AddressSanitizer.cpp revision 086a472dbed9911f83add781e020cb49c89829d0 
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/CallSite.h"
39#include "llvm/Support/CommandLine.h"
40#include "llvm/Support/DataTypes.h"
41#include "llvm/Support/Debug.h"
42#include "llvm/Support/raw_ostream.h"
43#include "llvm/Support/system_error.h"
44#include "llvm/Target/TargetMachine.h"
45#include "llvm/Transforms/Utils/BasicBlockUtils.h"
46#include "llvm/Transforms/Utils/BlackList.h"
47#include "llvm/Transforms/Utils/Local.h"
48#include "llvm/Transforms/Utils/ModuleUtils.h"
49#include <algorithm>
50#include <string>
51
52using namespace llvm;
53
54static const uint64_t kDefaultShadowScale = 3;
55static const uint64_t kDefaultShadowOffset32 = 1ULL << 29;
56static const uint64_t kDefaultShadowOffset64 = 1ULL << 44;
57static const uint64_t kDefaultShort64bitShadowOffset = 0x7FFF8000;  // < 2G.
58static const uint64_t kPPC64_ShadowOffset64 = 1ULL << 41;
59
60static const size_t kMaxStackMallocSize = 1 << 16;  // 64K
61static const uintptr_t kCurrentStackFrameMagic = 0x41B58AB3;
62static const uintptr_t kRetiredStackFrameMagic = 0x45E0360E;
63
64static const char *kAsanModuleCtorName = "asan.module_ctor";
65static const char *kAsanModuleDtorName = "asan.module_dtor";
66static const int   kAsanCtorAndCtorPriority = 1;
67static const char *kAsanReportErrorTemplate = "__asan_report_";
68static const char *kAsanReportLoadN = "__asan_report_load_n";
69static const char *kAsanReportStoreN = "__asan_report_store_n";
70static const char *kAsanRegisterGlobalsName = "__asan_register_globals";
71static const char *kAsanUnregisterGlobalsName = "__asan_unregister_globals";
72static const char *kAsanPoisonGlobalsName = "__asan_before_dynamic_init";
73static const char *kAsanUnpoisonGlobalsName = "__asan_after_dynamic_init";
74static const char *kAsanInitName = "__asan_init_v2";
75static const char *kAsanHandleNoReturnName = "__asan_handle_no_return";
76static const char *kAsanMappingOffsetName = "__asan_mapping_offset";
77static const char *kAsanMappingScaleName = "__asan_mapping_scale";
78static const char *kAsanStackMallocName = "__asan_stack_malloc";
79static const char *kAsanStackFreeName = "__asan_stack_free";
80static const char *kAsanGenPrefix = "__asan_gen_";
81static const char *kAsanPoisonStackMemoryName = "__asan_poison_stack_memory";
82static const char *kAsanUnpoisonStackMemoryName =
83    "__asan_unpoison_stack_memory";
84
85static const int kAsanStackLeftRedzoneMagic = 0xf1;
86static const int kAsanStackMidRedzoneMagic = 0xf2;
87static const int kAsanStackRightRedzoneMagic = 0xf3;
88static const int kAsanStackPartialRedzoneMagic = 0xf4;
89
90// Accesses sizes are powers of two: 1, 2, 4, 8, 16.
91static const size_t kNumberOfAccessSizes = 5;
92
93// Command-line flags.
94
95// This flag may need to be replaced with -f[no-]asan-reads.
96static cl::opt<bool> ClInstrumentReads("asan-instrument-reads",
97       cl::desc("instrument read instructions"), cl::Hidden, cl::init(true));
98static cl::opt<bool> ClInstrumentWrites("asan-instrument-writes",
99       cl::desc("instrument write instructions"), cl::Hidden, cl::init(true));
100static cl::opt<bool> ClInstrumentAtomics("asan-instrument-atomics",
101       cl::desc("instrument atomic instructions (rmw, cmpxchg)"),
102       cl::Hidden, cl::init(true));
103static cl::opt<bool> ClAlwaysSlowPath("asan-always-slow-path",
104       cl::desc("use instrumentation with slow path for all accesses"),
105       cl::Hidden, cl::init(false));
106// This flag limits the number of instructions to be instrumented
107// in any given BB. Normally, this should be set to unlimited (INT_MAX),
108// but due to http://llvm.org/bugs/show_bug.cgi?id=12652 we temporary
109// set it to 10000.
110static cl::opt<int> ClMaxInsnsToInstrumentPerBB("asan-max-ins-per-bb",
111       cl::init(10000),
112       cl::desc("maximal number of instructions to instrument in any given BB"),
113       cl::Hidden);
114// This flag may need to be replaced with -f[no]asan-stack.
115static cl::opt<bool> ClStack("asan-stack",
116       cl::desc("Handle stack memory"), cl::Hidden, cl::init(true));
117// This flag may need to be replaced with -f[no]asan-use-after-return.
118static cl::opt<bool> ClUseAfterReturn("asan-use-after-return",
119       cl::desc("Check return-after-free"), cl::Hidden, cl::init(false));
120// This flag may need to be replaced with -f[no]asan-globals.
121static cl::opt<bool> ClGlobals("asan-globals",
122       cl::desc("Handle global objects"), cl::Hidden, cl::init(true));
123static cl::opt<bool> ClInitializers("asan-initialization-order",
124       cl::desc("Handle C++ initializer order"), cl::Hidden, cl::init(false));
125static cl::opt<bool> ClMemIntrin("asan-memintrin",
126       cl::desc("Handle memset/memcpy/memmove"), cl::Hidden, cl::init(true));
127static cl::opt<bool> ClRealignStack("asan-realign-stack",
128       cl::desc("Realign stack to 32"), cl::Hidden, cl::init(true));
129static cl::opt<std::string> ClBlacklistFile("asan-blacklist",
130       cl::desc("File containing the list of objects to ignore "
131                "during instrumentation"), cl::Hidden);
132
133// These flags allow to change the shadow mapping.
134// The shadow mapping looks like
135//    Shadow = (Mem >> scale) + (1 << offset_log)
136static cl::opt<int> ClMappingScale("asan-mapping-scale",
137       cl::desc("scale of asan shadow mapping"), cl::Hidden, cl::init(0));
138static cl::opt<int> ClMappingOffsetLog("asan-mapping-offset-log",
139       cl::desc("offset of asan shadow mapping"), cl::Hidden, cl::init(-1));
140static cl::opt<bool> ClShort64BitOffset("asan-short-64bit-mapping-offset",
141       cl::desc("Use short immediate constant as the mapping offset for 64bit"),
142       cl::Hidden, cl::init(true));
143
144// Optimization flags. Not user visible, used mostly for testing
145// and benchmarking the tool.
146static cl::opt<bool> ClOpt("asan-opt",
147       cl::desc("Optimize instrumentation"), cl::Hidden, cl::init(true));
148static cl::opt<bool> ClOptSameTemp("asan-opt-same-temp",
149       cl::desc("Instrument the same temp just once"), cl::Hidden,
150       cl::init(true));
151static cl::opt<bool> ClOptGlobals("asan-opt-globals",
152       cl::desc("Don't instrument scalar globals"), cl::Hidden, cl::init(true));
153
154static cl::opt<bool> ClCheckLifetime("asan-check-lifetime",
155       cl::desc("Use llvm.lifetime intrinsics to insert extra checks"),
156       cl::Hidden, cl::init(false));
157
158// Debug flags.
159static cl::opt<int> ClDebug("asan-debug", cl::desc("debug"), cl::Hidden,
160                            cl::init(0));
161static cl::opt<int> ClDebugStack("asan-debug-stack", cl::desc("debug stack"),
162                                 cl::Hidden, cl::init(0));
163static cl::opt<std::string> ClDebugFunc("asan-debug-func",
164                                        cl::Hidden, cl::desc("Debug func"));
165static cl::opt<int> ClDebugMin("asan-debug-min", cl::desc("Debug min inst"),
166                               cl::Hidden, cl::init(-1));
167static cl::opt<int> ClDebugMax("asan-debug-max", cl::desc("Debug man inst"),
168                               cl::Hidden, cl::init(-1));
169
170namespace {
171/// A set of dynamically initialized globals extracted from metadata.
172class SetOfDynamicallyInitializedGlobals {
173 public:
174  void Init(Module& M) {
175    // Clang generates metadata identifying all dynamically initialized globals.
176    NamedMDNode *DynamicGlobals =
177        M.getNamedMetadata("llvm.asan.dynamically_initialized_globals");
178    if (!DynamicGlobals)
179      return;
180    for (int i = 0, n = DynamicGlobals->getNumOperands(); i < n; ++i) {
181      MDNode *MDN = DynamicGlobals->getOperand(i);
182      assert(MDN->getNumOperands() == 1);
183      Value *VG = MDN->getOperand(0);
184      // The optimizer may optimize away a global entirely, in which case we
185      // cannot instrument access to it.
