AddressSanitizer.cpp revision c8a196ae8fad3cba7a777e2e7916fd36ebf70fe6 
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 "llvm/ADT/ArrayRef.h"
20#include "llvm/ADT/DenseMap.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/DIBuilder.h"
29#include "llvm/IR/DataLayout.h"
30#include "llvm/IR/Function.h"
31#include "llvm/IR/IRBuilder.h"
32#include "llvm/IR/InlineAsm.h"
33#include "llvm/IR/IntrinsicInst.h"
34#include "llvm/IR/LLVMContext.h"
35#include "llvm/IR/Module.h"
36#include "llvm/IR/Type.h"
37#include "llvm/InstVisitor.h"
38#include "llvm/Support/CommandLine.h"
39#include "llvm/Support/DataTypes.h"
40#include "llvm/Support/Debug.h"
41#include "llvm/Support/raw_ostream.h"
42#include "llvm/Support/system_error.h"
43#include "llvm/Target/TargetMachine.h"
44#include "llvm/Transforms/Utils/BasicBlockUtils.h"
45#include "llvm/Transforms/Utils/BlackList.h"
46#include "llvm/Transforms/Utils/Local.h"
47#include "llvm/Transforms/Utils/ModuleUtils.h"
48#include <algorithm>
49#include <string>
50
51using namespace llvm;
52
53static const uint64_t kDefaultShadowScale = 3;
54static const uint64_t kDefaultShadowOffset32 = 1ULL << 29;
55static const uint64_t kDefaultShadowOffset64 = 1ULL << 44;
56static const uint64_t kDefaultShort64bitShadowOffset = 0x7FFF8000;  // < 2G.
57static const uint64_t kPPC64_ShadowOffset64 = 1ULL << 41;
58
59static const size_t kMaxStackMallocSize = 1 << 16;  // 64K
60static const uintptr_t kCurrentStackFrameMagic = 0x41B58AB3;
61static const uintptr_t kRetiredStackFrameMagic = 0x45E0360E;
62
63static const char *kAsanModuleCtorName = "asan.module_ctor";
64static const char *kAsanModuleDtorName = "asan.module_dtor";
65static const int   kAsanCtorAndCtorPriority = 1;
66static const char *kAsanReportErrorTemplate = "__asan_report_";
67static const char *kAsanRegisterGlobalsName = "__asan_register_globals";
68static const char *kAsanUnregisterGlobalsName = "__asan_unregister_globals";
69static const char *kAsanPoisonGlobalsName = "__asan_before_dynamic_init";
70static const char *kAsanUnpoisonGlobalsName = "__asan_after_dynamic_init";
71static const char *kAsanInitName = "__asan_init_v1";
72static const char *kAsanHandleNoReturnName = "__asan_handle_no_return";
73static const char *kAsanMappingOffsetName = "__asan_mapping_offset";
74static const char *kAsanMappingScaleName = "__asan_mapping_scale";
75static const char *kAsanStackMallocName = "__asan_stack_malloc";
76static const char *kAsanStackFreeName = "__asan_stack_free";
77static const char *kAsanGenPrefix = "__asan_gen_";
78static const char *kAsanPoisonStackMemoryName = "__asan_poison_stack_memory";
79static const char *kAsanUnpoisonStackMemoryName =
80    "__asan_unpoison_stack_memory";
81
82static const int kAsanStackLeftRedzoneMagic = 0xf1;
83static const int kAsanStackMidRedzoneMagic = 0xf2;
84static const int kAsanStackRightRedzoneMagic = 0xf3;
85static const int kAsanStackPartialRedzoneMagic = 0xf4;
86
87// Accesses sizes are powers of two: 1, 2, 4, 8, 16.
88static const size_t kNumberOfAccessSizes = 5;
89
90// Command-line flags.
91
92// This flag may need to be replaced with -f[no-]asan-reads.
93static cl::opt<bool> ClInstrumentReads("asan-instrument-reads",
94       cl::desc("instrument read instructions"), cl::Hidden, cl::init(true));
95static cl::opt<bool> ClInstrumentWrites("asan-instrument-writes",
96       cl::desc("instrument write instructions"), cl::Hidden, cl::init(true));
97static cl::opt<bool> ClInstrumentAtomics("asan-instrument-atomics",
98       cl::desc("instrument atomic instructions (rmw, cmpxchg)"),
99       cl::Hidden, cl::init(true));
100static cl::opt<bool> ClAlwaysSlowPath("asan-always-slow-path",
101       cl::desc("use instrumentation with slow path for all accesses"),
102       cl::Hidden, cl::init(false));
103// This flag limits the number of instructions to be instrumented
104// in any given BB. Normally, this should be set to unlimited (INT_MAX),
105// but due to http://llvm.org/bugs/show_bug.cgi?id=12652 we temporary
106// set it to 10000.
107static cl::opt<int> ClMaxInsnsToInstrumentPerBB("asan-max-ins-per-bb",
108       cl::init(10000),
109       cl::desc("maximal number of instructions to instrument in any given BB"),
110       cl::Hidden);
111// This flag may need to be replaced with -f[no]asan-stack.
112static cl::opt<bool> ClStack("asan-stack",
113       cl::desc("Handle stack memory"), cl::Hidden, cl::init(true));
114// This flag may need to be replaced with -f[no]asan-use-after-return.
115static cl::opt<bool> ClUseAfterReturn("asan-use-after-return",
116       cl::desc("Check return-after-free"), cl::Hidden, cl::init(false));
117// This flag may need to be replaced with -f[no]asan-globals.
118static cl::opt<bool> ClGlobals("asan-globals",
119       cl::desc("Handle global objects"), cl::Hidden, cl::init(true));
120static cl::opt<bool> ClInitializers("asan-initialization-order",
121       cl::desc("Handle C++ initializer order"), cl::Hidden, cl::init(false));
122static cl::opt<bool> ClMemIntrin("asan-memintrin",
123       cl::desc("Handle memset/memcpy/memmove"), cl::Hidden, cl::init(true));
124static cl::opt<bool> ClRealignStack("asan-realign-stack",
125       cl::desc("Realign stack to 32"), cl::Hidden, cl::init(true));
126static cl::opt<std::string> ClBlacklistFile("asan-blacklist",
127       cl::desc("File containing the list of objects to ignore "
128                "during instrumentation"), cl::Hidden);
129
130// These flags allow to change the shadow mapping.
131// The shadow mapping looks like
132//    Shadow = (Mem >> scale) + (1 << offset_log)
133static cl::opt<int> ClMappingScale("asan-mapping-scale",
134       cl::desc("scale of asan shadow mapping"), cl::Hidden, cl::init(0));
135static cl::opt<int> ClMappingOffsetLog("asan-mapping-offset-log",
136       cl::desc("offset of asan shadow mapping"), cl::Hidden, cl::init(-1));
137static cl::opt<bool> ClShort64BitOffset("asan-short-64bit-mapping-offset",
138       cl::desc("Use short immediate constant as the mapping offset for 64bit"),
139       cl::Hidden, cl::init(true));
140
141// Optimization flags. Not user visible, used mostly for testing
142// and benchmarking the tool.
143static cl::opt<bool> ClOpt("asan-opt",
144       cl::desc("Optimize instrumentation"), cl::Hidden, cl::init(true));
145static cl::opt<bool> ClOptSameTemp("asan-opt-same-temp",
146       cl::desc("Instrument the same temp just once"), cl::Hidden,
147       cl::init(true));
148static cl::opt<bool> ClOptGlobals("asan-opt-globals",
149       cl::desc("Don't instrument scalar globals"), cl::Hidden, cl::init(true));
150
151static cl::opt<bool> ClCheckLifetime("asan-check-lifetime",
152       cl::desc("Use llvm.lifetime intrinsics to insert extra checks"),
153       cl::Hidden, cl::init(false));
154
155// Debug flags.
156static cl::opt<int> ClDebug("asan-debug", cl::desc("debug"), cl::Hidden,
157                            cl::init(0));
158static cl::opt<int> ClDebugStack("asan-debug-stack", cl::desc("debug stack"),
159                                 cl::Hidden, cl::init(0));
160static cl::opt<std::string> ClDebugFunc("asan-debug-func",
161                                        cl::Hidden, cl::desc("Debug func"));
162static cl::opt<int> ClDebugMin("asan-debug-min", cl::desc("Debug min inst"),
163                               cl::Hidden, cl::init(-1));
164static cl::opt<int> ClDebugMax("asan-debug-max", cl::desc("Debug man inst"),
165                               cl::Hidden, cl::init(-1));
166
167namespace {
168/// A set of dynamically initialized globals extracted from metadata.
169class SetOfDynamicallyInitializedGlobals {
170 public:
171  void Init(Module& M) {
172    // Clang generates metadata identifying all dynamically initialized globals.
