AddressSanitizer.cpp revision 263508 
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/Statistic.h"
27#include "llvm/ADT/StringExtras.h"
28#include "llvm/ADT/Triple.h"
29#include "llvm/DIBuilder.h"
30#include "llvm/IR/DataLayout.h"
31#include "llvm/IR/Function.h"
32#include "llvm/IR/IRBuilder.h"
33#include "llvm/IR/InlineAsm.h"
34#include "llvm/IR/IntrinsicInst.h"
35#include "llvm/IR/LLVMContext.h"
36#include "llvm/IR/Module.h"
37#include "llvm/IR/Type.h"
38#include "llvm/InstVisitor.h"
39#include "llvm/Support/CallSite.h"
40#include "llvm/Support/CommandLine.h"
41#include "llvm/Support/DataTypes.h"
42#include "llvm/Support/Debug.h"
43#include "llvm/Support/Endian.h"
44#include "llvm/Support/raw_ostream.h"
45#include "llvm/Support/system_error.h"
46#include "llvm/Transforms/Utils/BasicBlockUtils.h"
47#include "llvm/Transforms/Utils/Cloning.h"
48#include "llvm/Transforms/Utils/Local.h"
49#include "llvm/Transforms/Utils/ModuleUtils.h"
50#include "llvm/Transforms/Utils/SpecialCaseList.h"
51#include <algorithm>
52#include <string>
53
54using namespace llvm;
55
56static const uint64_t kDefaultShadowScale = 3;
57static const uint64_t kDefaultShadowOffset32 = 1ULL << 29;
58static const uint64_t kDefaultShadowOffset64 = 1ULL << 44;
59static const uint64_t kDefaultShort64bitShadowOffset = 0x7FFF8000;  // < 2G.
60static const uint64_t kPPC64_ShadowOffset64 = 1ULL << 41;
61static const uint64_t kMIPS32_ShadowOffset32 = 0x0aaa8000;
62
63static const size_t kMinStackMallocSize = 1 << 6;  // 64B
64static const size_t kMaxStackMallocSize = 1 << 16;  // 64K
65static const uintptr_t kCurrentStackFrameMagic = 0x41B58AB3;
66static const uintptr_t kRetiredStackFrameMagic = 0x45E0360E;
67
68static const char *const kAsanModuleCtorName = "asan.module_ctor";
69static const char *const kAsanModuleDtorName = "asan.module_dtor";
70static const int         kAsanCtorAndCtorPriority = 1;
71static const char *const kAsanReportErrorTemplate = "__asan_report_";
72static const char *const kAsanReportLoadN = "__asan_report_load_n";
73static const char *const kAsanReportStoreN = "__asan_report_store_n";
74static const char *const kAsanRegisterGlobalsName = "__asan_register_globals";
75static const char *const kAsanUnregisterGlobalsName =
76    "__asan_unregister_globals";
77static const char *const kAsanPoisonGlobalsName = "__asan_before_dynamic_init";
78static const char *const kAsanUnpoisonGlobalsName = "__asan_after_dynamic_init";
79static const char *const kAsanInitName = "__asan_init_v3";
80static const char *const kAsanCovName = "__sanitizer_cov";
81static const char *const kAsanHandleNoReturnName = "__asan_handle_no_return";
82static const char *const kAsanMappingOffsetName = "__asan_mapping_offset";
83static const char *const kAsanMappingScaleName = "__asan_mapping_scale";
84static const int         kMaxAsanStackMallocSizeClass = 10;
85static const char *const kAsanStackMallocNameTemplate = "__asan_stack_malloc_";
86static const char *const kAsanStackFreeNameTemplate = "__asan_stack_free_";
87static const char *const kAsanGenPrefix = "__asan_gen_";
88static const char *const kAsanPoisonStackMemoryName =
89    "__asan_poison_stack_memory";
90static const char *const kAsanUnpoisonStackMemoryName =
91    "__asan_unpoison_stack_memory";
92
93static const char *const kAsanOptionDetectUAR =
94    "__asan_option_detect_stack_use_after_return";
95
96// These constants must match the definitions in the run-time library.
97static const int kAsanStackLeftRedzoneMagic = 0xf1;
98static const int kAsanStackMidRedzoneMagic = 0xf2;
99static const int kAsanStackRightRedzoneMagic = 0xf3;
100static const int kAsanStackPartialRedzoneMagic = 0xf4;
101#ifndef NDEBUG
102static const int kAsanStackAfterReturnMagic = 0xf5;
103#endif
104
105// Accesses sizes are powers of two: 1, 2, 4, 8, 16.