186      if (!VG)
187        continue;
188      DynInitGlobals.insert(cast<GlobalVariable>(VG));
189    }
190  }
191  bool Contains(GlobalVariable *G) { return DynInitGlobals.count(G) != 0; }
192 private:
193  SmallSet<GlobalValue*, 32> DynInitGlobals;
194};
195
196/// This struct defines the shadow mapping using the rule:
197///   shadow = (mem >> Scale) ADD-or-OR Offset.
198struct ShadowMapping {
199  int Scale;
200  uint64_t Offset;
201  bool OrShadowOffset;
202};
203
204static ShadowMapping getShadowMapping(const Module &M, int LongSize,
205                                      bool ZeroBaseShadow) {
206  llvm::Triple TargetTriple(M.getTargetTriple());
207  bool IsAndroid = TargetTriple.getEnvironment() == llvm::Triple::Android;
208  bool IsMacOSX = TargetTriple.getOS() == llvm::Triple::MacOSX;
209  bool IsPPC64 = TargetTriple.getArch() == llvm::Triple::ppc64;
210  bool IsX86_64 = TargetTriple.getArch() == llvm::Triple::x86_64;
211
212  ShadowMapping Mapping;
213
214  // OR-ing shadow offset if more efficient (at least on x86),
215  // but on ppc64 we have to use add since the shadow offset is not neccesary
216  // 1/8-th of the address space.
217  Mapping.OrShadowOffset = !IsPPC64 && !ClShort64BitOffset;
218
219  Mapping.Offset = (IsAndroid || ZeroBaseShadow) ? 0 :
220      (LongSize == 32 ? kDefaultShadowOffset32 :
221       IsPPC64 ? kPPC64_ShadowOffset64 : kDefaultShadowOffset64);
222  if (!ZeroBaseShadow && ClShort64BitOffset && IsX86_64 && !IsMacOSX) {
223    assert(LongSize == 64);
224    Mapping.Offset = kDefaultShort64bitShadowOffset;
225  }
226  if (!ZeroBaseShadow && ClMappingOffsetLog >= 0) {
227    // Zero offset log is the special case.
228    Mapping.Offset = (ClMappingOffsetLog == 0) ? 0 : 1ULL << ClMappingOffsetLog;
229  }
230
231  Mapping.Scale = kDefaultShadowScale;
232  if (ClMappingScale) {
233    Mapping.Scale = ClMappingScale;
234  }
235
236  return Mapping;
237}
238
239static size_t RedzoneSizeForScale(int MappingScale) {
240  // Redzone used for stack and globals is at least 32 bytes.
241  // For scales 6 and 7, the redzone has to be 64 and 128 bytes respectively.
242  return std::max(32U, 1U << MappingScale);
243}
244
245/// AddressSanitizer: instrument the code in module to find memory bugs.
246struct AddressSanitizer : public FunctionPass {
247  AddressSanitizer(bool CheckInitOrder = true,
248                   bool CheckUseAfterReturn = false,
249                   bool CheckLifetime = false,
250                   StringRef BlacklistFile = StringRef(),
251                   bool ZeroBaseShadow = false)
252      : FunctionPass(ID),
253        CheckInitOrder(CheckInitOrder || ClInitializers),
254        CheckUseAfterReturn(CheckUseAfterReturn || ClUseAfterReturn),
255        CheckLifetime(CheckLifetime || ClCheckLifetime),
256        BlacklistFile(BlacklistFile.empty() ? ClBlacklistFile
257                                            : BlacklistFile),
258        ZeroBaseShadow(ZeroBaseShadow) {}
259  virtual const char *getPassName() const {
260    return "AddressSanitizerFunctionPass";
261  }
262  void instrumentMop(Instruction *I);
263  void instrumentAddress(Instruction *OrigIns, Instruction *InsertBefore,
264                         Value *Addr, uint32_t TypeSize, bool IsWrite,
265                         Value *SizeArgument);
266  Value *createSlowPathCmp(IRBuilder<> &IRB, Value *AddrLong,
267                           Value *ShadowValue, uint32_t TypeSize);
268  Instruction *generateCrashCode(Instruction *InsertBefore, Value *Addr,
269                                 bool IsWrite, size_t AccessSizeIndex,
270                                 Value *SizeArgument);
271  bool instrumentMemIntrinsic(MemIntrinsic *MI);
272  void instrumentMemIntrinsicParam(Instruction *OrigIns, Value *Addr,
273                                   Value *Size,
274                                   Instruction *InsertBefore, bool IsWrite);
275  Value *memToShadow(Value *Shadow, IRBuilder<> &IRB);
276  bool runOnFunction(Function &F);
277  void createInitializerPoisonCalls(Module &M,
278                                    Value *FirstAddr, Value *LastAddr);
279  bool maybeInsertAsanInitAtFunctionEntry(Function &F);
280  void emitShadowMapping(Module &M, IRBuilder<> &IRB) const;
281  virtual bool doInitialization(Module &M);
282  static char ID;  // Pass identification, replacement for typeid
283
284 private:
285  void initializeCallbacks(Module &M);
286
287  bool ShouldInstrumentGlobal(GlobalVariable *G);
288  bool LooksLikeCodeInBug11395(Instruction *I);
289  void FindDynamicInitializers(Module &M);
290
291  bool CheckInitOrder;
292  bool CheckUseAfterReturn;
293  bool CheckLifetime;
294  SmallString<64> BlacklistFile;
295  bool ZeroBaseShadow;
296
297  LLVMContext *C;
298  DataLayout *TD;
299  int LongSize;
300  Type *IntptrTy;
301  ShadowMapping Mapping;
302  Function *AsanCtorFunction;
303  Function *AsanInitFunction;
304  Function *AsanHandleNoReturnFunc;
305  OwningPtr<BlackList> BL;
306  // This array is indexed by AccessIsWrite and log2(AccessSize).
307  Function *AsanErrorCallback[2][kNumberOfAccessSizes];
308  // This array is indexed by AccessIsWrite.
309  Function *AsanErrorCallbackSized[2];
310  InlineAsm *EmptyAsm;
311  SetOfDynamicallyInitializedGlobals DynamicallyInitializedGlobals;
312
313  friend struct FunctionStackPoisoner;
314};
315
316class AddressSanitizerModule : public ModulePass {
317 public:
318  AddressSanitizerModule(bool CheckInitOrder = true,
319                         StringRef BlacklistFile = StringRef(),
320                         bool ZeroBaseShadow = false)
321      : ModulePass(ID),
322        CheckInitOrder(CheckInitOrder || ClInitializers),
323        BlacklistFile(BlacklistFile.empty() ? ClBlacklistFile
324                                            : BlacklistFile),
325        ZeroBaseShadow(ZeroBaseShadow) {}
326  bool runOnModule(Module &M);
327  static char ID;  // Pass identification, replacement for typeid
328  virtual const char *getPassName() const {
329    return "AddressSanitizerModule";
330  }
331
332 private:
333  void initializeCallbacks(Module &M);
334
335  bool ShouldInstrumentGlobal(GlobalVariable *G);
336  void createInitializerPoisonCalls(Module &M, Value *FirstAddr,
337                                    Value *LastAddr);
338  size_t RedzoneSize() const {
339    return RedzoneSizeForScale(Mapping.Scale);
340  }
341
342  bool CheckInitOrder;
343  SmallString<64> BlacklistFile;
344  bool ZeroBaseShadow;
345
346  OwningPtr<BlackList> BL;
347  SetOfDynamicallyInitializedGlobals DynamicallyInitializedGlobals;
348  Type *IntptrTy;
349  LLVMContext *C;
350  DataLayout *TD;
351  ShadowMapping Mapping;
352  Function *AsanPoisonGlobals;
353  Function *AsanUnpoisonGlobals;
354  Function *AsanRegisterGlobals;
355  Function *AsanUnregisterGlobals;
356};
357
358// Stack poisoning does not play well with exception handling.
359// When an exception is thrown, we essentially bypass the code
360// that unpoisones the stack. This is why the run-time library has
361// to intercept __cxa_throw (as well as longjmp, etc) and unpoison the entire
362// stack in the interceptor. This however does not work inside the
363// actual function which catches the exception. Most likely because the
364// compiler hoists the load of the shadow value somewhere too high.
365// This causes asan to report a non-existing bug on 453.povray.
366// It sounds like an LLVM bug.
367struct FunctionStackPoisoner : public InstVisitor<FunctionStackPoisoner> {
368  Function &F;
369  AddressSanitizer &ASan;
370  DIBuilder DIB;
371  LLVMContext *C;
372  Type *IntptrTy;
373  Type *IntptrPtrTy;
374  ShadowMapping Mapping;
375
376  SmallVector<AllocaInst*, 16> AllocaVec;
377  SmallVector<Instruction*, 8> RetVec;
378  uint64_t TotalStackSize;
379  unsigned StackAlignment;
380
381  Function *AsanStackMallocFunc, *AsanStackFreeFunc;
382  Function *AsanPoisonStackMemoryFunc, *AsanUnpoisonStackMemoryFunc;
383
384  // Stores a place and arguments of poisoning/unpoisoning call for alloca.