173    NamedMDNode *DynamicGlobals =
174        M.getNamedMetadata("llvm.asan.dynamically_initialized_globals");
175    if (!DynamicGlobals)
176      return;
177    for (int i = 0, n = DynamicGlobals->getNumOperands(); i < n; ++i) {
178      MDNode *MDN = DynamicGlobals->getOperand(i);
179      assert(MDN->getNumOperands() == 1);
180      Value *VG = MDN->getOperand(0);
181      // The optimizer may optimize away a global entirely, in which case we
182      // cannot instrument access to it.
183      if (!VG)
184        continue;
185      DynInitGlobals.insert(cast<GlobalVariable>(VG));
186    }
187  }
188  bool Contains(GlobalVariable *G) { return DynInitGlobals.count(G) != 0; }
189 private:
190  SmallSet<GlobalValue*, 32> DynInitGlobals;
191};
192
193/// This struct defines the shadow mapping using the rule:
194///   shadow = (mem >> Scale) ADD-or-OR Offset.
195struct ShadowMapping {
196  int Scale;
197  uint64_t Offset;
198  bool OrShadowOffset;
199};
200
201static ShadowMapping getShadowMapping(const Module &M, int LongSize,
202                                      bool ZeroBaseShadow) {
203  llvm::Triple TargetTriple(M.getTargetTriple());
204  bool IsAndroid = TargetTriple.getEnvironment() == llvm::Triple::Android;
205  bool IsMacOSX = TargetTriple.getOS() == llvm::Triple::MacOSX;
206  bool IsPPC64 = TargetTriple.getArch() == llvm::Triple::ppc64;
207  bool IsX86_64 = TargetTriple.getArch() == llvm::Triple::x86_64;
208
209  ShadowMapping Mapping;
210
211  // OR-ing shadow offset if more efficient (at least on x86),
212  // but on ppc64 we have to use add since the shadow offset is not neccesary
213  // 1/8-th of the address space.
214  Mapping.OrShadowOffset = !IsPPC64 && !ClShort64BitOffset;
215
216  Mapping.Offset = (IsAndroid || ZeroBaseShadow) ? 0 :
217      (LongSize == 32 ? kDefaultShadowOffset32 :
218       IsPPC64 ? kPPC64_ShadowOffset64 : kDefaultShadowOffset64);
219  if (!ZeroBaseShadow && ClShort64BitOffset && IsX86_64 && !IsMacOSX) {
220    assert(LongSize == 64);
221    Mapping.Offset = kDefaultShort64bitShadowOffset;
222  } if (!ZeroBaseShadow && ClMappingOffsetLog >= 0) {
223    // Zero offset log is the special case.
224    Mapping.Offset = (ClMappingOffsetLog == 0) ? 0 : 1ULL << ClMappingOffsetLog;
225  }
226
227  Mapping.Scale = kDefaultShadowScale;
228  if (ClMappingScale) {
229    Mapping.Scale = ClMappingScale;
230  }
231
232  return Mapping;
233}
234
235static size_t RedzoneSizeForScale(int MappingScale) {
236  // Redzone used for stack and globals is at least 32 bytes.
237  // For scales 6 and 7, the redzone has to be 64 and 128 bytes respectively.
238  return std::max(32U, 1U << MappingScale);
239}
240
241/// AddressSanitizer: instrument the code in module to find memory bugs.
242struct AddressSanitizer : public FunctionPass {
243  AddressSanitizer(bool CheckInitOrder = false,
244                   bool CheckUseAfterReturn = false,
245                   bool CheckLifetime = false,
246                   StringRef BlacklistFile = StringRef(),
247                   bool ZeroBaseShadow = false)
248      : FunctionPass(ID),
249        CheckInitOrder(CheckInitOrder || ClInitializers),
250        CheckUseAfterReturn(CheckUseAfterReturn || ClUseAfterReturn),
251        CheckLifetime(CheckLifetime || ClCheckLifetime),
252        BlacklistFile(BlacklistFile.empty() ? ClBlacklistFile
253                                            : BlacklistFile),
254        ZeroBaseShadow(ZeroBaseShadow) {}
255  virtual const char *getPassName() const {
256    return "AddressSanitizerFunctionPass";
257  }
258  void instrumentMop(Instruction *I);
259  void instrumentAddress(Instruction *OrigIns, IRBuilder<> &IRB,
260                         Value *Addr, uint32_t TypeSize, bool IsWrite);
261  Value *createSlowPathCmp(IRBuilder<> &IRB, Value *AddrLong,
262                           Value *ShadowValue, uint32_t TypeSize);
263  Instruction *generateCrashCode(Instruction *InsertBefore, Value *Addr,
264                                 bool IsWrite, size_t AccessSizeIndex);
265  bool instrumentMemIntrinsic(MemIntrinsic *MI);
266  void instrumentMemIntrinsicParam(Instruction *OrigIns, Value *Addr,
267                                   Value *Size,
268                                   Instruction *InsertBefore, bool IsWrite);
269  Value *memToShadow(Value *Shadow, IRBuilder<> &IRB);
270  bool runOnFunction(Function &F);
271  void createInitializerPoisonCalls(Module &M,
272                                    Value *FirstAddr, Value *LastAddr);
273  bool maybeInsertAsanInitAtFunctionEntry(Function &F);
274  void emitShadowMapping(Module &M, IRBuilder<> &IRB) const;
275  virtual bool doInitialization(Module &M);
276  static char ID;  // Pass identification, replacement for typeid
277
278 private:
279  void initializeCallbacks(Module &M);
280
281  bool ShouldInstrumentGlobal(GlobalVariable *G);
282  bool LooksLikeCodeInBug11395(Instruction *I);
283  void FindDynamicInitializers(Module &M);
284
285  bool CheckInitOrder;
286  bool CheckUseAfterReturn;
287  bool CheckLifetime;
288  SmallString<64> BlacklistFile;
289  bool ZeroBaseShadow;
290
291  LLVMContext *C;
292  DataLayout *TD;
293  int LongSize;
294  Type *IntptrTy;
295  ShadowMapping Mapping;
296  Function *AsanCtorFunction;
297  Function *AsanInitFunction;
298  Function *AsanHandleNoReturnFunc;
299  OwningPtr<BlackList> BL;
300  // This array is indexed by AccessIsWrite and log2(AccessSize).
301  Function *AsanErrorCallback[2][kNumberOfAccessSizes];
302  InlineAsm *EmptyAsm;
303  SetOfDynamicallyInitializedGlobals DynamicallyInitializedGlobals;
304
305  friend struct FunctionStackPoisoner;
306};
307
308class AddressSanitizerModule : public ModulePass {
309 public:
310  AddressSanitizerModule(bool CheckInitOrder = false,
311                         StringRef BlacklistFile = StringRef(),
312                         bool ZeroBaseShadow = false)
313      : ModulePass(ID),
314        CheckInitOrder(CheckInitOrder || ClInitializers),
315        BlacklistFile(BlacklistFile.empty() ? ClBlacklistFile
316                                            : BlacklistFile),
317        ZeroBaseShadow(ZeroBaseShadow) {}
318  bool runOnModule(Module &M);
319  static char ID;  // Pass identification, replacement for typeid
320  virtual const char *getPassName() const {
321    return "AddressSanitizerModule";
322  }
323
324 private:
325  void initializeCallbacks(Module &M);
326
327  bool ShouldInstrumentGlobal(GlobalVariable *G);
328  void createInitializerPoisonCalls(Module &M, Value *FirstAddr,
329                                    Value *LastAddr);
330  size_t RedzoneSize() const {
331    return RedzoneSizeForScale(Mapping.Scale);
332  }
333
334  bool CheckInitOrder;
335  SmallString<64> BlacklistFile;
336  bool ZeroBaseShadow;
337
338  OwningPtr<BlackList> BL;
339  SetOfDynamicallyInitializedGlobals DynamicallyInitializedGlobals;
340  Type *IntptrTy;
341  LLVMContext *C;
342  DataLayout *TD;
343  ShadowMapping Mapping;
344  Function *AsanPoisonGlobals;
345  Function *AsanUnpoisonGlobals;
346  Function *AsanRegisterGlobals;
347  Function *AsanUnregisterGlobals;
348};
349
350// Stack poisoning does not play well with exception handling.
351// When an exception is thrown, we essentially bypass the code
352// that unpoisones the stack. This is why the run-time library has
353// to intercept __cxa_throw (as well as longjmp, etc) and unpoison the entire
354// stack in the interceptor. This however does not work inside the
355// actual function which catches the exception. Most likely because the
356// compiler hoists the load of the shadow value somewhere too high.
357// This causes asan to report a non-existing bug on 453.povray.
358// It sounds like an LLVM bug.