106static const size_t kNumberOfAccessSizes = 5;
107
108// Command-line flags.
109
110// This flag may need to be replaced with -f[no-]asan-reads.
111static cl::opt<bool> ClInstrumentReads("asan-instrument-reads",
112       cl::desc("instrument read instructions"), cl::Hidden, cl::init(true));
113static cl::opt<bool> ClInstrumentWrites("asan-instrument-writes",
114       cl::desc("instrument write instructions"), cl::Hidden, cl::init(true));
115static cl::opt<bool> ClInstrumentAtomics("asan-instrument-atomics",
116       cl::desc("instrument atomic instructions (rmw, cmpxchg)"),
117       cl::Hidden, cl::init(true));
118static cl::opt<bool> ClAlwaysSlowPath("asan-always-slow-path",
119       cl::desc("use instrumentation with slow path for all accesses"),
120       cl::Hidden, cl::init(false));
121// This flag limits the number of instructions to be instrumented
122// in any given BB. Normally, this should be set to unlimited (INT_MAX),
123// but due to http://llvm.org/bugs/show_bug.cgi?id=12652 we temporary
124// set it to 10000.
125static cl::opt<int> ClMaxInsnsToInstrumentPerBB("asan-max-ins-per-bb",
126       cl::init(10000),
127       cl::desc("maximal number of instructions to instrument in any given BB"),
128       cl::Hidden);
129// This flag may need to be replaced with -f[no]asan-stack.
130static cl::opt<bool> ClStack("asan-stack",
131       cl::desc("Handle stack memory"), cl::Hidden, cl::init(true));
132// This flag may need to be replaced with -f[no]asan-use-after-return.
133static cl::opt<bool> ClUseAfterReturn("asan-use-after-return",
134       cl::desc("Check return-after-free"), cl::Hidden, cl::init(false));
135// This flag may need to be replaced with -f[no]asan-globals.
136static cl::opt<bool> ClGlobals("asan-globals",
137       cl::desc("Handle global objects"), cl::Hidden, cl::init(true));
138static cl::opt<bool> ClCoverage("asan-coverage",
139       cl::desc("ASan coverage"), cl::Hidden, cl::init(false));
140static cl::opt<bool> ClInitializers("asan-initialization-order",
141       cl::desc("Handle C++ initializer order"), cl::Hidden, cl::init(false));
142static cl::opt<bool> ClMemIntrin("asan-memintrin",
143       cl::desc("Handle memset/memcpy/memmove"), cl::Hidden, cl::init(true));
144static cl::opt<bool> ClRealignStack("asan-realign-stack",
145       cl::desc("Realign stack to 32"), cl::Hidden, cl::init(true));
146static cl::opt<std::string> ClBlacklistFile("asan-blacklist",
147       cl::desc("File containing the list of objects to ignore "
148                "during instrumentation"), cl::Hidden);
149
150// This is an experimental feature that will allow to choose between
151// instrumented and non-instrumented code at link-time.
152// If this option is on, just before instrumenting a function we create its
153// clone; if the function is not changed by asan the clone is deleted.
154// If we end up with a clone, we put the instrumented function into a section
155// called "ASAN" and the uninstrumented function into a section called "NOASAN".
156//
157// This is still a prototype, we need to figure out a way to keep two copies of
158// a function so that the linker can easily choose one of them.
159static cl::opt<bool> ClKeepUninstrumented("asan-keep-uninstrumented-functions",
160       cl::desc("Keep uninstrumented copies of functions"),
161       cl::Hidden, cl::init(false));
162
163// These flags allow to change the shadow mapping.
164// The shadow mapping looks like
165//    Shadow = (Mem >> scale) + (1 << offset_log)
166static cl::opt<int> ClMappingScale("asan-mapping-scale",
167       cl::desc("scale of asan shadow mapping"), cl::Hidden, cl::init(0));
168static cl::opt<int> ClMappingOffsetLog("asan-mapping-offset-log",
169       cl::desc("offset of asan shadow mapping"), cl::Hidden, cl::init(-1));
170static cl::opt<bool> ClShort64BitOffset("asan-short-64bit-mapping-offset",
171       cl::desc("Use short immediate constant as the mapping offset for 64bit"),
172       cl::Hidden, cl::init(true));
173
174// Optimization flags. Not user visible, used mostly for testing
175// and benchmarking the tool.