385  struct AllocaPoisonCall {
386    IntrinsicInst *InsBefore;
387    uint64_t Size;
388    bool DoPoison;
389  };
390  SmallVector<AllocaPoisonCall, 8> AllocaPoisonCallVec;
391
392  // Maps Value to an AllocaInst from which the Value is originated.
393  typedef DenseMap<Value*, AllocaInst*> AllocaForValueMapTy;
394  AllocaForValueMapTy AllocaForValue;
395
396  FunctionStackPoisoner(Function &F, AddressSanitizer &ASan)
397      : F(F), ASan(ASan), DIB(*F.getParent()), C(ASan.C),
398        IntptrTy(ASan.IntptrTy), IntptrPtrTy(PointerType::get(IntptrTy, 0)),
399        Mapping(ASan.Mapping),
400        TotalStackSize(0), StackAlignment(1 << Mapping.Scale) {}
401
402  bool runOnFunction() {
403    if (!ClStack) return false;
404    // Collect alloca, ret, lifetime instructions etc.
405    for (df_iterator<BasicBlock*> DI = df_begin(&F.getEntryBlock()),
406         DE = df_end(&F.getEntryBlock()); DI != DE; ++DI) {
407      BasicBlock *BB = *DI;
408      visit(*BB);
409    }
410    if (AllocaVec.empty()) return false;
411
412    initializeCallbacks(*F.getParent());
413
414    poisonStack();
415
416    if (ClDebugStack) {
417      DEBUG(dbgs() << F);
418    }
419    return true;
420  }
421
422  // Finds all static Alloca instructions and puts
423  // poisoned red zones around all of them.
424  // Then unpoison everything back before the function returns.
425  void poisonStack();
426
427  // ----------------------- Visitors.
428  /// \brief Collect all Ret instructions.
429  void visitReturnInst(ReturnInst &RI) {
430    RetVec.push_back(&RI);
431  }
432
433  /// \brief Collect Alloca instructions we want (and can) handle.
434  void visitAllocaInst(AllocaInst &AI) {
435    if (!isInterestingAlloca(AI)) return;
436
437    StackAlignment = std::max(StackAlignment, AI.getAlignment());
438    AllocaVec.push_back(&AI);
439    uint64_t AlignedSize =  getAlignedAllocaSize(&AI);
440    TotalStackSize += AlignedSize;
441  }
442
443  /// \brief Collect lifetime intrinsic calls to check for use-after-scope
444  /// errors.
445  void visitIntrinsicInst(IntrinsicInst &II) {
446    if (!ASan.CheckLifetime) return;
447    Intrinsic::ID ID = II.getIntrinsicID();
448    if (ID != Intrinsic::lifetime_start &&
449        ID != Intrinsic::lifetime_end)
450      return;
451    // Found lifetime intrinsic, add ASan instrumentation if necessary.
452    ConstantInt *Size = dyn_cast<ConstantInt>(II.getArgOperand(0));
453    // If size argument is undefined, don't do anything.
454    if (Size->isMinusOne()) return;
455    // Check that size doesn't saturate uint64_t and can
456    // be stored in IntptrTy.
457    const uint64_t SizeValue = Size->getValue().getLimitedValue();
458    if (SizeValue == ~0ULL ||
459        !ConstantInt::isValueValidForType(IntptrTy, SizeValue))
460      return;
461    // Find alloca instruction that corresponds to llvm.lifetime argument.
462    AllocaInst *AI = findAllocaForValue(II.getArgOperand(1));
463    if (!AI) return;
464    bool DoPoison = (ID == Intrinsic::lifetime_end);
465    AllocaPoisonCall APC = {&II, SizeValue, DoPoison};
466    AllocaPoisonCallVec.push_back(APC);
467  }
468
469  // ---------------------- Helpers.
470  void initializeCallbacks(Module &M);
471
472  // Check if we want (and can) handle this alloca.
473  bool isInterestingAlloca(AllocaInst &AI) {
474    return (!AI.isArrayAllocation() &&
475            AI.isStaticAlloca() &&
476            AI.getAllocatedType()->isSized());
477  }
478
479  size_t RedzoneSize() const {
480    return RedzoneSizeForScale(Mapping.Scale);
481  }
482  uint64_t getAllocaSizeInBytes(AllocaInst *AI) {
483    Type *Ty = AI->getAllocatedType();
484    uint64_t SizeInBytes = ASan.TD->getTypeAllocSize(Ty);
485    return SizeInBytes;
486  }
487  uint64_t getAlignedSize(uint64_t SizeInBytes) {
488    size_t RZ = RedzoneSize();
489    return ((SizeInBytes + RZ - 1) / RZ) * RZ;
490  }
491  uint64_t getAlignedAllocaSize(AllocaInst *AI) {
492    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
493    return getAlignedSize(SizeInBytes);
494  }
495  /// Finds alloca where the value comes from.
496  AllocaInst *findAllocaForValue(Value *V);
497  void poisonRedZones(const ArrayRef<AllocaInst*> &AllocaVec, IRBuilder<> IRB,
498                      Value *ShadowBase, bool DoPoison);
499  void poisonAlloca(Value *V, uint64_t Size, IRBuilder<> IRB, bool DoPoison);
500};
501
502}  // namespace
503
504char AddressSanitizer::ID = 0;
505INITIALIZE_PASS(AddressSanitizer, "asan",
506    "AddressSanitizer: detects use-after-free and out-of-bounds bugs.",
507    false, false)
508FunctionPass *llvm::createAddressSanitizerFunctionPass(
509    bool CheckInitOrder, bool CheckUseAfterReturn, bool CheckLifetime,
510    StringRef BlacklistFile, bool ZeroBaseShadow) {
511  return new AddressSanitizer(CheckInitOrder, CheckUseAfterReturn,
512                              CheckLifetime, BlacklistFile, ZeroBaseShadow);
513}
514
515char AddressSanitizerModule::ID = 0;
516INITIALIZE_PASS(AddressSanitizerModule, "asan-module",
517    "AddressSanitizer: detects use-after-free and out-of-bounds bugs."
518    "ModulePass", false, false)
519ModulePass *llvm::createAddressSanitizerModulePass(
520    bool CheckInitOrder, StringRef BlacklistFile, bool ZeroBaseShadow) {
521  return new AddressSanitizerModule(CheckInitOrder, BlacklistFile,
522                                    ZeroBaseShadow);
523}
524
525static size_t TypeSizeToSizeIndex(uint32_t TypeSize) {
526  size_t Res = CountTrailingZeros_32(TypeSize / 8);
527  assert(Res < kNumberOfAccessSizes);
528  return Res;
529}
530
531// Create a constant for Str so that we can pass it to the run-time lib.
532static GlobalVariable *createPrivateGlobalForString(Module &M, StringRef Str) {
533  Constant *StrConst = ConstantDataArray::getString(M.getContext(), Str);
534  return new GlobalVariable(M, StrConst->getType(), true,
535                            GlobalValue::PrivateLinkage, StrConst,
536                            kAsanGenPrefix);
537}
538
539static bool GlobalWasGeneratedByAsan(GlobalVariable *G) {
540  return G->getName().find(kAsanGenPrefix) == 0;
541}
542
543Value *AddressSanitizer::memToShadow(Value *Shadow, IRBuilder<> &IRB) {
544  // Shadow >> scale
545  Shadow = IRB.CreateLShr(Shadow, Mapping.Scale);
546  if (Mapping.Offset == 0)
547    return Shadow;
548  // (Shadow >> scale) | offset
549  if (Mapping.OrShadowOffset)
550    return IRB.CreateOr(Shadow, ConstantInt::get(IntptrTy, Mapping.Offset));
551  else
552    return IRB.CreateAdd(Shadow, ConstantInt::get(IntptrTy, Mapping.Offset));
553}
554
555void AddressSanitizer::instrumentMemIntrinsicParam(
556    Instruction *OrigIns,
557    Value *Addr, Value *Size, Instruction *InsertBefore, bool IsWrite) {
558  IRBuilder<> IRB(InsertBefore);
559  if (Size->getType() != IntptrTy)
560    Size = IRB.CreateIntCast(Size, IntptrTy, false);
561  // Check the first byte.
562  instrumentAddress(OrigIns, InsertBefore, Addr, 8, IsWrite, Size);
563  // Check the last byte.
564  IRB.SetInsertPoint(InsertBefore);
565  Value *SizeMinusOne = IRB.CreateSub(Size, ConstantInt::get(IntptrTy, 1));
566  Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
567  Value *AddrLast = IRB.CreateAdd(AddrLong, SizeMinusOne);
568  instrumentAddress(OrigIns, InsertBefore, AddrLast, 8, IsWrite, Size);
569}
570
571// Instrument memset/memmove/memcpy
572bool AddressSanitizer::instrumentMemIntrinsic(MemIntrinsic *MI) {
573  Value *Dst = MI->getDest();
574  MemTransferInst *MemTran = dyn_cast<MemTransferInst>(MI);
575  Value *Src = MemTran ? MemTran->getSource() : 0;
576  Value *Length = MI->getLength();
577
578  Constant *ConstLength = dyn_cast<Constant>(Length);
579  Instruction *InsertBefore = MI;
580  if (ConstLength) {
581    if (ConstLength->isNullValue()) return false;
582  } else {
583    // The size is not a constant so it could be zero -- check at run-time.