359struct FunctionStackPoisoner : public InstVisitor<FunctionStackPoisoner> {
360  Function &F;
361  AddressSanitizer &ASan;
362  DIBuilder DIB;
363  LLVMContext *C;
364  Type *IntptrTy;
365  Type *IntptrPtrTy;
366  ShadowMapping Mapping;
367
368  SmallVector<AllocaInst*, 16> AllocaVec;
369  SmallVector<Instruction*, 8> RetVec;
370  uint64_t TotalStackSize;
371  unsigned StackAlignment;
372
373  Function *AsanStackMallocFunc, *AsanStackFreeFunc;
374  Function *AsanPoisonStackMemoryFunc, *AsanUnpoisonStackMemoryFunc;
375
376  // Stores a place and arguments of poisoning/unpoisoning call for alloca.
377  struct AllocaPoisonCall {
378    IntrinsicInst *InsBefore;
379    uint64_t Size;
380    bool DoPoison;
381  };
382  SmallVector<AllocaPoisonCall, 8> AllocaPoisonCallVec;
383
384  // Maps Value to an AllocaInst from which the Value is originated.
385  typedef DenseMap<Value*, AllocaInst*> AllocaForValueMapTy;
386  AllocaForValueMapTy AllocaForValue;
387
388  FunctionStackPoisoner(Function &F, AddressSanitizer &ASan)
389      : F(F), ASan(ASan), DIB(*F.getParent()), C(ASan.C),
390        IntptrTy(ASan.IntptrTy), IntptrPtrTy(PointerType::get(IntptrTy, 0)),
391        Mapping(ASan.Mapping),
392        TotalStackSize(0), StackAlignment(1 << Mapping.Scale) {}
393
394  bool runOnFunction() {
395    if (!ClStack) return false;
396    // Collect alloca, ret, lifetime instructions etc.
397    for (df_iterator<BasicBlock*> DI = df_begin(&F.getEntryBlock()),
398         DE = df_end(&F.getEntryBlock()); DI != DE; ++DI) {
399      BasicBlock *BB = *DI;
400      visit(*BB);
401    }
402    if (AllocaVec.empty()) return false;
403
404    initializeCallbacks(*F.getParent());
405
406    poisonStack();
407
408    if (ClDebugStack) {
409      DEBUG(dbgs() << F);
410    }
411    return true;
412  }
413
414  // Finds all static Alloca instructions and puts
415  // poisoned red zones around all of them.
416  // Then unpoison everything back before the function returns.
417  void poisonStack();
418
419  // ----------------------- Visitors.
420  /// \brief Collect all Ret instructions.
421  void visitReturnInst(ReturnInst &RI) {
422    RetVec.push_back(&RI);
423  }
424
425  /// \brief Collect Alloca instructions we want (and can) handle.
426  void visitAllocaInst(AllocaInst &AI) {
427    if (!isInterestingAlloca(AI)) return;
428
429    StackAlignment = std::max(StackAlignment, AI.getAlignment());
430    AllocaVec.push_back(&AI);
431    uint64_t AlignedSize =  getAlignedAllocaSize(&AI);
432    TotalStackSize += AlignedSize;
433  }
434
435  /// \brief Collect lifetime intrinsic calls to check for use-after-scope
436  /// errors.
437  void visitIntrinsicInst(IntrinsicInst &II) {
438    if (!ASan.CheckLifetime) return;
439    Intrinsic::ID ID = II.getIntrinsicID();
440    if (ID != Intrinsic::lifetime_start &&
441        ID != Intrinsic::lifetime_end)
442      return;
443    // Found lifetime intrinsic, add ASan instrumentation if necessary.
444    ConstantInt *Size = dyn_cast<ConstantInt>(II.getArgOperand(0));
445    // If size argument is undefined, don't do anything.
446    if (Size->isMinusOne()) return;
447    // Check that size doesn't saturate uint64_t and can
448    // be stored in IntptrTy.
449    const uint64_t SizeValue = Size->getValue().getLimitedValue();
450    if (SizeValue == ~0ULL ||
451        !ConstantInt::isValueValidForType(IntptrTy, SizeValue))
452      return;
453    // Find alloca instruction that corresponds to llvm.lifetime argument.
454    AllocaInst *AI = findAllocaForValue(II.getArgOperand(1));
455    if (!AI) return;
456    bool DoPoison = (ID == Intrinsic::lifetime_end);
457    AllocaPoisonCall APC = {&II, SizeValue, DoPoison};
458    AllocaPoisonCallVec.push_back(APC);
459  }
460
461  // ---------------------- Helpers.
462  void initializeCallbacks(Module &M);
463
464  // Check if we want (and can) handle this alloca.
465  bool isInterestingAlloca(AllocaInst &AI) {
466    return (!AI.isArrayAllocation() &&
467            AI.isStaticAlloca() &&
468            AI.getAllocatedType()->isSized());
469  }
470
471  size_t RedzoneSize() const {
472    return RedzoneSizeForScale(Mapping.Scale);
473  }
474  uint64_t getAllocaSizeInBytes(AllocaInst *AI) {
475    Type *Ty = AI->getAllocatedType();
476    uint64_t SizeInBytes = ASan.TD->getTypeAllocSize(Ty);
477    return SizeInBytes;
478  }
479  uint64_t getAlignedSize(uint64_t SizeInBytes) {
480    size_t RZ = RedzoneSize();
481    return ((SizeInBytes + RZ - 1) / RZ) * RZ;
482  }
483  uint64_t getAlignedAllocaSize(AllocaInst *AI) {
484    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
485    return getAlignedSize(SizeInBytes);
486  }
487  /// Finds alloca where the value comes from.
488  AllocaInst *findAllocaForValue(Value *V);
489  void poisonRedZones(const ArrayRef<AllocaInst*> &AllocaVec, IRBuilder<> IRB,
490                      Value *ShadowBase, bool DoPoison);
491  void poisonAlloca(Value *V, uint64_t Size, IRBuilder<> IRB, bool DoPoison);
492};
493
494}  // namespace
495
496char AddressSanitizer::ID = 0;
497INITIALIZE_PASS(AddressSanitizer, "asan",
498    "AddressSanitizer: detects use-after-free and out-of-bounds bugs.",
499    false, false)
500FunctionPass *llvm::createAddressSanitizerFunctionPass(
501    bool CheckInitOrder, bool CheckUseAfterReturn, bool CheckLifetime,
502    StringRef BlacklistFile, bool ZeroBaseShadow) {
503  return new AddressSanitizer(CheckInitOrder, CheckUseAfterReturn,
504                              CheckLifetime, BlacklistFile, ZeroBaseShadow);
505}
506
507char AddressSanitizerModule::ID = 0;
508INITIALIZE_PASS(AddressSanitizerModule, "asan-module",
509    "AddressSanitizer: detects use-after-free and out-of-bounds bugs."
510    "ModulePass", false, false)
511ModulePass *llvm::createAddressSanitizerModulePass(
512    bool CheckInitOrder, StringRef BlacklistFile, bool ZeroBaseShadow) {
513  return new AddressSanitizerModule(CheckInitOrder, BlacklistFile,
514                                    ZeroBaseShadow);
515}
516
517static size_t TypeSizeToSizeIndex(uint32_t TypeSize) {
518  size_t Res = CountTrailingZeros_32(TypeSize / 8);
519  assert(Res < kNumberOfAccessSizes);
520  return Res;
521}
522
523// Create a constant for Str so that we can pass it to the run-time lib.
524static GlobalVariable *createPrivateGlobalForString(Module &M, StringRef Str) {
525  Constant *StrConst = ConstantDataArray::getString(M.getContext(), Str);
526  return new GlobalVariable(M, StrConst->getType(), true,
527                            GlobalValue::PrivateLinkage, StrConst,
528                            kAsanGenPrefix);
529}
530
531static bool GlobalWasGeneratedByAsan(GlobalVariable *G) {
532  return G->getName().find(kAsanGenPrefix) == 0;
533}
534
535Value *AddressSanitizer::memToShadow(Value *Shadow, IRBuilder<> &IRB) {
536  // Shadow >> scale
537  Shadow = IRB.CreateLShr(Shadow, Mapping.Scale);
538  if (Mapping.Offset == 0)
539    return Shadow;
540  // (Shadow >> scale) | offset
541  if (Mapping.OrShadowOffset)
542    return IRB.CreateOr(Shadow, ConstantInt::get(IntptrTy, Mapping.Offset));
543  else
544    return IRB.CreateAdd(Shadow, ConstantInt::get(IntptrTy, Mapping.Offset));
545}
546
547void AddressSanitizer::instrumentMemIntrinsicParam(
548    Instruction *OrigIns,
549    Value *Addr, Value *Size, Instruction *InsertBefore, bool IsWrite) {
550  // Check the first byte.
551  {
552    IRBuilder<> IRB(InsertBefore);
553    instrumentAddress(OrigIns, IRB, Addr, 8, IsWrite);
554  }
555  // Check the last byte.