176static cl::opt<bool> ClOpt("asan-opt",
177       cl::desc("Optimize instrumentation"), cl::Hidden, cl::init(true));
178static cl::opt<bool> ClOptSameTemp("asan-opt-same-temp",
179       cl::desc("Instrument the same temp just once"), cl::Hidden,
180       cl::init(true));
181static cl::opt<bool> ClOptGlobals("asan-opt-globals",
182       cl::desc("Don't instrument scalar globals"), cl::Hidden, cl::init(true));
183
184static cl::opt<bool> ClCheckLifetime("asan-check-lifetime",
185       cl::desc("Use llvm.lifetime intrinsics to insert extra checks"),
186       cl::Hidden, cl::init(false));
187
188// Debug flags.
189static cl::opt<int> ClDebug("asan-debug", cl::desc("debug"), cl::Hidden,
190                            cl::init(0));
191static cl::opt<int> ClDebugStack("asan-debug-stack", cl::desc("debug stack"),
192                                 cl::Hidden, cl::init(0));
193static cl::opt<std::string> ClDebugFunc("asan-debug-func",
194                                        cl::Hidden, cl::desc("Debug func"));
195static cl::opt<int> ClDebugMin("asan-debug-min", cl::desc("Debug min inst"),
196                               cl::Hidden, cl::init(-1));
197static cl::opt<int> ClDebugMax("asan-debug-max", cl::desc("Debug man inst"),
198                               cl::Hidden, cl::init(-1));
199
200STATISTIC(NumInstrumentedReads, "Number of instrumented reads");
201STATISTIC(NumInstrumentedWrites, "Number of instrumented writes");
202STATISTIC(NumOptimizedAccessesToGlobalArray,
203          "Number of optimized accesses to global arrays");
204STATISTIC(NumOptimizedAccessesToGlobalVar,
205          "Number of optimized accesses to global vars");
206
207namespace {
208/// A set of dynamically initialized globals extracted from metadata.
209class SetOfDynamicallyInitializedGlobals {
210 public:
211  void Init(Module& M) {
212    // Clang generates metadata identifying all dynamically initialized globals.
213    NamedMDNode *DynamicGlobals =
214        M.getNamedMetadata("llvm.asan.dynamically_initialized_globals");
215    if (!DynamicGlobals)
216      return;
217    for (int i = 0, n = DynamicGlobals->getNumOperands(); i < n; ++i) {
218      MDNode *MDN = DynamicGlobals->getOperand(i);
219      assert(MDN->getNumOperands() == 1);
220      Value *VG = MDN->getOperand(0);
221      // The optimizer may optimize away a global entirely, in which case we
222      // cannot instrument access to it.
223      if (!VG)
224        continue;
225      DynInitGlobals.insert(cast<GlobalVariable>(VG));
226    }
227  }
228  bool Contains(GlobalVariable *G) { return DynInitGlobals.count(G) != 0; }
229 private:
230  SmallSet<GlobalValue*, 32> DynInitGlobals;
231};
232
233/// This struct defines the shadow mapping using the rule:
234///   shadow = (mem >> Scale) ADD-or-OR Offset.
235struct ShadowMapping {
236  int Scale;
237  uint64_t Offset;
238  bool OrShadowOffset;
239};
240
241static ShadowMapping getShadowMapping(const Module &M, int LongSize,
242                                      bool ZeroBaseShadow) {
243  llvm::Triple TargetTriple(M.getTargetTriple());
244  bool IsAndroid = TargetTriple.getEnvironment() == llvm::Triple::Android;
245  bool IsMacOSX = TargetTriple.getOS() == llvm::Triple::MacOSX;
246  bool IsPPC64 = TargetTriple.getArch() == llvm::Triple::ppc64 ||
247                 TargetTriple.getArch() == llvm::Triple::ppc64le;
248  bool IsX86_64 = TargetTriple.getArch() == llvm::Triple::x86_64;
249  bool IsMIPS32 = TargetTriple.getArch() == llvm::Triple::mips ||
250                  TargetTriple.getArch() == llvm::Triple::mipsel;
251
252  ShadowMapping Mapping;
253
254  // OR-ing shadow offset if more efficient (at least on x86),
255  // but on ppc64 we have to use add since the shadow offset is not neccesary
256  // 1/8-th of the address space.
257  Mapping.OrShadowOffset = !IsPPC64 && !ClShort64BitOffset;
258
259  Mapping.Offset = (IsAndroid || ZeroBaseShadow) ? 0 :
260      (LongSize == 32 ?