584    IRBuilder<> IRB(InsertBefore);
585
586    Value *Cmp = IRB.CreateICmpNE(Length,
587                                  Constant::getNullValue(Length->getType()));
588    InsertBefore = SplitBlockAndInsertIfThen(cast<Instruction>(Cmp), false);
589  }
590
591  instrumentMemIntrinsicParam(MI, Dst, Length, InsertBefore, true);
592  if (Src)
593    instrumentMemIntrinsicParam(MI, Src, Length, InsertBefore, false);
594  return true;
595}
596
597// If I is an interesting memory access, return the PointerOperand
598// and set IsWrite. Otherwise return NULL.
599static Value *isInterestingMemoryAccess(Instruction *I, bool *IsWrite) {
600  if (LoadInst *LI = dyn_cast<LoadInst>(I)) {
601    if (!ClInstrumentReads) return NULL;
602    *IsWrite = false;
603    return LI->getPointerOperand();
604  }
605  if (StoreInst *SI = dyn_cast<StoreInst>(I)) {
606    if (!ClInstrumentWrites) return NULL;
607    *IsWrite = true;
608    return SI->getPointerOperand();
609  }
610  if (AtomicRMWInst *RMW = dyn_cast<AtomicRMWInst>(I)) {
611    if (!ClInstrumentAtomics) return NULL;
612    *IsWrite = true;
613    return RMW->getPointerOperand();
614  }
615  if (AtomicCmpXchgInst *XCHG = dyn_cast<AtomicCmpXchgInst>(I)) {
616    if (!ClInstrumentAtomics) return NULL;
617    *IsWrite = true;
618    return XCHG->getPointerOperand();
619  }
620  return NULL;
621}
622
623void AddressSanitizer::instrumentMop(Instruction *I) {
624  bool IsWrite = false;
625  Value *Addr = isInterestingMemoryAccess(I, &IsWrite);
626  assert(Addr);
627  if (ClOpt && ClOptGlobals) {
628    if (GlobalVariable *G = dyn_cast<GlobalVariable>(Addr)) {
629      // If initialization order checking is disabled, a simple access to a
630      // dynamically initialized global is always valid.
631      if (!CheckInitOrder)
632        return;
633      // If a global variable does not have dynamic initialization we don't
634      // have to instrument it.  However, if a global does not have initailizer
635      // at all, we assume it has dynamic initializer (in other TU).
636      if (G->hasInitializer() && !DynamicallyInitializedGlobals.Contains(G))
637        return;
638    }
639  }
640
641  Type *OrigPtrTy = Addr->getType();
642  Type *OrigTy = cast<PointerType>(OrigPtrTy)->getElementType();
643
644  assert(OrigTy->isSized());
645  uint32_t TypeSize = TD->getTypeStoreSizeInBits(OrigTy);
646
647  assert((TypeSize % 8) == 0);
648
649  // Instrument a 1-, 2-, 4-, 8-, or 16- byte access with one check.
650  if (TypeSize == 8  || TypeSize == 16 ||
651      TypeSize == 32 || TypeSize == 64 || TypeSize == 128)
652    return instrumentAddress(I, I, Addr, TypeSize, IsWrite, 0);
653  // Instrument unusual size (but still multiple of 8).
654  // We can not do it with a single check, so we do 1-byte check for the first
655  // and the last bytes. We call __asan_report_*_n(addr, real_size) to be able
656  // to report the actual access size.
657  IRBuilder<> IRB(I);
658  Value *LastByte =  IRB.CreateIntToPtr(
659      IRB.CreateAdd(IRB.CreatePointerCast(Addr, IntptrTy),
660                    ConstantInt::get(IntptrTy, TypeSize / 8 - 1)),
661      OrigPtrTy);
662  Value *Size = ConstantInt::get(IntptrTy, TypeSize / 8);
663  instrumentAddress(I, I, Addr, 8, IsWrite, Size);
664  instrumentAddress(I, I, LastByte, 8, IsWrite, Size);
665}
666
667// Validate the result of Module::getOrInsertFunction called for an interface
668// function of AddressSanitizer. If the instrumented module defines a function
669// with the same name, their prototypes must match, otherwise
670// getOrInsertFunction returns a bitcast.
671static Function *checkInterfaceFunction(Constant *FuncOrBitcast) {
672  if (isa<Function>(FuncOrBitcast)) return cast<Function>(FuncOrBitcast);
673  FuncOrBitcast->dump();
674  report_fatal_error("trying to redefine an AddressSanitizer "
675                     "interface function");
676}
677
678Instruction *AddressSanitizer::generateCrashCode(
679    Instruction *InsertBefore, Value *Addr,
680    bool IsWrite, size_t AccessSizeIndex, Value *SizeArgument) {
681  IRBuilder<> IRB(InsertBefore);
682  CallInst *Call = SizeArgument
683    ? IRB.CreateCall2(AsanErrorCallbackSized[IsWrite], Addr, SizeArgument)
684    : IRB.CreateCall(AsanErrorCallback[IsWrite][AccessSizeIndex], Addr);
685
686  // We don't do Call->setDoesNotReturn() because the BB already has
687  // UnreachableInst at the end.
688  // This EmptyAsm is required to avoid callback merge.
689  IRB.CreateCall(EmptyAsm);
690  return Call;
691}
692
693Value *AddressSanitizer::createSlowPathCmp(IRBuilder<> &IRB, Value *AddrLong,
694                                            Value *ShadowValue,
695                                            uint32_t TypeSize) {
696  size_t Granularity = 1 << Mapping.Scale;
697  // Addr & (Granularity - 1)
698  Value *LastAccessedByte = IRB.CreateAnd(
699      AddrLong, ConstantInt::get(IntptrTy, Granularity - 1));
700  // (Addr & (Granularity - 1)) + size - 1
701  if (TypeSize / 8 > 1)
702    LastAccessedByte = IRB.CreateAdd(
703        LastAccessedByte, ConstantInt::get(IntptrTy, TypeSize / 8 - 1));
704  // (uint8_t) ((Addr & (Granularity-1)) + size - 1)
705  LastAccessedByte = IRB.CreateIntCast(
706      LastAccessedByte, ShadowValue->getType(), false);
707  // ((uint8_t) ((Addr & (Granularity-1)) + size - 1)) >= ShadowValue
708  return IRB.CreateICmpSGE(LastAccessedByte, ShadowValue);
709}
710
711void AddressSanitizer::instrumentAddress(Instruction *OrigIns,
712                                         Instruction *InsertBefore,
713                                         Value *Addr, uint32_t TypeSize,
714                                         bool IsWrite, Value *SizeArgument) {
715  IRBuilder<> IRB(InsertBefore);
716  Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
717
718  Type *ShadowTy  = IntegerType::get(
719      *C, std::max(8U, TypeSize >> Mapping.Scale));
720  Type *ShadowPtrTy = PointerType::get(ShadowTy, 0);
721  Value *ShadowPtr = memToShadow(AddrLong, IRB);
722  Value *CmpVal = Constant::getNullValue(ShadowTy);
723  Value *ShadowValue = IRB.CreateLoad(
724      IRB.CreateIntToPtr(ShadowPtr, ShadowPtrTy));
725
726  Value *Cmp = IRB.CreateICmpNE(ShadowValue, CmpVal);
727  size_t AccessSizeIndex = TypeSizeToSizeIndex(TypeSize);
728  size_t Granularity = 1 << Mapping.Scale;
729  TerminatorInst *CrashTerm = 0;
730
731  if (ClAlwaysSlowPath || (TypeSize < 8 * Granularity)) {
732    TerminatorInst *CheckTerm =
733        SplitBlockAndInsertIfThen(cast<Instruction>(Cmp), false);
734    assert(dyn_cast<BranchInst>(CheckTerm)->isUnconditional());
735    BasicBlock *NextBB = CheckTerm->getSuccessor(0);
736    IRB.SetInsertPoint(CheckTerm);
737    Value *Cmp2 = createSlowPathCmp(IRB, AddrLong, ShadowValue, TypeSize);
738    BasicBlock *CrashBlock =
739        BasicBlock::Create(*C, "", NextBB->getParent(), NextBB);
740    CrashTerm = new UnreachableInst(*C, CrashBlock);
741    BranchInst *NewTerm = BranchInst::Create(CrashBlock, NextBB, Cmp2);
742    ReplaceInstWithInst(CheckTerm, NewTerm);
743  } else {
744    CrashTerm = SplitBlockAndInsertIfThen(cast<Instruction>(Cmp), true);
745  }
746
747  Instruction *Crash = generateCrashCode(
748      CrashTerm, AddrLong, IsWrite, AccessSizeIndex, SizeArgument);
749  Crash->setDebugLoc(OrigIns->getDebugLoc());
750}
751
752void AddressSanitizerModule::createInitializerPoisonCalls(
753    Module &M, Value *FirstAddr, Value *LastAddr) {
754  // We do all of our poisoning and unpoisoning within _GLOBAL__I_a.