556  {
557    IRBuilder<> IRB(InsertBefore);
558    Value *SizeMinusOne = IRB.CreateSub(
559        Size, ConstantInt::get(Size->getType(), 1));
560    SizeMinusOne = IRB.CreateIntCast(SizeMinusOne, IntptrTy, false);
561    Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
562    Value *AddrPlusSizeMinisOne = IRB.CreateAdd(AddrLong, SizeMinusOne);
563    instrumentAddress(OrigIns, IRB, AddrPlusSizeMinisOne, 8, IsWrite);
564  }
565}
566
567// Instrument memset/memmove/memcpy
568bool AddressSanitizer::instrumentMemIntrinsic(MemIntrinsic *MI) {
569  Value *Dst = MI->getDest();
570  MemTransferInst *MemTran = dyn_cast<MemTransferInst>(MI);
571  Value *Src = MemTran ? MemTran->getSource() : 0;
572  Value *Length = MI->getLength();
573
574  Constant *ConstLength = dyn_cast<Constant>(Length);
575  Instruction *InsertBefore = MI;
576  if (ConstLength) {
577    if (ConstLength->isNullValue()) return false;
578  } else {
579    // The size is not a constant so it could be zero -- check at run-time.
580    IRBuilder<> IRB(InsertBefore);
581
582    Value *Cmp = IRB.CreateICmpNE(Length,
583                                  Constant::getNullValue(Length->getType()));
584    InsertBefore = SplitBlockAndInsertIfThen(cast<Instruction>(Cmp), false);
585  }
586
587  instrumentMemIntrinsicParam(MI, Dst, Length, InsertBefore, true);
588  if (Src)
589    instrumentMemIntrinsicParam(MI, Src, Length, InsertBefore, false);
590  return true;
591}
592
593// If I is an interesting memory access, return the PointerOperand
594// and set IsWrite. Otherwise return NULL.
595static Value *isInterestingMemoryAccess(Instruction *I, bool *IsWrite) {
596  if (LoadInst *LI = dyn_cast<LoadInst>(I)) {
597    if (!ClInstrumentReads) return NULL;
598    *IsWrite = false;
599    return LI->getPointerOperand();
600  }
601  if (StoreInst *SI = dyn_cast<StoreInst>(I)) {
602    if (!ClInstrumentWrites) return NULL;
603    *IsWrite = true;
604    return SI->getPointerOperand();
605  }
606  if (AtomicRMWInst *RMW = dyn_cast<AtomicRMWInst>(I)) {
607    if (!ClInstrumentAtomics) return NULL;
608    *IsWrite = true;
609    return RMW->getPointerOperand();
610  }
611  if (AtomicCmpXchgInst *XCHG = dyn_cast<AtomicCmpXchgInst>(I)) {
612    if (!ClInstrumentAtomics) return NULL;
613    *IsWrite = true;
614    return XCHG->getPointerOperand();
615  }
616  return NULL;
617}
618
619void AddressSanitizer::instrumentMop(Instruction *I) {
620  bool IsWrite = false;
621  Value *Addr = isInterestingMemoryAccess(I, &IsWrite);
622  assert(Addr);
623  if (ClOpt && ClOptGlobals) {
624    if (GlobalVariable *G = dyn_cast<GlobalVariable>(Addr)) {
625      // If initialization order checking is disabled, a simple access to a
626      // dynamically initialized global is always valid.
627      if (!CheckInitOrder)
628        return;
629      // If a global variable does not have dynamic initialization we don't
630      // have to instrument it.  However, if a global does not have initailizer
631      // at all, we assume it has dynamic initializer (in other TU).
632      if (G->hasInitializer() && !DynamicallyInitializedGlobals.Contains(G))
633        return;
634    }
635  }
636
637  Type *OrigPtrTy = Addr->getType();
638  Type *OrigTy = cast<PointerType>(OrigPtrTy)->getElementType();
639
640  assert(OrigTy->isSized());
641  uint32_t TypeSize = TD->getTypeStoreSizeInBits(OrigTy);
642
643  if (TypeSize != 8  && TypeSize != 16 &&
644      TypeSize != 32 && TypeSize != 64 && TypeSize != 128) {
645    // Ignore all unusual sizes.
646    return;
647  }
648
649  IRBuilder<> IRB(I);
650  instrumentAddress(I, IRB, Addr, TypeSize, IsWrite);
651}
652
653// Validate the result of Module::getOrInsertFunction called for an interface
654// function of AddressSanitizer. If the instrumented module defines a function
655// with the same name, their prototypes must match, otherwise
656// getOrInsertFunction returns a bitcast.
657static Function *checkInterfaceFunction(Constant *FuncOrBitcast) {
658  if (isa<Function>(FuncOrBitcast)) return cast<Function>(FuncOrBitcast);
659  FuncOrBitcast->dump();
660  report_fatal_error("trying to redefine an AddressSanitizer "
661                     "interface function");
662}
663
664Instruction *AddressSanitizer::generateCrashCode(
665    Instruction *InsertBefore, Value *Addr,
666    bool IsWrite, size_t AccessSizeIndex) {
667  IRBuilder<> IRB(InsertBefore);
668  CallInst *Call = IRB.CreateCall(AsanErrorCallback[IsWrite][AccessSizeIndex],
669                                  Addr);
670  // We don't do Call->setDoesNotReturn() because the BB already has
671  // UnreachableInst at the end.
672  // This EmptyAsm is required to avoid callback merge.
673  IRB.CreateCall(EmptyAsm);
674  return Call;
675}
676
677Value *AddressSanitizer::createSlowPathCmp(IRBuilder<> &IRB, Value *AddrLong,
678                                            Value *ShadowValue,
679                                            uint32_t TypeSize) {
680  size_t Granularity = 1 << Mapping.Scale;
681  // Addr & (Granularity - 1)
682  Value *LastAccessedByte = IRB.CreateAnd(
683      AddrLong, ConstantInt::get(IntptrTy, Granularity - 1));
684  // (Addr & (Granularity - 1)) + size - 1
685  if (TypeSize / 8 > 1)
686    LastAccessedByte = IRB.CreateAdd(
687        LastAccessedByte, ConstantInt::get(IntptrTy, TypeSize / 8 - 1));
688  // (uint8_t) ((Addr & (Granularity-1)) + size - 1)
689  LastAccessedByte = IRB.CreateIntCast(
690      LastAccessedByte, ShadowValue->getType(), false);
691  // ((uint8_t) ((Addr & (Granularity-1)) + size - 1)) >= ShadowValue
692  return IRB.CreateICmpSGE(LastAccessedByte, ShadowValue);
693}
694
695void AddressSanitizer::instrumentAddress(Instruction *OrigIns,
696                                         IRBuilder<> &IRB, Value *Addr,
697                                         uint32_t TypeSize, bool IsWrite) {
698  Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
699
700  Type *ShadowTy  = IntegerType::get(
701      *C, std::max(8U, TypeSize >> Mapping.Scale));
702  Type *ShadowPtrTy = PointerType::get(ShadowTy, 0);
703  Value *ShadowPtr = memToShadow(AddrLong, IRB);
704  Value *CmpVal = Constant::getNullValue(ShadowTy);
705  Value *ShadowValue = IRB.CreateLoad(
706      IRB.CreateIntToPtr(ShadowPtr, ShadowPtrTy));
707
708  Value *Cmp = IRB.CreateICmpNE(ShadowValue, CmpVal);
709  size_t AccessSizeIndex = TypeSizeToSizeIndex(TypeSize);
710  size_t Granularity = 1 << Mapping.Scale;
711  TerminatorInst *CrashTerm = 0;
712
713  if (ClAlwaysSlowPath || (TypeSize < 8 * Granularity)) {
714    TerminatorInst *CheckTerm =
715        SplitBlockAndInsertIfThen(cast<Instruction>(Cmp), false);
716    assert(dyn_cast<BranchInst>(CheckTerm)->isUnconditional());
717    BasicBlock *NextBB = CheckTerm->getSuccessor(0);
718    IRB.SetInsertPoint(CheckTerm);
719    Value *Cmp2 = createSlowPathCmp(IRB, AddrLong, ShadowValue, TypeSize);
720    BasicBlock *CrashBlock =
721        BasicBlock::Create(*C, "", NextBB->getParent(), NextBB);
722    CrashTerm = new UnreachableInst(*C, CrashBlock);
723    BranchInst *NewTerm = BranchInst::Create(CrashBlock, NextBB, Cmp2);
724    ReplaceInstWithInst(CheckTerm, NewTerm);
725  } else {
726    CrashTerm = SplitBlockAndInsertIfThen(cast<Instruction>(Cmp), true);
727  }
728
729  Instruction *Crash =
730      generateCrashCode(CrashTerm, AddrLong, IsWrite, AccessSizeIndex);
731  Crash->setDebugLoc(OrigIns->getDebugLoc());
732}
733
734void AddressSanitizerModule::createInitializerPoisonCalls(
735    Module &M, Value *FirstAddr, Value *LastAddr) {
736  // We do all of our poisoning and unpoisoning within _GLOBAL__I_a.