261       (IsMIPS32 ? kMIPS32_ShadowOffset32 : kDefaultShadowOffset32) :
262       IsPPC64 ? kPPC64_ShadowOffset64 : kDefaultShadowOffset64);
263  if (!ZeroBaseShadow && ClShort64BitOffset && IsX86_64 && !IsMacOSX) {
264    assert(LongSize == 64);
265    Mapping.Offset = kDefaultShort64bitShadowOffset;
266  }
267  if (!ZeroBaseShadow && ClMappingOffsetLog >= 0) {
268    // Zero offset log is the special case.
269    Mapping.Offset = (ClMappingOffsetLog == 0) ? 0 : 1ULL << ClMappingOffsetLog;
270  }
271
272  Mapping.Scale = kDefaultShadowScale;
273  if (ClMappingScale) {
274    Mapping.Scale = ClMappingScale;
275  }
276
277  return Mapping;
278}
279
280static size_t RedzoneSizeForScale(int MappingScale) {
281  // Redzone used for stack and globals is at least 32 bytes.
282  // For scales 6 and 7, the redzone has to be 64 and 128 bytes respectively.
283  return std::max(32U, 1U << MappingScale);
284}
285
286/// AddressSanitizer: instrument the code in module to find memory bugs.
287struct AddressSanitizer : public FunctionPass {
288  AddressSanitizer(bool CheckInitOrder = true,
289                   bool CheckUseAfterReturn = false,
290                   bool CheckLifetime = false,
291                   StringRef BlacklistFile = StringRef(),
292                   bool ZeroBaseShadow = false)
293      : FunctionPass(ID),
294        CheckInitOrder(CheckInitOrder || ClInitializers),
295        CheckUseAfterReturn(CheckUseAfterReturn || ClUseAfterReturn),
296        CheckLifetime(CheckLifetime || ClCheckLifetime),
297        BlacklistFile(BlacklistFile.empty() ? ClBlacklistFile
298                                            : BlacklistFile),
299        ZeroBaseShadow(ZeroBaseShadow) {}
300  virtual const char *getPassName() const {
301    return "AddressSanitizerFunctionPass";
302  }
303  void instrumentMop(Instruction *I);
304  void instrumentAddress(Instruction *OrigIns, Instruction *InsertBefore,
305                         Value *Addr, uint32_t TypeSize, bool IsWrite,
306                         Value *SizeArgument);
307  Value *createSlowPathCmp(IRBuilder<> &IRB, Value *AddrLong,
308                           Value *ShadowValue, uint32_t TypeSize);
309  Instruction *generateCrashCode(Instruction *InsertBefore, Value *Addr,
310                                 bool IsWrite, size_t AccessSizeIndex,
311                                 Value *SizeArgument);
312  bool instrumentMemIntrinsic(MemIntrinsic *MI);
313  void instrumentMemIntrinsicParam(Instruction *OrigIns, Value *Addr,
314                                   Value *Size,
315                                   Instruction *InsertBefore, bool IsWrite);
316  Value *memToShadow(Value *Shadow, IRBuilder<> &IRB);
317  bool runOnFunction(Function &F);
318  bool maybeInsertAsanInitAtFunctionEntry(Function &F);
319  void emitShadowMapping(Module &M, IRBuilder<> &IRB) const;
320  virtual bool doInitialization(Module &M);
321  static char ID;  // Pass identification, replacement for typeid
322
323 private:
324  void initializeCallbacks(Module &M);
325
326  bool ShouldInstrumentGlobal(GlobalVariable *G);
327  bool LooksLikeCodeInBug11395(Instruction *I);
328  void FindDynamicInitializers(Module &M);
329  bool GlobalIsLinkerInitialized(GlobalVariable *G);
330  bool InjectCoverage(Function &F);
331
332  bool CheckInitOrder;
333  bool CheckUseAfterReturn;
334  bool CheckLifetime;
335  SmallString<64> BlacklistFile;
336  bool ZeroBaseShadow;
337
338  LLVMContext *C;
339  DataLayout *TD;
340  int LongSize;
341  Type *IntptrTy;
342  ShadowMapping Mapping;
343  Function *AsanCtorFunction;
344  Function *AsanInitFunction;
345  Function *AsanHandleNoReturnFunc;
346  Function *AsanCovFunction;
347  OwningPtr<SpecialCaseList> BL;
348  // This array is indexed by AccessIsWrite and log2(AccessSize).
349  Function *AsanErrorCallback[2][kNumberOfAccessSizes];
350  // This array is indexed by AccessIsWrite.