755  Function *GlobalInit = M.getFunction("_GLOBAL__I_a");
756  // If that function is not present, this TU contains no globals, or they have
757  // all been optimized away
758  if (!GlobalInit)
759    return;
760
761  // Set up the arguments to our poison/unpoison functions.
762  IRBuilder<> IRB(GlobalInit->begin()->getFirstInsertionPt());
763
764  // Add a call to poison all external globals before the given function starts.
765  IRB.CreateCall2(AsanPoisonGlobals, FirstAddr, LastAddr);
766
767  // Add calls to unpoison all globals before each return instruction.
768  for (Function::iterator I = GlobalInit->begin(), E = GlobalInit->end();
769      I != E; ++I) {
770    if (ReturnInst *RI = dyn_cast<ReturnInst>(I->getTerminator())) {
771      CallInst::Create(AsanUnpoisonGlobals, "", RI);
772    }
773  }
774}
775
776bool AddressSanitizerModule::ShouldInstrumentGlobal(GlobalVariable *G) {
777  Type *Ty = cast<PointerType>(G->getType())->getElementType();
778  DEBUG(dbgs() << "GLOBAL: " << *G << "\n");
779
780  if (BL->isIn(*G)) return false;
781  if (!Ty->isSized()) return false;
782  if (!G->hasInitializer()) return false;
783  if (GlobalWasGeneratedByAsan(G)) return false;  // Our own global.
784  // Touch only those globals that will not be defined in other modules.
785  // Don't handle ODR type linkages since other modules may be built w/o asan.
786  if (G->getLinkage() != GlobalVariable::ExternalLinkage &&
787      G->getLinkage() != GlobalVariable::PrivateLinkage &&
788      G->getLinkage() != GlobalVariable::InternalLinkage)
789    return false;
790  // Two problems with thread-locals:
791  //   - The address of the main thread's copy can't be computed at link-time.
792  //   - Need to poison all copies, not just the main thread's one.
793  if (G->isThreadLocal())
794    return false;
795  // For now, just ignore this Alloca if the alignment is large.
796  if (G->getAlignment() > RedzoneSize()) return false;
797
798  // Ignore all the globals with the names starting with "\01L_OBJC_".
799  // Many of those are put into the .cstring section. The linker compresses
800  // that section by removing the spare \0s after the string terminator, so
801  // our redzones get broken.
802  if ((G->getName().find("\01L_OBJC_") == 0) ||
803      (G->getName().find("\01l_OBJC_") == 0)) {
804    DEBUG(dbgs() << "Ignoring \\01L_OBJC_* global: " << *G);
805    return false;
806  }
807
808  if (G->hasSection()) {
809    StringRef Section(G->getSection());
810    // Ignore the globals from the __OBJC section. The ObjC runtime assumes
811    // those conform to /usr/lib/objc/runtime.h, so we can't add redzones to
812    // them.
813    if ((Section.find("__OBJC,") == 0) ||
814        (Section.find("__DATA, __objc_") == 0)) {
815      DEBUG(dbgs() << "Ignoring ObjC runtime global: " << *G);
816      return false;
817    }
818    // See http://code.google.com/p/address-sanitizer/issues/detail?id=32
819    // Constant CFString instances are compiled in the following way:
820    //  -- the string buffer is emitted into
821    //     __TEXT,__cstring,cstring_literals
822    //  -- the constant NSConstantString structure referencing that buffer
823    //     is placed into __DATA,__cfstring
824    // Therefore there's no point in placing redzones into __DATA,__cfstring.
825    // Moreover, it causes the linker to crash on OS X 10.7
826    if (Section.find("__DATA,__cfstring") == 0) {
827      DEBUG(dbgs() << "Ignoring CFString: " << *G);
828      return false;
829    }
830  }
831
832  return true;
833}
834
835void AddressSanitizerModule::initializeCallbacks(Module &M) {
836  IRBuilder<> IRB(*C);
837  // Declare our poisoning and unpoisoning functions.
838  AsanPoisonGlobals = checkInterfaceFunction(M.getOrInsertFunction(
839      kAsanPoisonGlobalsName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
840  AsanPoisonGlobals->setLinkage(Function::ExternalLinkage);
841  AsanUnpoisonGlobals = checkInterfaceFunction(M.getOrInsertFunction(
842      kAsanUnpoisonGlobalsName, IRB.getVoidTy(), NULL));
843  AsanUnpoisonGlobals->setLinkage(Function::ExternalLinkage);
844  // Declare functions that register/unregister globals.
845  AsanRegisterGlobals = checkInterfaceFunction(M.getOrInsertFunction(
846      kAsanRegisterGlobalsName, IRB.getVoidTy(),
847      IntptrTy, IntptrTy, NULL));
848  AsanRegisterGlobals->setLinkage(Function::ExternalLinkage);
849  AsanUnregisterGlobals = checkInterfaceFunction(M.getOrInsertFunction(
850      kAsanUnregisterGlobalsName,
851      IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
852  AsanUnregisterGlobals->setLinkage(Function::ExternalLinkage);
853}
854
855// This function replaces all global variables with new variables that have
856// trailing redzones. It also creates a function that poisons
857// redzones and inserts this function into llvm.global_ctors.
858bool AddressSanitizerModule::runOnModule(Module &M) {
859  if (!ClGlobals) return false;
860  TD = getAnalysisIfAvailable<DataLayout>();
861  if (!TD)
862    return false;
863  BL.reset(new BlackList(BlacklistFile));
864  if (BL->isIn(M)) return false;
865  C = &(M.getContext());
866  int LongSize = TD->getPointerSizeInBits();
867  IntptrTy = Type::getIntNTy(*C, LongSize);
868  Mapping = getShadowMapping(M, LongSize, ZeroBaseShadow);
869  initializeCallbacks(M);
870  DynamicallyInitializedGlobals.Init(M);
871
872  SmallVector<GlobalVariable *, 16> GlobalsToChange;
873
874  for (Module::GlobalListType::iterator G = M.global_begin(),
875       E = M.global_end(); G != E; ++G) {
876    if (ShouldInstrumentGlobal(G))
877      GlobalsToChange.push_back(G);
878  }
879
880  size_t n = GlobalsToChange.size();
881  if (n == 0) return false;
882
883  // A global is described by a structure
884  //   size_t beg;
885  //   size_t size;
886  //   size_t size_with_redzone;
887  //   const char *name;
888  //   const char *module_name;
889  //   size_t has_dynamic_init;
890  // We initialize an array of such structures and pass it to a run-time call.
891  StructType *GlobalStructTy = StructType::get(IntptrTy, IntptrTy,
892                                               IntptrTy, IntptrTy,
893                                               IntptrTy, IntptrTy, NULL);
894  SmallVector<Constant *, 16> Initializers(n), DynamicInit;
895
896
897  Function *CtorFunc = M.getFunction(kAsanModuleCtorName);
898  assert(CtorFunc);
899  IRBuilder<> IRB(CtorFunc->getEntryBlock().getTerminator());
900
901  // The addresses of the first and last dynamically initialized globals in
902  // this TU.  Used in initialization order checking.
903  Value *FirstDynamic = 0, *LastDynamic = 0;
904
905  GlobalVariable *ModuleName = createPrivateGlobalForString(
906      M, M.getModuleIdentifier());
907
908  for (size_t i = 0; i < n; i++) {
909    static const uint64_t kMaxGlobalRedzone = 1 << 18;
910    GlobalVariable *G = GlobalsToChange[i];
911    PointerType *PtrTy = cast<PointerType>(G->getType());
912    Type *Ty = PtrTy->getElementType();
913    uint64_t SizeInBytes = TD->getTypeAllocSize(Ty);
914    uint64_t MinRZ = RedzoneSize();
915    // MinRZ <= RZ <= kMaxGlobalRedzone
916    // and trying to make RZ to be ~ 1/4 of SizeInBytes.
917    uint64_t RZ = std::max(MinRZ,
918                         std::min(kMaxGlobalRedzone,
919                                  (SizeInBytes / MinRZ / 4) * MinRZ));
920    uint64_t RightRedzoneSize = RZ;
921    // Round up to MinRZ
922    if (SizeInBytes % MinRZ)
923      RightRedzoneSize += MinRZ - (SizeInBytes % MinRZ);
924    assert(((RightRedzoneSize + SizeInBytes) % MinRZ) == 0);
925    Type *RightRedZoneTy = ArrayType::get(IRB.getInt8Ty(), RightRedzoneSize);
926    // Determine whether this global should be poisoned in initialization.
927    bool GlobalHasDynamicInitializer =
928        DynamicallyInitializedGlobals.Contains(G);
929    // Don't check initialization order if this global is blacklisted.