737  Function *GlobalInit = M.getFunction("_GLOBAL__I_a");
738  // If that function is not present, this TU contains no globals, or they have
739  // all been optimized away
740  if (!GlobalInit)
741    return;
742
743  // Set up the arguments to our poison/unpoison functions.
744  IRBuilder<> IRB(GlobalInit->begin()->getFirstInsertionPt());
745
746  // Add a call to poison all external globals before the given function starts.
747  IRB.CreateCall2(AsanPoisonGlobals, FirstAddr, LastAddr);
748
749  // Add calls to unpoison all globals before each return instruction.
750  for (Function::iterator I = GlobalInit->begin(), E = GlobalInit->end();
751      I != E; ++I) {
752    if (ReturnInst *RI = dyn_cast<ReturnInst>(I->getTerminator())) {
753      CallInst::Create(AsanUnpoisonGlobals, "", RI);
754    }
755  }
756}
757
758bool AddressSanitizerModule::ShouldInstrumentGlobal(GlobalVariable *G) {
759  Type *Ty = cast<PointerType>(G->getType())->getElementType();
760  DEBUG(dbgs() << "GLOBAL: " << *G << "\n");
761
762  if (BL->isIn(*G)) return false;
763  if (!Ty->isSized()) return false;
764  if (!G->hasInitializer()) return false;
765  if (GlobalWasGeneratedByAsan(G)) return false;  // Our own global.
766  // Touch only those globals that will not be defined in other modules.
767  // Don't handle ODR type linkages since other modules may be built w/o asan.
768  if (G->getLinkage() != GlobalVariable::ExternalLinkage &&
769      G->getLinkage() != GlobalVariable::PrivateLinkage &&
770      G->getLinkage() != GlobalVariable::InternalLinkage)
771    return false;
772  // Two problems with thread-locals:
773  //   - The address of the main thread's copy can't be computed at link-time.
774  //   - Need to poison all copies, not just the main thread's one.
775  if (G->isThreadLocal())
776    return false;
777  // For now, just ignore this Alloca if the alignment is large.
778  if (G->getAlignment() > RedzoneSize()) return false;
779
780  // Ignore all the globals with the names starting with "\01L_OBJC_".
781  // Many of those are put into the .cstring section. The linker compresses
782  // that section by removing the spare \0s after the string terminator, so
783  // our redzones get broken.
784  if ((G->getName().find("\01L_OBJC_") == 0) ||
785      (G->getName().find("\01l_OBJC_") == 0)) {
786    DEBUG(dbgs() << "Ignoring \\01L_OBJC_* global: " << *G);
787    return false;
788  }
789
790  if (G->hasSection()) {
791    StringRef Section(G->getSection());
792    // Ignore the globals from the __OBJC section. The ObjC runtime assumes
793    // those conform to /usr/lib/objc/runtime.h, so we can't add redzones to
794    // them.
795    if ((Section.find("__OBJC,") == 0) ||
796        (Section.find("__DATA, __objc_") == 0)) {
797      DEBUG(dbgs() << "Ignoring ObjC runtime global: " << *G);
798      return false;
799    }
800    // See http://code.google.com/p/address-sanitizer/issues/detail?id=32
801    // Constant CFString instances are compiled in the following way:
802    //  -- the string buffer is emitted into
803    //     __TEXT,__cstring,cstring_literals
804    //  -- the constant NSConstantString structure referencing that buffer
805    //     is placed into __DATA,__cfstring
806    // Therefore there's no point in placing redzones into __DATA,__cfstring.
807    // Moreover, it causes the linker to crash on OS X 10.7
808    if (Section.find("__DATA,__cfstring") == 0) {
809      DEBUG(dbgs() << "Ignoring CFString: " << *G);
810      return false;
811    }
812  }
813
814  return true;
815}
816
817void AddressSanitizerModule::initializeCallbacks(Module &M) {
818  IRBuilder<> IRB(*C);
819  // Declare our poisoning and unpoisoning functions.
820  AsanPoisonGlobals = checkInterfaceFunction(M.getOrInsertFunction(
821      kAsanPoisonGlobalsName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
822  AsanPoisonGlobals->setLinkage(Function::ExternalLinkage);
823  AsanUnpoisonGlobals = checkInterfaceFunction(M.getOrInsertFunction(
824      kAsanUnpoisonGlobalsName, IRB.getVoidTy(), NULL));
825  AsanUnpoisonGlobals->setLinkage(Function::ExternalLinkage);
826  // Declare functions that register/unregister globals.
827  AsanRegisterGlobals = checkInterfaceFunction(M.getOrInsertFunction(
828      kAsanRegisterGlobalsName, IRB.getVoidTy(),
829      IntptrTy, IntptrTy, NULL));
830  AsanRegisterGlobals->setLinkage(Function::ExternalLinkage);
831  AsanUnregisterGlobals = checkInterfaceFunction(M.getOrInsertFunction(
832      kAsanUnregisterGlobalsName,
833      IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
834  AsanUnregisterGlobals->setLinkage(Function::ExternalLinkage);
835}
836
837// This function replaces all global variables with new variables that have
838// trailing redzones. It also creates a function that poisons
839// redzones and inserts this function into llvm.global_ctors.
840bool AddressSanitizerModule::runOnModule(Module &M) {
841  if (!ClGlobals) return false;
842  TD = getAnalysisIfAvailable<DataLayout>();
843  if (!TD)
844    return false;
845  BL.reset(new BlackList(BlacklistFile));
846  if (BL->isIn(M)) return false;
847  C = &(M.getContext());
848  int LongSize = TD->getPointerSizeInBits();
849  IntptrTy = Type::getIntNTy(*C, LongSize);
850  Mapping = getShadowMapping(M, LongSize, ZeroBaseShadow);
851  initializeCallbacks(M);
852  DynamicallyInitializedGlobals.Init(M);
853
854  SmallVector<GlobalVariable *, 16> GlobalsToChange;
855
856  for (Module::GlobalListType::iterator G = M.global_begin(),
857       E = M.global_end(); G != E; ++G) {
858    if (ShouldInstrumentGlobal(G))
859      GlobalsToChange.push_back(G);
860  }
861
862  size_t n = GlobalsToChange.size();
863  if (n == 0) return false;
864
865  // A global is described by a structure
866  //   size_t beg;
867  //   size_t size;
868  //   size_t size_with_redzone;
869  //   const char *name;
870  //   size_t has_dynamic_init;
871  // We initialize an array of such structures and pass it to a run-time call.
872  StructType *GlobalStructTy = StructType::get(IntptrTy, IntptrTy,
873                                               IntptrTy, IntptrTy,
874                                               IntptrTy, NULL);
875  SmallVector<Constant *, 16> Initializers(n), DynamicInit;
876
877
878  Function *CtorFunc = M.getFunction(kAsanModuleCtorName);
879  assert(CtorFunc);
880  IRBuilder<> IRB(CtorFunc->getEntryBlock().getTerminator());
881
882  // The addresses of the first and last dynamically initialized globals in
883  // this TU.  Used in initialization order checking.
884  Value *FirstDynamic = 0, *LastDynamic = 0;
885
886  for (size_t i = 0; i < n; i++) {
887    static const uint64_t kMaxGlobalRedzone = 1 << 18;
888    GlobalVariable *G = GlobalsToChange[i];
889    PointerType *PtrTy = cast<PointerType>(G->getType());
890    Type *Ty = PtrTy->getElementType();
891    uint64_t SizeInBytes = TD->getTypeAllocSize(Ty);
892    uint64_t MinRZ = RedzoneSize();
893    // MinRZ <= RZ <= kMaxGlobalRedzone
894    // and trying to make RZ to be ~ 1/4 of SizeInBytes.
895    uint64_t RZ = std::max(MinRZ,
896                         std::min(kMaxGlobalRedzone,
897                                  (SizeInBytes / MinRZ / 4) * MinRZ));
898    uint64_t RightRedzoneSize = RZ;
899    // Round up to MinRZ
900    if (SizeInBytes % MinRZ)
901      RightRedzoneSize += MinRZ - (SizeInBytes % MinRZ);
902    assert(((RightRedzoneSize + SizeInBytes) % MinRZ) == 0);
903    Type *RightRedZoneTy = ArrayType::get(IRB.getInt8Ty(), RightRedzoneSize);
904    // Determine whether this global should be poisoned in initialization.
905    bool GlobalHasDynamicInitializer =
906        DynamicallyInitializedGlobals.Contains(G);
907    // Don't check initialization order if this global is blacklisted.