351  Function *AsanErrorCallbackSized[2];
352  InlineAsm *EmptyAsm;
353  SetOfDynamicallyInitializedGlobals DynamicallyInitializedGlobals;
354
355  friend struct FunctionStackPoisoner;
356};
357
358class AddressSanitizerModule : public ModulePass {
359 public:
360  AddressSanitizerModule(bool CheckInitOrder = true,
361                         StringRef BlacklistFile = StringRef(),
362                         bool ZeroBaseShadow = false)
363      : ModulePass(ID),
364        CheckInitOrder(CheckInitOrder || ClInitializers),
365        BlacklistFile(BlacklistFile.empty() ? ClBlacklistFile
366                                            : BlacklistFile),
367        ZeroBaseShadow(ZeroBaseShadow) {}
368  bool runOnModule(Module &M);
369  static char ID;  // Pass identification, replacement for typeid
370  virtual const char *getPassName() const {
371    return "AddressSanitizerModule";
372  }
373
374 private:
375  void initializeCallbacks(Module &M);
376
377  bool ShouldInstrumentGlobal(GlobalVariable *G);
378  void createInitializerPoisonCalls(Module &M, GlobalValue *ModuleName);
379  size_t RedzoneSize() const {
380    return RedzoneSizeForScale(Mapping.Scale);
381  }
382
383  bool CheckInitOrder;
384  SmallString<64> BlacklistFile;
385  bool ZeroBaseShadow;
386
387  OwningPtr<SpecialCaseList> BL;
388  SetOfDynamicallyInitializedGlobals DynamicallyInitializedGlobals;
389  Type *IntptrTy;
390  LLVMContext *C;
391  DataLayout *TD;
392  ShadowMapping Mapping;
393  Function *AsanPoisonGlobals;
394  Function *AsanUnpoisonGlobals;
395  Function *AsanRegisterGlobals;
396  Function *AsanUnregisterGlobals;
397};
398
399// Stack poisoning does not play well with exception handling.
400// When an exception is thrown, we essentially bypass the code
401// that unpoisones the stack. This is why the run-time library has
402// to intercept __cxa_throw (as well as longjmp, etc) and unpoison the entire
403// stack in the interceptor. This however does not work inside the
404// actual function which catches the exception. Most likely because the
405// compiler hoists the load of the shadow value somewhere too high.
406// This causes asan to report a non-existing bug on 453.povray.
407// It sounds like an LLVM bug.
408struct FunctionStackPoisoner : public InstVisitor<FunctionStackPoisoner> {
409  Function &F;
410  AddressSanitizer &ASan;
411  DIBuilder DIB;
412  LLVMContext *C;
413  Type *IntptrTy;
414  Type *IntptrPtrTy;
415  ShadowMapping Mapping;
416
417  SmallVector<AllocaInst*, 16> AllocaVec;
418  SmallVector<Instruction*, 8> RetVec;
419  uint64_t TotalStackSize;
420  unsigned StackAlignment;
421
422  Function *AsanStackMallocFunc[kMaxAsanStackMallocSizeClass + 1],
423           *AsanStackFreeFunc[kMaxAsanStackMallocSizeClass + 1];
424  Function *AsanPoisonStackMemoryFunc, *AsanUnpoisonStackMemoryFunc;
425
426  // Stores a place and arguments of poisoning/unpoisoning call for alloca.
427  struct AllocaPoisonCall {
428    IntrinsicInst *InsBefore;
429    AllocaInst *AI;
430    uint64_t Size;
431    bool DoPoison;
432  };
433  SmallVector<AllocaPoisonCall, 8> AllocaPoisonCallVec;
434
435  // Maps Value to an AllocaInst from which the Value is originated.
436  typedef DenseMap<Value*, AllocaInst*> AllocaForValueMapTy;
437  AllocaForValueMapTy AllocaForValue;
438
439  FunctionStackPoisoner(Function &F, AddressSanitizer &ASan)
440      : F(F), ASan(ASan), DIB(*F.getParent()), C(ASan.C),
441        IntptrTy(ASan.IntptrTy), IntptrPtrTy(PointerType::get(IntptrTy, 0)),
442        Mapping(ASan.Mapping),
443        TotalStackSize(0), StackAlignment(1 << Mapping.Scale) {}
444
445  bool runOnFunction() {
446    if (!ClStack) return false;
447    // Collect alloca, ret, lifetime instructions etc.