930    GlobalHasDynamicInitializer &= !BL->isInInit(*G);
931
932    StructType *NewTy = StructType::get(Ty, RightRedZoneTy, NULL);
933    Constant *NewInitializer = ConstantStruct::get(
934        NewTy, G->getInitializer(),
935        Constant::getNullValue(RightRedZoneTy), NULL);
936
937    GlobalVariable *Name = createPrivateGlobalForString(M, G->getName());
938
939    // Create a new global variable with enough space for a redzone.
940    GlobalVariable *NewGlobal = new GlobalVariable(
941        M, NewTy, G->isConstant(), G->getLinkage(),
942        NewInitializer, "", G, G->getThreadLocalMode());
943    NewGlobal->copyAttributesFrom(G);
944    NewGlobal->setAlignment(MinRZ);
945
946    Value *Indices2[2];
947    Indices2[0] = IRB.getInt32(0);
948    Indices2[1] = IRB.getInt32(0);
949
950    G->replaceAllUsesWith(
951        ConstantExpr::getGetElementPtr(NewGlobal, Indices2, true));
952    NewGlobal->takeName(G);
953    G->eraseFromParent();
954
955    Initializers[i] = ConstantStruct::get(
956        GlobalStructTy,
957        ConstantExpr::getPointerCast(NewGlobal, IntptrTy),
958        ConstantInt::get(IntptrTy, SizeInBytes),
959        ConstantInt::get(IntptrTy, SizeInBytes + RightRedzoneSize),
960        ConstantExpr::getPointerCast(Name, IntptrTy),
961        ConstantExpr::getPointerCast(ModuleName, IntptrTy),
962        ConstantInt::get(IntptrTy, GlobalHasDynamicInitializer),
963        NULL);
964
965    // Populate the first and last globals declared in this TU.
966    if (CheckInitOrder && GlobalHasDynamicInitializer) {
967      LastDynamic = ConstantExpr::getPointerCast(NewGlobal, IntptrTy);
968      if (FirstDynamic == 0)
969        FirstDynamic = LastDynamic;
970    }
971
972    DEBUG(dbgs() << "NEW GLOBAL: " << *NewGlobal << "\n");
973  }
974
975  ArrayType *ArrayOfGlobalStructTy = ArrayType::get(GlobalStructTy, n);
976  GlobalVariable *AllGlobals = new GlobalVariable(
977      M, ArrayOfGlobalStructTy, false, GlobalVariable::PrivateLinkage,
978      ConstantArray::get(ArrayOfGlobalStructTy, Initializers), "");
979
980  // Create calls for poisoning before initializers run and unpoisoning after.
981  if (CheckInitOrder && FirstDynamic && LastDynamic)
982    createInitializerPoisonCalls(M, FirstDynamic, LastDynamic);
983  IRB.CreateCall2(AsanRegisterGlobals,
984                  IRB.CreatePointerCast(AllGlobals, IntptrTy),
985                  ConstantInt::get(IntptrTy, n));
986
987  // We also need to unregister globals at the end, e.g. when a shared library
988  // gets closed.
989  Function *AsanDtorFunction = Function::Create(
990      FunctionType::get(Type::getVoidTy(*C), false),
991      GlobalValue::InternalLinkage, kAsanModuleDtorName, &M);
992  BasicBlock *AsanDtorBB = BasicBlock::Create(*C, "", AsanDtorFunction);
993  IRBuilder<> IRB_Dtor(ReturnInst::Create(*C, AsanDtorBB));
994  IRB_Dtor.CreateCall2(AsanUnregisterGlobals,
995                       IRB.CreatePointerCast(AllGlobals, IntptrTy),
996                       ConstantInt::get(IntptrTy, n));
997  appendToGlobalDtors(M, AsanDtorFunction, kAsanCtorAndCtorPriority);
998
999  DEBUG(dbgs() << M);
1000  return true;
1001}
1002
1003void AddressSanitizer::initializeCallbacks(Module &M) {
1004  IRBuilder<> IRB(*C);
1005  // Create __asan_report* callbacks.
1006  for (size_t AccessIsWrite = 0; AccessIsWrite <= 1; AccessIsWrite++) {
1007    for (size_t AccessSizeIndex = 0; AccessSizeIndex < kNumberOfAccessSizes;
1008         AccessSizeIndex++) {
1009      // IsWrite and TypeSize are encoded in the function name.
1010      std::string FunctionName = std::string(kAsanReportErrorTemplate) +
1011          (AccessIsWrite ? "store" : "load") + itostr(1 << AccessSizeIndex);
1012      // If we are merging crash callbacks, they have two parameters.
1013      AsanErrorCallback[AccessIsWrite][AccessSizeIndex] =
1014          checkInterfaceFunction(M.getOrInsertFunction(
1015              FunctionName, IRB.getVoidTy(), IntptrTy, NULL));
1016    }
1017  }
1018  AsanErrorCallbackSized[0] = checkInterfaceFunction(M.getOrInsertFunction(
1019              kAsanReportLoadN, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
1020  AsanErrorCallbackSized[1] = checkInterfaceFunction(M.getOrInsertFunction(
1021              kAsanReportStoreN, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
1022
1023  AsanHandleNoReturnFunc = checkInterfaceFunction(M.getOrInsertFunction(
1024      kAsanHandleNoReturnName, IRB.getVoidTy(), NULL));
1025  // We insert an empty inline asm after __asan_report* to avoid callback merge.
1026  EmptyAsm = InlineAsm::get(FunctionType::get(IRB.getVoidTy(), false),
1027                            StringRef(""), StringRef(""),
1028                            /*hasSideEffects=*/true);
1029}
1030
1031void AddressSanitizer::emitShadowMapping(Module &M, IRBuilder<> &IRB) const {
1032  // Tell the values of mapping offset and scale to the run-time.
1033  GlobalValue *asan_mapping_offset =
1034      new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
1035                     ConstantInt::get(IntptrTy, Mapping.Offset),
1036                     kAsanMappingOffsetName);
1037  // Read the global, otherwise it may be optimized away.
1038  IRB.CreateLoad(asan_mapping_offset, true);
1039
1040  GlobalValue *asan_mapping_scale =
1041      new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
1042                         ConstantInt::get(IntptrTy, Mapping.Scale),
1043                         kAsanMappingScaleName);
1044  // Read the global, otherwise it may be optimized away.
1045  IRB.CreateLoad(asan_mapping_scale, true);
1046}
1047
1048// virtual
1049bool AddressSanitizer::doInitialization(Module &M) {
1050  // Initialize the private fields. No one has accessed them before.
1051  TD = getAnalysisIfAvailable<DataLayout>();
1052
1053  if (!TD)
1054    return false;
1055  BL.reset(new BlackList(BlacklistFile));
1056  DynamicallyInitializedGlobals.Init(M);
1057
1058  C = &(M.getContext());
1059  LongSize = TD->getPointerSizeInBits();
1060  IntptrTy = Type::getIntNTy(*C, LongSize);
1061
1062  AsanCtorFunction = Function::Create(
1063      FunctionType::get(Type::getVoidTy(*C), false),
1064      GlobalValue::InternalLinkage, kAsanModuleCtorName, &M);
1065  BasicBlock *AsanCtorBB = BasicBlock::Create(*C, "", AsanCtorFunction);
1066  // call __asan_init in the module ctor.
1067  IRBuilder<> IRB(ReturnInst::Create(*C, AsanCtorBB));
1068  AsanInitFunction = checkInterfaceFunction(
1069      M.getOrInsertFunction(kAsanInitName, IRB.getVoidTy(), NULL));
1070  AsanInitFunction->setLinkage(Function::ExternalLinkage);
1071  IRB.CreateCall(AsanInitFunction);
1072
1073  Mapping = getShadowMapping(M, LongSize, ZeroBaseShadow);
1074  emitShadowMapping(M, IRB);
1075
1076  appendToGlobalCtors(M, AsanCtorFunction, kAsanCtorAndCtorPriority);
1077  return true;
1078}
1079
1080bool AddressSanitizer::maybeInsertAsanInitAtFunctionEntry(Function &F) {
1081  // For each NSObject descendant having a +load method, this method is invoked
1082  // by the ObjC runtime before any of the static constructors is called.
1083  // Therefore we need to instrument such methods with a call to __asan_init
1084  // at the beginning in order to initialize our runtime before any access to
1085  // the shadow memory.
1086  // We cannot just ignore these methods, because they may call other
1087  // instrumented functions.
1088  if (F.getName().find(" load]") != std::string::npos) {
1089    IRBuilder<> IRB(F.begin()->begin());
1090    IRB.CreateCall(AsanInitFunction);
1091    return true;
1092  }
1093  return false;
1094}
1095
1096bool AddressSanitizer::runOnFunction(Function &F) {
1097  if (BL->isIn(F)) return false;
1098  if (&F == AsanCtorFunction) return false;
1099  if (F.getLinkage() == GlobalValue::AvailableExternallyLinkage) return false;
1100  DEBUG(dbgs() << "ASAN instrumenting:\n" << F << "\n");
1101  initializeCallbacks(*F.getParent());
1102
1103  // If needed, insert __asan_init before checking for SanitizeAddress attr.