908    GlobalHasDynamicInitializer &= !BL->isInInit(*G);
909
910    StructType *NewTy = StructType::get(Ty, RightRedZoneTy, NULL);
911    Constant *NewInitializer = ConstantStruct::get(
912        NewTy, G->getInitializer(),
913        Constant::getNullValue(RightRedZoneTy), NULL);
914
915    SmallString<2048> DescriptionOfGlobal = G->getName();
916    DescriptionOfGlobal += " (";
917    DescriptionOfGlobal += M.getModuleIdentifier();
918    DescriptionOfGlobal += ")";
919    GlobalVariable *Name = createPrivateGlobalForString(M, DescriptionOfGlobal);
920
921    // Create a new global variable with enough space for a redzone.
922    GlobalVariable *NewGlobal = new GlobalVariable(
923        M, NewTy, G->isConstant(), G->getLinkage(),
924        NewInitializer, "", G, G->getThreadLocalMode());
925    NewGlobal->copyAttributesFrom(G);
926    NewGlobal->setAlignment(MinRZ);
927
928    Value *Indices2[2];
929    Indices2[0] = IRB.getInt32(0);
930    Indices2[1] = IRB.getInt32(0);
931
932    G->replaceAllUsesWith(
933        ConstantExpr::getGetElementPtr(NewGlobal, Indices2, true));
934    NewGlobal->takeName(G);
935    G->eraseFromParent();
936
937    Initializers[i] = ConstantStruct::get(
938        GlobalStructTy,
939        ConstantExpr::getPointerCast(NewGlobal, IntptrTy),
940        ConstantInt::get(IntptrTy, SizeInBytes),
941        ConstantInt::get(IntptrTy, SizeInBytes + RightRedzoneSize),
942        ConstantExpr::getPointerCast(Name, IntptrTy),
943        ConstantInt::get(IntptrTy, GlobalHasDynamicInitializer),
944        NULL);
945
946    // Populate the first and last globals declared in this TU.
947    if (CheckInitOrder && GlobalHasDynamicInitializer) {
948      LastDynamic = ConstantExpr::getPointerCast(NewGlobal, IntptrTy);
949      if (FirstDynamic == 0)
950        FirstDynamic = LastDynamic;
951    }
952
953    DEBUG(dbgs() << "NEW GLOBAL: " << *NewGlobal << "\n");
954  }
955
956  ArrayType *ArrayOfGlobalStructTy = ArrayType::get(GlobalStructTy, n);
957  GlobalVariable *AllGlobals = new GlobalVariable(
958      M, ArrayOfGlobalStructTy, false, GlobalVariable::PrivateLinkage,
959      ConstantArray::get(ArrayOfGlobalStructTy, Initializers), "");
960
961  // Create calls for poisoning before initializers run and unpoisoning after.
962  if (CheckInitOrder && FirstDynamic && LastDynamic)
963    createInitializerPoisonCalls(M, FirstDynamic, LastDynamic);
964  IRB.CreateCall2(AsanRegisterGlobals,
965                  IRB.CreatePointerCast(AllGlobals, IntptrTy),
966                  ConstantInt::get(IntptrTy, n));
967
968  // We also need to unregister globals at the end, e.g. when a shared library
969  // gets closed.
970  Function *AsanDtorFunction = Function::Create(
971      FunctionType::get(Type::getVoidTy(*C), false),
972      GlobalValue::InternalLinkage, kAsanModuleDtorName, &M);
973  BasicBlock *AsanDtorBB = BasicBlock::Create(*C, "", AsanDtorFunction);
974  IRBuilder<> IRB_Dtor(ReturnInst::Create(*C, AsanDtorBB));
975  IRB_Dtor.CreateCall2(AsanUnregisterGlobals,
976                       IRB.CreatePointerCast(AllGlobals, IntptrTy),
977                       ConstantInt::get(IntptrTy, n));
978  appendToGlobalDtors(M, AsanDtorFunction, kAsanCtorAndCtorPriority);
979
980  DEBUG(dbgs() << M);
981  return true;
982}
983
984void AddressSanitizer::initializeCallbacks(Module &M) {
985  IRBuilder<> IRB(*C);
986  // Create __asan_report* callbacks.
987  for (size_t AccessIsWrite = 0; AccessIsWrite <= 1; AccessIsWrite++) {
988    for (size_t AccessSizeIndex = 0; AccessSizeIndex < kNumberOfAccessSizes;
989         AccessSizeIndex++) {
990      // IsWrite and TypeSize are encoded in the function name.
991      std::string FunctionName = std::string(kAsanReportErrorTemplate) +
992          (AccessIsWrite ? "store" : "load") + itostr(1 << AccessSizeIndex);
993      // If we are merging crash callbacks, they have two parameters.
994      AsanErrorCallback[AccessIsWrite][AccessSizeIndex] =
995          checkInterfaceFunction(M.getOrInsertFunction(
996              FunctionName, IRB.getVoidTy(), IntptrTy, NULL));
997    }
998  }
999
1000  AsanHandleNoReturnFunc = checkInterfaceFunction(M.getOrInsertFunction(
1001      kAsanHandleNoReturnName, IRB.getVoidTy(), NULL));
1002  // We insert an empty inline asm after __asan_report* to avoid callback merge.
1003  EmptyAsm = InlineAsm::get(FunctionType::get(IRB.getVoidTy(), false),
1004                            StringRef(""), StringRef(""),
1005                            /*hasSideEffects=*/true);
1006}
1007
1008void AddressSanitizer::emitShadowMapping(Module &M, IRBuilder<> &IRB) const {
1009  // Tell the values of mapping offset and scale to the run-time.
1010  GlobalValue *asan_mapping_offset =
1011      new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
1012                     ConstantInt::get(IntptrTy, Mapping.Offset),
1013                     kAsanMappingOffsetName);
1014  // Read the global, otherwise it may be optimized away.
1015  IRB.CreateLoad(asan_mapping_offset, true);
1016
1017  GlobalValue *asan_mapping_scale =
1018      new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
1019                         ConstantInt::get(IntptrTy, Mapping.Scale),
1020                         kAsanMappingScaleName);
1021  // Read the global, otherwise it may be optimized away.
1022  IRB.CreateLoad(asan_mapping_scale, true);
1023}
1024
1025// virtual
1026bool AddressSanitizer::doInitialization(Module &M) {
1027  // Initialize the private fields. No one has accessed them before.
1028  TD = getAnalysisIfAvailable<DataLayout>();
1029
1030  if (!TD)
1031    return false;
1032  BL.reset(new BlackList(BlacklistFile));
1033  DynamicallyInitializedGlobals.Init(M);
1034
1035  C = &(M.getContext());
1036  LongSize = TD->getPointerSizeInBits();
1037  IntptrTy = Type::getIntNTy(*C, LongSize);
1038
1039  AsanCtorFunction = Function::Create(
1040      FunctionType::get(Type::getVoidTy(*C), false),
1041      GlobalValue::InternalLinkage, kAsanModuleCtorName, &M);
1042  BasicBlock *AsanCtorBB = BasicBlock::Create(*C, "", AsanCtorFunction);
1043  // call __asan_init in the module ctor.
1044  IRBuilder<> IRB(ReturnInst::Create(*C, AsanCtorBB));
1045  AsanInitFunction = checkInterfaceFunction(
1046      M.getOrInsertFunction(kAsanInitName, IRB.getVoidTy(), NULL));
1047  AsanInitFunction->setLinkage(Function::ExternalLinkage);
1048  IRB.CreateCall(AsanInitFunction);
1049
1050  Mapping = getShadowMapping(M, LongSize, ZeroBaseShadow);
1051  emitShadowMapping(M, IRB);
1052
1053  appendToGlobalCtors(M, AsanCtorFunction, kAsanCtorAndCtorPriority);
1054  return true;
1055}
1056
1057bool AddressSanitizer::maybeInsertAsanInitAtFunctionEntry(Function &F) {
1058  // For each NSObject descendant having a +load method, this method is invoked
1059  // by the ObjC runtime before any of the static constructors is called.
1060  // Therefore we need to instrument such methods with a call to __asan_init
1061  // at the beginning in order to initialize our runtime before any access to
1062  // the shadow memory.
1063  // We cannot just ignore these methods, because they may call other
1064  // instrumented functions.
1065  if (F.getName().find(" load]") != std::string::npos) {
1066    IRBuilder<> IRB(F.begin()->begin());
1067    IRB.CreateCall(AsanInitFunction);
1068    return true;
1069  }
1070  return false;
1071}
1072
1073bool AddressSanitizer::runOnFunction(Function &F) {
1074  if (BL->isIn(F)) return false;
1075  if (&F == AsanCtorFunction) return false;
1076  DEBUG(dbgs() << "ASAN instrumenting:\n" << F << "\n");
1077  initializeCallbacks(*F.getParent());
1078
1079  // If needed, insert __asan_init before checking for AddressSafety attr.
1080  maybeInsertAsanInitAtFunctionEntry(F);
1081
1082  if (!F.getAttributes().hasAttribute(AttributeSet::FunctionIndex,
1083                                      Attribute::AddressSafety))
1084    return false;
1085
1086  if (!ClDebugFunc.empty() && ClDebugFunc != F.getName())
1087    return false;
1088
1089  // We want to instrument every address only once per basic block (unless there
1090  // are calls between uses).