448    for (df_iterator<BasicBlock*> DI = df_begin(&F.getEntryBlock()),
449         DE = df_end(&F.getEntryBlock()); DI != DE; ++DI) {
450      BasicBlock *BB = *DI;
451      visit(*BB);
452    }
453    if (AllocaVec.empty()) return false;
454
455    initializeCallbacks(*F.getParent());
456
457    poisonStack();
458
459    if (ClDebugStack) {
460      DEBUG(dbgs() << F);
461    }
462    return true;
463  }
464
465  // Finds all static Alloca instructions and puts
466  // poisoned red zones around all of them.
467  // Then unpoison everything back before the function returns.
468  void poisonStack();
469
470  // ----------------------- Visitors.
471  /// \brief Collect all Ret instructions.
472  void visitReturnInst(ReturnInst &RI) {
473    RetVec.push_back(&RI);
474  }
475
476  /// \brief Collect Alloca instructions we want (and can) handle.
477  void visitAllocaInst(AllocaInst &AI) {
478    if (!isInterestingAlloca(AI)) return;
479
480    StackAlignment = std::max(StackAlignment, AI.getAlignment());
481    AllocaVec.push_back(&AI);
482    uint64_t AlignedSize = getAlignedAllocaSize(&AI);
483    TotalStackSize += AlignedSize;
484  }
485
486  /// \brief Collect lifetime intrinsic calls to check for use-after-scope
487  /// errors.
488  void visitIntrinsicInst(IntrinsicInst &II) {
489    if (!ASan.CheckLifetime) return;
490    Intrinsic::ID ID = II.getIntrinsicID();
491    if (ID != Intrinsic::lifetime_start &&
492        ID != Intrinsic::lifetime_end)
493      return;
494    // Found lifetime intrinsic, add ASan instrumentation if necessary.
495    ConstantInt *Size = dyn_cast<ConstantInt>(II.getArgOperand(0));
496    // If size argument is undefined, don't do anything.
497    if (Size->isMinusOne()) return;
498    // Check that size doesn't saturate uint64_t and can
499    // be stored in IntptrTy.
500    const uint64_t SizeValue = Size->getValue().getLimitedValue();
501    if (SizeValue == ~0ULL ||
502        !ConstantInt::isValueValidForType(IntptrTy, SizeValue))
503      return;
504    // Find alloca instruction that corresponds to llvm.lifetime argument.
505    AllocaInst *AI = findAllocaForValue(II.getArgOperand(1));
506    if (!AI) return;
507    bool DoPoison = (ID == Intrinsic::lifetime_end);
508    AllocaPoisonCall APC = {&II, AI, SizeValue, DoPoison};
509    AllocaPoisonCallVec.push_back(APC);
510  }
511
512  // ---------------------- Helpers.
513  void initializeCallbacks(Module &M);
514
515  // Check if we want (and can) handle this alloca.
516  bool isInterestingAlloca(AllocaInst &AI) const {
517    return (!AI.isArrayAllocation() &&
518            AI.isStaticAlloca() &&
519            AI.getAlignment() <= RedzoneSize() &&
520            AI.getAllocatedType()->isSized());
521  }
522
523  size_t RedzoneSize() const {
524    return RedzoneSizeForScale(Mapping.Scale);
525  }
526  uint64_t getAllocaSizeInBytes(AllocaInst *AI) const {
527    Type *Ty = AI->getAllocatedType();
528    uint64_t SizeInBytes = ASan.TD->getTypeAllocSize(Ty);
529    return SizeInBytes;
530  }
531  uint64_t getAlignedSize(uint64_t SizeInBytes) const {
532    size_t RZ = RedzoneSize();
533    return ((SizeInBytes + RZ - 1) / RZ) * RZ;
534  }
535  uint64_t getAlignedAllocaSize(AllocaInst *AI) const {
536    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
537    return getAlignedSize(SizeInBytes);
538  }
539  /// Finds alloca where the value comes from.
540  AllocaInst *findAllocaForValue(Value *V);
541  void poisonRedZones(const ArrayRef<AllocaInst*> &AllocaVec, IRBuilder<> &IRB,
542                      Value *ShadowBase, bool DoPoison);
543  void poisonAlloca(Value *V, uint64_t Size, IRBuilder<> &IRB, bool DoPoison);
544
545  void SetShadowToStackAfterReturnInlined(IRBuilder<> &IRB, Value *ShadowBase,
546                                          int Size);
547};
548
549}  // namespace
550
551char AddressSanitizer::ID = 0;
552INITIALIZE_PASS(AddressSanitizer, "asan",
553    "AddressSanitizer: detects use-after-free and out-of-bounds bugs.",
554    false, false)
555FunctionPass *llvm::createAddressSanitizerFunctionPass(
556    bool CheckInitOrder, bool CheckUseAfterReturn, bool CheckLifetime,
557    StringRef BlacklistFile, bool ZeroBaseShadow) {
558  return new AddressSanitizer(CheckInitOrder, CheckUseAfterReturn,
559                              CheckLifetime, BlacklistFile, ZeroBaseShadow);
560}
561
562char AddressSanitizerModule::ID = 0;
563INITIALIZE_PASS(AddressSanitizerModule, "asan-module",
564    "AddressSanitizer: detects use-after-free and out-of-bounds bugs."