1104  maybeInsertAsanInitAtFunctionEntry(F);
1105
1106  if (!F.getAttributes().hasAttribute(AttributeSet::FunctionIndex,
1107                                      Attribute::SanitizeAddress))
1108    return false;
1109
1110  if (!ClDebugFunc.empty() && ClDebugFunc != F.getName())
1111    return false;
1112
1113  // We want to instrument every address only once per basic block (unless there
1114  // are calls between uses).
1115  SmallSet<Value*, 16> TempsToInstrument;
1116  SmallVector<Instruction*, 16> ToInstrument;
1117  SmallVector<Instruction*, 8> NoReturnCalls;
1118  bool IsWrite;
1119
1120  // Fill the set of memory operations to instrument.
1121  for (Function::iterator FI = F.begin(), FE = F.end();
1122       FI != FE; ++FI) {
1123    TempsToInstrument.clear();
1124    int NumInsnsPerBB = 0;
1125    for (BasicBlock::iterator BI = FI->begin(), BE = FI->end();
1126         BI != BE; ++BI) {
1127      if (LooksLikeCodeInBug11395(BI)) return false;
1128      if (Value *Addr = isInterestingMemoryAccess(BI, &IsWrite)) {
1129        if (ClOpt && ClOptSameTemp) {
1130          if (!TempsToInstrument.insert(Addr))
1131            continue;  // We've seen this temp in the current BB.
1132        }
1133      } else if (isa<MemIntrinsic>(BI) && ClMemIntrin) {
1134        // ok, take it.
1135      } else {
1136        CallSite CS(BI);
1137        if (CS) {
1138          // A call inside BB.
1139          TempsToInstrument.clear();
1140          if (CS.doesNotReturn())
1141            NoReturnCalls.push_back(CS.getInstruction());
1142        }
1143        continue;
1144      }
1145      ToInstrument.push_back(BI);
1146      NumInsnsPerBB++;
1147      if (NumInsnsPerBB >= ClMaxInsnsToInstrumentPerBB)
1148        break;
1149    }
1150  }
1151
1152  // Instrument.
1153  int NumInstrumented = 0;
1154  for (size_t i = 0, n = ToInstrument.size(); i != n; i++) {
1155    Instruction *Inst = ToInstrument[i];
1156    if (ClDebugMin < 0 || ClDebugMax < 0 ||
1157        (NumInstrumented >= ClDebugMin && NumInstrumented <= ClDebugMax)) {
1158      if (isInterestingMemoryAccess(Inst, &IsWrite))
1159        instrumentMop(Inst);
1160      else
1161        instrumentMemIntrinsic(cast<MemIntrinsic>(Inst));
1162    }
1163    NumInstrumented++;
1164  }
1165
1166  FunctionStackPoisoner FSP(F, *this);
1167  bool ChangedStack = FSP.runOnFunction();
1168
1169  // We must unpoison the stack before every NoReturn call (throw, _exit, etc).
1170  // See e.g. http://code.google.com/p/address-sanitizer/issues/detail?id=37
1171  for (size_t i = 0, n = NoReturnCalls.size(); i != n; i++) {
1172    Instruction *CI = NoReturnCalls[i];
1173    IRBuilder<> IRB(CI);
1174    IRB.CreateCall(AsanHandleNoReturnFunc);
1175  }
1176  DEBUG(dbgs() << "ASAN done instrumenting:\n" << F << "\n");
1177
1178  return NumInstrumented > 0 || ChangedStack || !NoReturnCalls.empty();
1179}
1180
1181static uint64_t ValueForPoison(uint64_t PoisonByte, size_t ShadowRedzoneSize) {
1182  if (ShadowRedzoneSize == 1) return PoisonByte;
1183  if (ShadowRedzoneSize == 2) return (PoisonByte << 8) + PoisonByte;
1184  if (ShadowRedzoneSize == 4)
1185    return (PoisonByte << 24) + (PoisonByte << 16) +
1186        (PoisonByte << 8) + (PoisonByte);
1187  llvm_unreachable("ShadowRedzoneSize is either 1, 2 or 4");
1188}
1189
1190static void PoisonShadowPartialRightRedzone(uint8_t *Shadow,
1191                                            size_t Size,
1192                                            size_t RZSize,
1193                                            size_t ShadowGranularity,
1194                                            uint8_t Magic) {
1195  for (size_t i = 0; i < RZSize;
1196       i+= ShadowGranularity, Shadow++) {
1197    if (i + ShadowGranularity <= Size) {
1198      *Shadow = 0;  // fully addressable
1199    } else if (i >= Size) {
1200      *Shadow = Magic;  // unaddressable
1201    } else {
1202      *Shadow = Size - i;  // first Size-i bytes are addressable
1203    }
1204  }
1205}
1206
1207// Workaround for bug 11395: we don't want to instrument stack in functions
1208// with large assembly blobs (32-bit only), otherwise reg alloc may crash.
1209// FIXME: remove once the bug 11395 is fixed.
1210bool AddressSanitizer::LooksLikeCodeInBug11395(Instruction *I) {
1211  if (LongSize != 32) return false;
1212  CallInst *CI = dyn_cast<CallInst>(I);
1213  if (!CI || !CI->isInlineAsm()) return false;
1214  if (CI->getNumArgOperands() <= 5) return false;
1215  // We have inline assembly with quite a few arguments.
1216  return true;
1217}
1218
1219void FunctionStackPoisoner::initializeCallbacks(Module &M) {
1220  IRBuilder<> IRB(*C);
1221  AsanStackMallocFunc = checkInterfaceFunction(M.getOrInsertFunction(
1222      kAsanStackMallocName, IntptrTy, IntptrTy, IntptrTy, NULL));
1223  AsanStackFreeFunc = checkInterfaceFunction(M.getOrInsertFunction(
1224      kAsanStackFreeName, IRB.getVoidTy(),
1225      IntptrTy, IntptrTy, IntptrTy, NULL));
1226  AsanPoisonStackMemoryFunc = checkInterfaceFunction(M.getOrInsertFunction(
1227      kAsanPoisonStackMemoryName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
1228  AsanUnpoisonStackMemoryFunc = checkInterfaceFunction(M.getOrInsertFunction(
1229      kAsanUnpoisonStackMemoryName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
1230}
1231
1232void FunctionStackPoisoner::poisonRedZones(
1233  const ArrayRef<AllocaInst*> &AllocaVec, IRBuilder<> IRB, Value *ShadowBase,
1234  bool DoPoison) {
1235  size_t ShadowRZSize = RedzoneSize() >> Mapping.Scale;
1236  assert(ShadowRZSize >= 1 && ShadowRZSize <= 4);
1237  Type *RZTy = Type::getIntNTy(*C, ShadowRZSize * 8);
1238  Type *RZPtrTy = PointerType::get(RZTy, 0);
1239
1240  Value *PoisonLeft  = ConstantInt::get(RZTy,
1241    ValueForPoison(DoPoison ? kAsanStackLeftRedzoneMagic : 0LL, ShadowRZSize));
1242  Value *PoisonMid   = ConstantInt::get(RZTy,
1243    ValueForPoison(DoPoison ? kAsanStackMidRedzoneMagic : 0LL, ShadowRZSize));
1244  Value *PoisonRight = ConstantInt::get(RZTy,
1245    ValueForPoison(DoPoison ? kAsanStackRightRedzoneMagic : 0LL, ShadowRZSize));
1246
1247  // poison the first red zone.
1248  IRB.CreateStore(PoisonLeft, IRB.CreateIntToPtr(ShadowBase, RZPtrTy));
1249
1250  // poison all other red zones.
1251  uint64_t Pos = RedzoneSize();
1252  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
1253    AllocaInst *AI = AllocaVec[i];
1254    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
1255    uint64_t AlignedSize = getAlignedAllocaSize(AI);
1256    assert(AlignedSize - SizeInBytes < RedzoneSize());
1257    Value *Ptr = NULL;
1258
1259    Pos += AlignedSize;
1260
1261    assert(ShadowBase->getType() == IntptrTy);
1262    if (SizeInBytes < AlignedSize) {
1263      // Poison the partial redzone at right
1264      Ptr = IRB.CreateAdd(
1265          ShadowBase, ConstantInt::get(IntptrTy,
1266                                       (Pos >> Mapping.Scale) - ShadowRZSize));
1267      size_t AddressableBytes = RedzoneSize() - (AlignedSize - SizeInBytes);
1268      uint32_t Poison = 0;
1269      if (DoPoison) {
1270        PoisonShadowPartialRightRedzone((uint8_t*)&Poison, AddressableBytes,
1271                                        RedzoneSize(),
1272                                        1ULL << Mapping.Scale,
1273                                        kAsanStackPartialRedzoneMagic);
1274      }
1275      Value *PartialPoison = ConstantInt::get(RZTy, Poison);
1276      IRB.CreateStore(PartialPoison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
1277    }
1278
1279    // Poison the full redzone at right.