1091  SmallSet<Value*, 16> TempsToInstrument;
1092  SmallVector<Instruction*, 16> ToInstrument;
1093  SmallVector<Instruction*, 8> NoReturnCalls;
1094  bool IsWrite;
1095
1096  // Fill the set of memory operations to instrument.
1097  for (Function::iterator FI = F.begin(), FE = F.end();
1098       FI != FE; ++FI) {
1099    TempsToInstrument.clear();
1100    int NumInsnsPerBB = 0;
1101    for (BasicBlock::iterator BI = FI->begin(), BE = FI->end();
1102         BI != BE; ++BI) {
1103      if (LooksLikeCodeInBug11395(BI)) return false;
1104      if (Value *Addr = isInterestingMemoryAccess(BI, &IsWrite)) {
1105        if (ClOpt && ClOptSameTemp) {
1106          if (!TempsToInstrument.insert(Addr))
1107            continue;  // We've seen this temp in the current BB.
1108        }
1109      } else if (isa<MemIntrinsic>(BI) && ClMemIntrin) {
1110        // ok, take it.
1111      } else {
1112        if (CallInst *CI = dyn_cast<CallInst>(BI)) {
1113          // A call inside BB.
1114          TempsToInstrument.clear();
1115          if (CI->doesNotReturn()) {
1116            NoReturnCalls.push_back(CI);
1117          }
1118        }
1119        continue;
1120      }
1121      ToInstrument.push_back(BI);
1122      NumInsnsPerBB++;
1123      if (NumInsnsPerBB >= ClMaxInsnsToInstrumentPerBB)
1124        break;
1125    }
1126  }
1127
1128  // Instrument.
1129  int NumInstrumented = 0;
1130  for (size_t i = 0, n = ToInstrument.size(); i != n; i++) {
1131    Instruction *Inst = ToInstrument[i];
1132    if (ClDebugMin < 0 || ClDebugMax < 0 ||
1133        (NumInstrumented >= ClDebugMin && NumInstrumented <= ClDebugMax)) {
1134      if (isInterestingMemoryAccess(Inst, &IsWrite))
1135        instrumentMop(Inst);
1136      else
1137        instrumentMemIntrinsic(cast<MemIntrinsic>(Inst));
1138    }
1139    NumInstrumented++;
1140  }
1141
1142  FunctionStackPoisoner FSP(F, *this);
1143  bool ChangedStack = FSP.runOnFunction();
1144
1145  // We must unpoison the stack before every NoReturn call (throw, _exit, etc).
1146  // See e.g. http://code.google.com/p/address-sanitizer/issues/detail?id=37
1147  for (size_t i = 0, n = NoReturnCalls.size(); i != n; i++) {
1148    Instruction *CI = NoReturnCalls[i];
1149    IRBuilder<> IRB(CI);
1150    IRB.CreateCall(AsanHandleNoReturnFunc);
1151  }
1152  DEBUG(dbgs() << "ASAN done instrumenting:\n" << F << "\n");
1153
1154  return NumInstrumented > 0 || ChangedStack || !NoReturnCalls.empty();
1155}
1156
1157static uint64_t ValueForPoison(uint64_t PoisonByte, size_t ShadowRedzoneSize) {
1158  if (ShadowRedzoneSize == 1) return PoisonByte;
1159  if (ShadowRedzoneSize == 2) return (PoisonByte << 8) + PoisonByte;
1160  if (ShadowRedzoneSize == 4)
1161    return (PoisonByte << 24) + (PoisonByte << 16) +
1162        (PoisonByte << 8) + (PoisonByte);
1163  llvm_unreachable("ShadowRedzoneSize is either 1, 2 or 4");
1164}
1165
1166static void PoisonShadowPartialRightRedzone(uint8_t *Shadow,
1167                                            size_t Size,
1168                                            size_t RZSize,
1169                                            size_t ShadowGranularity,
1170                                            uint8_t Magic) {
1171  for (size_t i = 0; i < RZSize;
1172       i+= ShadowGranularity, Shadow++) {
1173    if (i + ShadowGranularity <= Size) {
1174      *Shadow = 0;  // fully addressable
1175    } else if (i >= Size) {
1176      *Shadow = Magic;  // unaddressable
1177    } else {
1178      *Shadow = Size - i;  // first Size-i bytes are addressable
1179    }
1180  }
1181}
1182
1183// Workaround for bug 11395: we don't want to instrument stack in functions
1184// with large assembly blobs (32-bit only), otherwise reg alloc may crash.
1185// FIXME: remove once the bug 11395 is fixed.
1186bool AddressSanitizer::LooksLikeCodeInBug11395(Instruction *I) {
1187  if (LongSize != 32) return false;
1188  CallInst *CI = dyn_cast<CallInst>(I);
1189  if (!CI || !CI->isInlineAsm()) return false;
1190  if (CI->getNumArgOperands() <= 5) return false;
1191  // We have inline assembly with quite a few arguments.
1192  return true;
1193}
1194
1195void FunctionStackPoisoner::initializeCallbacks(Module &M) {
1196  IRBuilder<> IRB(*C);
1197  AsanStackMallocFunc = checkInterfaceFunction(M.getOrInsertFunction(
1198      kAsanStackMallocName, IntptrTy, IntptrTy, IntptrTy, NULL));
1199  AsanStackFreeFunc = checkInterfaceFunction(M.getOrInsertFunction(
1200      kAsanStackFreeName, IRB.getVoidTy(),
1201      IntptrTy, IntptrTy, IntptrTy, NULL));
1202  AsanPoisonStackMemoryFunc = checkInterfaceFunction(M.getOrInsertFunction(
1203      kAsanPoisonStackMemoryName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
1204  AsanUnpoisonStackMemoryFunc = checkInterfaceFunction(M.getOrInsertFunction(
1205      kAsanUnpoisonStackMemoryName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
1206}
1207
1208void FunctionStackPoisoner::poisonRedZones(
1209  const ArrayRef<AllocaInst*> &AllocaVec, IRBuilder<> IRB, Value *ShadowBase,
1210  bool DoPoison) {
1211  size_t ShadowRZSize = RedzoneSize() >> Mapping.Scale;
1212  assert(ShadowRZSize >= 1 && ShadowRZSize <= 4);
1213  Type *RZTy = Type::getIntNTy(*C, ShadowRZSize * 8);
1214  Type *RZPtrTy = PointerType::get(RZTy, 0);
1215
1216  Value *PoisonLeft  = ConstantInt::get(RZTy,
1217    ValueForPoison(DoPoison ? kAsanStackLeftRedzoneMagic : 0LL, ShadowRZSize));
1218  Value *PoisonMid   = ConstantInt::get(RZTy,
1219    ValueForPoison(DoPoison ? kAsanStackMidRedzoneMagic : 0LL, ShadowRZSize));
1220  Value *PoisonRight = ConstantInt::get(RZTy,
1221    ValueForPoison(DoPoison ? kAsanStackRightRedzoneMagic : 0LL, ShadowRZSize));
1222
1223  // poison the first red zone.
1224  IRB.CreateStore(PoisonLeft, IRB.CreateIntToPtr(ShadowBase, RZPtrTy));
1225
1226  // poison all other red zones.
1227  uint64_t Pos = RedzoneSize();
1228  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
1229    AllocaInst *AI = AllocaVec[i];
1230    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
1231    uint64_t AlignedSize = getAlignedAllocaSize(AI);
1232    assert(AlignedSize - SizeInBytes < RedzoneSize());
1233    Value *Ptr = NULL;
1234
1235    Pos += AlignedSize;
1236
1237    assert(ShadowBase->getType() == IntptrTy);
1238    if (SizeInBytes < AlignedSize) {
1239      // Poison the partial redzone at right
1240      Ptr = IRB.CreateAdd(
1241          ShadowBase, ConstantInt::get(IntptrTy,
1242                                       (Pos >> Mapping.Scale) - ShadowRZSize));
1243      size_t AddressableBytes = RedzoneSize() - (AlignedSize - SizeInBytes);
1244      uint32_t Poison = 0;
1245      if (DoPoison) {
1246        PoisonShadowPartialRightRedzone((uint8_t*)&Poison, AddressableBytes,
1247                                        RedzoneSize(),
1248                                        1ULL << Mapping.Scale,
1249                                        kAsanStackPartialRedzoneMagic);
1250      }
1251      Value *PartialPoison = ConstantInt::get(RZTy, Poison);
1252      IRB.CreateStore(PartialPoison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
1253    }
1254
1255    // Poison the full redzone at right.