565    "ModulePass", false, false)
566ModulePass *llvm::createAddressSanitizerModulePass(
567    bool CheckInitOrder, StringRef BlacklistFile, bool ZeroBaseShadow) {
568  return new AddressSanitizerModule(CheckInitOrder, BlacklistFile,
569                                    ZeroBaseShadow);
570}
571
572static size_t TypeSizeToSizeIndex(uint32_t TypeSize) {
573  size_t Res = countTrailingZeros(TypeSize / 8);
574  assert(Res < kNumberOfAccessSizes);
575  return Res;
576}
577
578// \brief Create a constant for Str so that we can pass it to the run-time lib.
579static GlobalVariable *createPrivateGlobalForString(Module &M, StringRef Str) {
580  Constant *StrConst = ConstantDataArray::getString(M.getContext(), Str);
581  GlobalVariable *GV = new GlobalVariable(M, StrConst->getType(), true,
582                            GlobalValue::InternalLinkage, StrConst,
583                            kAsanGenPrefix);
584  GV->setUnnamedAddr(true);  // Ok to merge these.
585  GV->setAlignment(1);  // Strings may not be merged w/o setting align 1.
586  return GV;
587}
588
589static bool GlobalWasGeneratedByAsan(GlobalVariable *G) {
590  return G->getName().find(kAsanGenPrefix) == 0;
591}
592
593Value *AddressSanitizer::memToShadow(Value *Shadow, IRBuilder<> &IRB) {
594  // Shadow >> scale
595  Shadow = IRB.CreateLShr(Shadow, Mapping.Scale);
596  if (Mapping.Offset == 0)
597    return Shadow;
598  // (Shadow >> scale) | offset
599  if (Mapping.OrShadowOffset)
600    return IRB.CreateOr(Shadow, ConstantInt::get(IntptrTy, Mapping.Offset));
601  else
602    return IRB.CreateAdd(Shadow, ConstantInt::get(IntptrTy, Mapping.Offset));
603}
604
605void AddressSanitizer::instrumentMemIntrinsicParam(
606    Instruction *OrigIns,
607    Value *Addr, Value *Size, Instruction *InsertBefore, bool IsWrite) {
608  IRBuilder<> IRB(InsertBefore);
609  if (Size->getType() != IntptrTy)
610    Size = IRB.CreateIntCast(Size, IntptrTy, false);
611  // Check the first byte.
612  instrumentAddress(OrigIns, InsertBefore, Addr, 8, IsWrite, Size);
613  // Check the last byte.
614  IRB.SetInsertPoint(InsertBefore);
615  Value *SizeMinusOne = IRB.CreateSub(Size, ConstantInt::get(IntptrTy, 1));
616  Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
617  Value *AddrLast = IRB.CreateAdd(AddrLong, SizeMinusOne);
618  instrumentAddress(OrigIns, InsertBefore, AddrLast, 8, IsWrite, Size);
619}
620
621// Instrument memset/memmove/memcpy
622bool AddressSanitizer::instrumentMemIntrinsic(MemIntrinsic *MI) {
623  Value *Dst = MI->getDest();
624  MemTransferInst *MemTran = dyn_cast<MemTransferInst>(MI);
625  Value *Src = MemTran ? MemTran->getSource() : 0;
626  Value *Length = MI->getLength();
627
628  Constant *ConstLength = dyn_cast<Constant>(Length);
629  Instruction *InsertBefore = MI;
630  if (ConstLength) {
631    if (ConstLength->isNullValue()) return false;
632  } else {
633    // The size is not a constant so it could be zero -- check at run-time.
634    IRBuilder<> IRB(InsertBefore);
635
636    Value *Cmp = IRB.CreateICmpNE(Length,
637                                  Constant::getNullValue(Length->getType()));
638    InsertBefore = SplitBlockAndInsertIfThen(cast<Instruction>(Cmp), false);
639  }
640
641  instrumentMemIntrinsicParam(MI, Dst, Length, InsertBefore, true);
642  if (Src)
643    instrumentMemIntrinsicParam(MI, Src, Length, InsertBefore, false);
644  return true;
645}
646
647// If I is an interesting memory access, return the PointerOperand
648// and set IsWrite. Otherwise return NULL.
