SizeArg = ConstantInt::get(IntptrTy, Size);
1261  IRB.CreateCall2(DoPoison ? AsanPoisonStackMemoryFunc
1262                           : AsanUnpoisonStackMemoryFunc,
1263                  AddrArg, SizeArg);
1264}
1265
1266// Handling llvm.lifetime intrinsics for a given %alloca:
1267// (1) collect all llvm.lifetime.xxx(%size, %value) describing the alloca.
1268// (2) if %size is constant, poison memory for llvm.lifetime.end (to detect
1269//     invalid accesses) and unpoison it for llvm.lifetime.start (the memory
1270//     could be poisoned by previous llvm.lifetime.end instruction, as the
1271//     variable may go in and out of scope several times, e.g. in loops).
1272// (3) if we poisoned at least one %alloca in a function,
1273//     unpoison the whole stack frame at function exit.
1274bool FunctionStackPoisoner::handleAllocaLifetime(Value *Alloca) {
1275  assert(ASan.CheckLifetime);
1276  Type *AllocaType = Alloca->getType();
1277  Type *Int8PtrTy = Type::getInt8PtrTy(AllocaType->getContext());
1278
1279  bool Res = false;
1280  // Typical code looks like this:
1281  // %alloca = alloca <type>, <alignment>
1282  // ... some code ...
1283  // %val1 = bitcast <type>* %alloca to i8*
1284  // call void @llvm.lifetime.start(i64 <size>, i8* %val1)
1285  // ... more code ...
1286  // %val2 = bitcast <type>* %alloca to i8*
1287  // call void @llvm.lifetime.start(i64 <size>, i8* %val2)
1288  // That is, to handle %alloca we must find all its casts to
1289  // i8* values, and find lifetime instructions for these values.
1290  if (AllocaType == Int8PtrTy)
1291    Res |= handleValueLifetime(Alloca);
1292  for (Value::use_iterator UI = Alloca->use_begin(), UE = Alloca->use_end();
1293       UI != UE; ++UI) {
1294    if (UI->getType() != Int8PtrTy) continue;
1295    if (UI->stripPointerCasts() != Alloca) continue;
1296    Res |= handleValueLifetime(*UI);
1297  }
1298  return Res;
1299}
1300
1301bool FunctionStackPoisoner::handleValueLifetime(Value *V) {
1302  assert(ASan.CheckLifetime);
1303  bool Res = false;
1304  for (Value::use_iterator UI = V->use_begin(), UE = V->use_end(); UI != UE;
1305       ++UI) {
1306    IntrinsicInst *II = dyn_cast<IntrinsicInst>(*UI);
1307    if (!II) continue;
1308    Intrinsic::ID ID = II->getIntrinsicID();
1309    if (ID != Intrinsic::lifetime_start &&
1310        ID != Intrinsic::lifetime_end)
1311      continue;
1312    if (V != II->getArgOperand(1))
1313      continue;
1314    // Found lifetime intrinsic, add ASan instrumentation if necessary.
1315    ConstantInt *Size = dyn_cast<ConstantInt>(II->getArgOperand(0));
1316    // If size argument is undefined, don't do anything.
1317    if (Size->isMinusOne())
1318      continue;
1319    // Check that size doesn't saturate uint64_t and can
1320    // be stored in IntptrTy.
1321    const uint64_t SizeValue = Size->getValue().getLimitedValue();
1322    if (SizeValue == ~0ULL ||
1323        !ConstantInt::isValueValidForType(IntptrTy, SizeValue)) {
1324      continue;
1325    }
1326    IRBuilder<> IRB(II);
1327    bool DoPoison = (ID == Intrinsic::lifetime_end);
1328    poisonAlloca(V, SizeValue, IRB, DoPoison);
1329    Res = true;
1330  }
1331  return Res;
1332}
1333