1256    Ptr = IRB.CreateAdd(ShadowBase,
1257                        ConstantInt::get(IntptrTy, Pos >> Mapping.Scale));
1258    bool LastAlloca = (i == AllocaVec.size() - 1);
1259    Value *Poison = LastAlloca ? PoisonRight : PoisonMid;
1260    IRB.CreateStore(Poison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
1261
1262    Pos += RedzoneSize();
1263  }
1264}
1265
1266void FunctionStackPoisoner::poisonStack() {
1267  uint64_t LocalStackSize = TotalStackSize +
1268                            (AllocaVec.size() + 1) * RedzoneSize();
1269
1270  bool DoStackMalloc = ASan.CheckUseAfterReturn
1271      && LocalStackSize <= kMaxStackMallocSize;
1272
1273  assert(AllocaVec.size() > 0);
1274  Instruction *InsBefore = AllocaVec[0];
1275  IRBuilder<> IRB(InsBefore);
1276
1277
1278  Type *ByteArrayTy = ArrayType::get(IRB.getInt8Ty(), LocalStackSize);
1279  AllocaInst *MyAlloca =
1280      new AllocaInst(ByteArrayTy, "MyAlloca", InsBefore);
1281  if (ClRealignStack && StackAlignment < RedzoneSize())
1282    StackAlignment = RedzoneSize();
1283  MyAlloca->setAlignment(StackAlignment);
1284  assert(MyAlloca->isStaticAlloca());
1285  Value *OrigStackBase = IRB.CreatePointerCast(MyAlloca, IntptrTy);
1286  Value *LocalStackBase = OrigStackBase;
1287
1288  if (DoStackMalloc) {
1289    LocalStackBase = IRB.CreateCall2(AsanStackMallocFunc,
1290        ConstantInt::get(IntptrTy, LocalStackSize), OrigStackBase);
1291  }
1292
1293  // This string will be parsed by the run-time (DescribeStackAddress).
1294  SmallString<2048> StackDescriptionStorage;
1295  raw_svector_ostream StackDescription(StackDescriptionStorage);
1296  StackDescription << F.getName() << " " << AllocaVec.size() << " ";
1297
1298  // Insert poison calls for lifetime intrinsics for alloca.
1299  bool HavePoisonedAllocas = false;
1300  for (size_t i = 0, n = AllocaPoisonCallVec.size(); i < n; i++) {
1301    const AllocaPoisonCall &APC = AllocaPoisonCallVec[i];
1302    IntrinsicInst *II = APC.InsBefore;
1303    AllocaInst *AI = findAllocaForValue(II->getArgOperand(1));
1304    assert(AI);
1305    IRBuilder<> IRB(II);
1306    poisonAlloca(AI, APC.Size, IRB, APC.DoPoison);
1307    HavePoisonedAllocas |= APC.DoPoison;
1308  }
1309
1310  uint64_t Pos = RedzoneSize();
1311  // Replace Alloca instructions with base+offset.
1312  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
1313    AllocaInst *AI = AllocaVec[i];
1314    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
1315    StringRef Name = AI->getName();
1316    StackDescription << Pos << " " << SizeInBytes << " "
1317                     << Name.size() << " " << Name << " ";
1318    uint64_t AlignedSize = getAlignedAllocaSize(AI);
1319    assert((AlignedSize % RedzoneSize()) == 0);
1320    Value *NewAllocaPtr = IRB.CreateIntToPtr(
1321            IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy, Pos)),
1322            AI->getType());
1323    replaceDbgDeclareForAlloca(AI, NewAllocaPtr, DIB);
1324    AI->replaceAllUsesWith(NewAllocaPtr);
1325    Pos += AlignedSize + RedzoneSize();
1326  }
1327  assert(Pos == LocalStackSize);
1328
1329  // Write the Magic value and the frame description constant to the redzone.
1330  Value *BasePlus0 = IRB.CreateIntToPtr(LocalStackBase, IntptrPtrTy);
1331  IRB.CreateStore(ConstantInt::get(IntptrTy, kCurrentStackFrameMagic),
1332                  BasePlus0);
1333  Value *BasePlus1 = IRB.CreateAdd(LocalStackBase,
1334                                   ConstantInt::get(IntptrTy,
1335                                                    ASan.LongSize/8));
1336  BasePlus1 = IRB.CreateIntToPtr(BasePlus1, IntptrPtrTy);
1337  GlobalVariable *StackDescriptionGlobal =
1338      createPrivateGlobalForString(*F.getParent(), StackDescription.str());
1339  Value *Description = IRB.CreatePointerCast(StackDescriptionGlobal,
1340                                             IntptrTy);
1341  IRB.CreateStore(Description, BasePlus1);
1342
1343  // Poison the stack redzones at the entry.
1344  Value *ShadowBase = ASan.memToShadow(LocalStackBase, IRB);
1345  poisonRedZones(AllocaVec, IRB, ShadowBase, true);
1346
1347  // Unpoison the stack before all ret instructions.
1348  for (size_t i = 0, n = RetVec.size(); i < n; i++) {
1349    Instruction *Ret = RetVec[i];
1350    IRBuilder<> IRBRet(Ret);
1351    // Mark the current frame as retired.
1352    IRBRet.CreateStore(ConstantInt::get(IntptrTy, kRetiredStackFrameMagic),
1353                       BasePlus0);
1354    // Unpoison the stack.
1355    poisonRedZones(AllocaVec, IRBRet, ShadowBase, false);
1356    if (DoStackMalloc) {
1357      // In use-after-return mode, mark the whole stack frame unaddressable.
1358      IRBRet.CreateCall3(AsanStackFreeFunc, LocalStackBase,
1359                         ConstantInt::get(IntptrTy, LocalStackSize),
1360                         OrigStackBase);
1361    } else if (HavePoisonedAllocas) {
1362      // If we poisoned some allocas in llvm.lifetime analysis,
1363      // unpoison whole stack frame now.
1364      assert(LocalStackBase == OrigStackBase);
1365      poisonAlloca(LocalStackBase, LocalStackSize, IRBRet, false);
1366    }
1367  }
1368
1369  // We are done. Remove the old unused alloca instructions.
1370  for (size_t i = 0, n = AllocaVec.size(); i < n; i++)
1371    AllocaVec[i]->eraseFromParent();
1372}
1373
1374void FunctionStackPoisoner::poisonAlloca(Value *V, uint64_t Size,
1375                                         IRBuilder<> IRB, bool DoPoison) {
1376  // For now just insert the call to ASan runtime.
1377  Value *AddrArg = IRB.CreatePointerCast(V, IntptrTy);
1378  Value *SizeArg = ConstantInt::get(IntptrTy, Size);
1379  IRB.CreateCall2(DoPoison ? AsanPoisonStackMemoryFunc
1380                           : AsanUnpoisonStackMemoryFunc,
1381                  AddrArg, SizeArg);
1382}
1383
1384// Handling llvm.lifetime intrinsics for a given %alloca:
1385// (1) collect all llvm.lifetime.xxx(%size, %value) describing the alloca.
1386// (2) if %size is constant, poison memory for llvm.lifetime.end (to detect
1387//     invalid accesses) and unpoison it for llvm.lifetime.start (the memory
1388//     could be poisoned by previous llvm.lifetime.end instruction, as the
1389//     variable may go in and out of scope several times, e.g. in loops).
1390// (3) if we poisoned at least one %alloca in a function,
1391//     unpoison the whole stack frame at function exit.
1392
1393AllocaInst *FunctionStackPoisoner::findAllocaForValue(Value *V) {
1394  if (AllocaInst *AI = dyn_cast<AllocaInst>(V))
1395    // We're intested only in allocas we can handle.
1396    return isInterestingAlloca(*AI) ? AI : 0;
1397  // See if we've already calculated (or started to calculate) alloca for a
1398  // given value.
1399  AllocaForValueMapTy::iterator I = AllocaForValue.find(V);
1400  if (I != AllocaForValue.end())
1401    return I->second;
1402  // Store 0 while we're calculating alloca for value V to avoid
1403  // infinite recursion if the value references itself.
1404  AllocaForValue[V] = 0;
1405  AllocaInst *Res = 0;
1406  if (CastInst *CI = dyn_cast<CastInst>(V))
1407    Res = findAllocaForValue(CI->getOperand(0));
1408  else if (PHINode *PN = dyn_cast<PHINode>(V)) {
1409    for (unsigned i = 0, e = PN->getNumIncomingValues(); i != e; ++i) {
1410      Value *IncValue = PN->getIncomingValue(i);
1411      // Allow self-referencing phi-nodes.
1412      if (IncValue == PN) continue;
1413      AllocaInst *IncValueAI = findAllocaForValue(IncValue);
1414      // AI for incoming values should exist and should all be equal.
1415      if (IncValueAI == 0 || (Res != 0 && IncValueAI != Res))
1416        return 0;
1417      Res = IncValueAI;
1418    }
1419  }
1420  if (Res != 0)
1421    AllocaForValue[V] = Res;
1422  return Res;
1423}
1424