1280    Ptr = IRB.CreateAdd(ShadowBase,
1281                        ConstantInt::get(IntptrTy, Pos >> Mapping.Scale));
1282    bool LastAlloca = (i == AllocaVec.size() - 1);
1283    Value *Poison = LastAlloca ? PoisonRight : PoisonMid;
1284    IRB.CreateStore(Poison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
1285
1286    Pos += RedzoneSize();
1287  }
1288}
1289
1290void FunctionStackPoisoner::poisonStack() {
1291  uint64_t LocalStackSize = TotalStackSize +
1292                            (AllocaVec.size() + 1) * RedzoneSize();
1293
1294  bool DoStackMalloc = ASan.CheckUseAfterReturn
1295      && LocalStackSize <= kMaxStackMallocSize;
1296
1297  assert(AllocaVec.size() > 0);
1298  Instruction *InsBefore = AllocaVec[0];
1299  IRBuilder<> IRB(InsBefore);
1300
1301
1302  Type *ByteArrayTy = ArrayType::get(IRB.getInt8Ty(), LocalStackSize);
1303  AllocaInst *MyAlloca =
1304      new AllocaInst(ByteArrayTy, "MyAlloca", InsBefore);
1305  if (ClRealignStack && StackAlignment < RedzoneSize())
1306    StackAlignment = RedzoneSize();
1307  MyAlloca->setAlignment(StackAlignment);
1308  assert(MyAlloca->isStaticAlloca());
1309  Value *OrigStackBase = IRB.CreatePointerCast(MyAlloca, IntptrTy);
1310  Value *LocalStackBase = OrigStackBase;
1311
1312  if (DoStackMalloc) {
1313    LocalStackBase = IRB.CreateCall2(AsanStackMallocFunc,
1314        ConstantInt::get(IntptrTy, LocalStackSize), OrigStackBase);
1315  }
1316
1317  // This string will be parsed by the run-time (DescribeStackAddress).
1318  SmallString<2048> StackDescriptionStorage;
1319  raw_svector_ostream StackDescription(StackDescriptionStorage);
1320  StackDescription << F.getName() << " " << AllocaVec.size() << " ";
1321
1322  // Insert poison calls for lifetime intrinsics for alloca.
1323  bool HavePoisonedAllocas = false;
1324  for (size_t i = 0, n = AllocaPoisonCallVec.size(); i < n; i++) {
1325    const AllocaPoisonCall &APC = AllocaPoisonCallVec[i];
1326    IntrinsicInst *II = APC.InsBefore;
1327    AllocaInst *AI = findAllocaForValue(II->getArgOperand(1));
1328    assert(AI);
1329    IRBuilder<> IRB(II);
1330    poisonAlloca(AI, APC.Size, IRB, APC.DoPoison);
1331    HavePoisonedAllocas |= APC.DoPoison;
1332  }
1333
1334  uint64_t Pos = RedzoneSize();
1335  // Replace Alloca instructions with base+offset.
1336  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
1337    AllocaInst *AI = AllocaVec[i];
1338    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
1339    StringRef Name = AI->getName();
1340    StackDescription << Pos << " " << SizeInBytes << " "
1341                     << Name.size() << " " << Name << " ";
1342    uint64_t AlignedSize = getAlignedAllocaSize(AI);
1343    assert((AlignedSize % RedzoneSize()) == 0);
1344    Value *NewAllocaPtr = IRB.CreateIntToPtr(
1345            IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy, Pos)),
1346            AI->getType());
1347    replaceDbgDeclareForAlloca(AI, NewAllocaPtr, DIB);
1348    AI->replaceAllUsesWith(NewAllocaPtr);
1349    Pos += AlignedSize + RedzoneSize();
1350  }
1351  assert(Pos == LocalStackSize);
1352
1353  // Write the Magic value and the frame description constant to the redzone.
1354  Value *BasePlus0 = IRB.CreateIntToPtr(LocalStackBase, IntptrPtrTy);
1355  IRB.CreateStore(ConstantInt::get(IntptrTy, kCurrentStackFrameMagic),
1356                  BasePlus0);
1357  Value *BasePlus1 = IRB.CreateAdd(LocalStackBase,
1358                                   ConstantInt::get(IntptrTy,
1359                                                    ASan.LongSize/8));
1360  BasePlus1 = IRB.CreateIntToPtr(BasePlus1, IntptrPtrTy);
1361  GlobalVariable *StackDescriptionGlobal =
1362      createPrivateGlobalForString(*F.getParent(), StackDescription.str());
1363  Value *Description = IRB.CreatePointerCast(StackDescriptionGlobal,
1364                                             IntptrTy);
1365  IRB.CreateStore(Description, BasePlus1);
1366
1367  // Poison the stack redzones at the entry.
1368  Value *ShadowBase = ASan.memToShadow(LocalStackBase, IRB);
1369  poisonRedZones(AllocaVec, IRB, ShadowBase, true);
1370
1371  // Unpoison the stack before all ret instructions.
1372  for (size_t i = 0, n = RetVec.size(); i < n; i++) {
1373    Instruction *Ret = RetVec[i];
1374    IRBuilder<> IRBRet(Ret);
1375    // Mark the current frame as retired.
1376    IRBRet.CreateStore(ConstantInt::get(IntptrTy, kRetiredStackFrameMagic),
1377                       BasePlus0);
1378    // Unpoison the stack.
1379    poisonRedZones(AllocaVec, IRBRet, ShadowBase, false);
1380    if (DoStackMalloc) {
1381      // In use-after-return mode, mark the whole stack frame unaddressable.
1382      IRBRet.CreateCall3(AsanStackFreeFunc, LocalStackBase,
1383                         ConstantInt::get(IntptrTy, LocalStackSize),
1384                         OrigStackBase);
1385    } else if (HavePoisonedAllocas) {
1386      // If we poisoned some allocas in llvm.lifetime analysis,
1387      // unpoison whole stack frame now.
1388      assert(LocalStackBase == OrigStackBase);
1389      poisonAlloca(LocalStackBase, LocalStackSize, IRBRet, false);
1390    }
1391  }
1392
1393  // We are done. Remove the old unused alloca instructions.
1394  for (size_t i = 0, n = AllocaVec.size(); i < n; i++)
1395    AllocaVec[i]->eraseFromParent();
1396}
1397
1398void FunctionStackPoisoner::poisonAlloca(Value *V, uint64_t Size,
1399                                         IRBuilder<> IRB, bool DoPoison) {
1400  // For now just insert the call to ASan runtime.
1401  Value *AddrArg = IRB.CreatePointerCast(V, IntptrTy);
1402  Value *SizeArg = ConstantInt::get(IntptrTy, Size);
1403  IRB.CreateCall2(DoPoison ? AsanPoisonStackMemoryFunc
1404                           : AsanUnpoisonStackMemoryFunc,
1405                  AddrArg, SizeArg);
1406}
1407
1408// Handling llvm.lifetime intrinsics for a given %alloca:
1409// (1) collect all llvm.lifetime.xxx(%size, %value) describing the alloca.
1410// (2) if %size is constant, poison memory for llvm.lifetime.end (to detect
1411//     invalid accesses) and unpoison it for llvm.lifetime.start (the memory
1412//     could be poisoned by previous llvm.lifetime.end instruction, as the
1413//     variable may go in and out of scope several times, e.g. in loops).
1414// (3) if we poisoned at least one %alloca in a function,
1415//     unpoison the whole stack frame at function exit.
1416
1417AllocaInst *FunctionStackPoisoner::findAllocaForValue(Value *V) {
1418  if (AllocaInst *AI = dyn_cast<AllocaInst>(V))
1419    // We're intested only in allocas we can handle.
1420    return isInterestingAlloca(*AI) ? AI : 0;
1421  // See if we've already calculated (or started to calculate) alloca for a
1422  // given value.
1423  AllocaForValueMapTy::iterator I = AllocaForValue.find(V);
1424  if (I != AllocaForValue.end())
1425    return I->second;
1426  // Store 0 while we're calculating alloca for value V to avoid
1427  // infinite recursion if the value references itself.
1428  AllocaForValue[V] = 0;
1429  AllocaInst *Res = 0;
1430  if (CastInst *CI = dyn_cast<CastInst>(V))
1431    Res = findAllocaForValue(CI->getOperand(0));
1432  else if (PHINode *PN = dyn_cast<PHINode>(V)) {
1433    for (unsigned i = 0, e = PN->getNumIncomingValues(); i != e; ++i) {
1434      Value *IncValue = PN->getIncomingValue(i);
1435      // Allow self-referencing phi-nodes.
1436      if (IncValue == PN) continue;
1437      AllocaInst *IncValueAI = findAllocaForValue(IncValue);
1438      // AI for incoming values should exist and should all be equal.
1439      if (IncValueAI == 0 || (Res != 0 && IncValueAI != Res))
1440        return 0;
1441      Res = IncValueAI;
1442    }
1443  }
1444  if (Res != 0)
1445    AllocaForValue[V] = Res;
1446  return Res;
1447}
1448