649static Value *isInterestingMemoryAccess(Instruction *I, bool *IsWrite) {
650  if (LoadInst *LI = dyn_cast<LoadInst>(I)) {
651    if (!ClInstrumentReads) return NULL;
652    *IsWrite = false;
653    return LI->getPointerOperand();
654  }
655  if (StoreInst *SI = dyn_cast<StoreInst>(I)) {
656    if (!ClInstrumentWrites) return NULL;
657    *IsWrite = true;
658    return SI->getPointerOperand();
659  }
660  if (AtomicRMWInst *RMW = dyn_cast<AtomicRMWInst>(I)) {
661    if (!ClInstrumentAtomics) return NULL;
662    *IsWrite = true;
663    return RMW->getPointerOperand();
664  }
665  if (AtomicCmpXchgInst *XCHG = dyn_cast<AtomicCmpXchgInst>(I)) {
666    if (!ClInstrumentAtomics) return NULL;
667    *IsWrite = true;
668    return XCHG->getPointerOperand();
669  }
670  return NULL;
671}
672
673bool AddressSanitizer::GlobalIsLinkerInitialized(GlobalVariable *G) {
674  // If a global variable does not have dynamic initialization we don't
675  // have to instrument it.  However, if a global does not have initializer
676  // at all, we assume it has dynamic initializer (in other TU).
677  return G->hasInitializer() && !DynamicallyInitializedGlobals.Contains(G);
678}
679
680void AddressSanitizer::instrumentMop(Instruction *I) {
681  bool IsWrite = false;
682  Value *Addr = isInterestingMemoryAccess(I, &IsWrite);
683  assert(Addr);
684  if (ClOpt && ClOptGlobals) {
685    if (GlobalVariable *G = dyn_cast<GlobalVariable>(Addr)) {
686      // If initialization order checking is disabled, a simple access to a
687      // dynamically initialized global is always valid.
688      if (!CheckInitOrder || GlobalIsLinkerInitialized(G)) {
689        NumOptimizedAccessesToGlobalVar++;
690        return;
691      }
692    }
693    ConstantExpr *CE = dyn_cast<ConstantExpr>(Addr);
694    if (CE && CE->isGEPWithNoNotionalOverIndexing()) {
695      if (GlobalVariable *G = dyn_cast<GlobalVariable>(CE->getOperand(0))) {
696        if (CE->getOperand(1)->isNullValue() && GlobalIsLinkerInitialized(G)) {
697          NumOptimizedAccessesToGlobalArray++;
698          return;
699        }
700      }
701    }
702  }
703
704  Type *OrigPtrTy = Addr->getType();
705  Type *OrigTy = cast<PointerType>(OrigPtrTy)->getElementType();
706
707  assert(OrigTy->isSized());
708  uint32_t TypeSize = TD->getTypeStoreSizeInBits(OrigTy);
709
710  assert((TypeSize % 8) == 0);
711
712  if (IsWrite)
713    NumInstrumentedWrites++;
714  else
715    NumInstrumentedReads++;
716
717  // Instrument a 1-, 2-, 4-, 8-, or 16- byte access with one check.
718  if (TypeSize == 8  || TypeSize == 16 ||
719      TypeSize == 32 || TypeSize == 64 || TypeSize == 128)
720    return instrumentAddress(I, I, Addr, TypeSize, IsWrite, 0);
721  // Instrument unusual size (but still multiple of 8).
722  // We can not do it with a single check, so we do 1-byte check for the first
723  // and the last bytes. We call __asan_report_*_n(addr, real_size) to be able
724  // to report the actual access size.
725  IRBuilder<> IRB(I);
726  Value *LastByte =  IRB.CreateIntToPtr(
727      IRB.CreateAdd(IRB.CreatePointerCast(Addr, IntptrTy),
728                    ConstantInt::get(IntptrTy, TypeSize / 8 - 1)),
729      OrigPtrTy);
730  Value *Size = ConstantInt::get(IntptrTy, TypeSize / 8);
731  instrumentAddress(I, I, Addr, 8, IsWrite, Size);
732  instrumentAddress(I, I, LastByte, 8, IsWrite, Size);
733}
734
735// Validate the result of Module::getOrInsertFunction called for an interface
736// function of AddressSanitizer. If the instrumented module defines a function
737// with the same name, their prototypes must match, otherwise
738// getOrInsertFunction returns a bitcast.
739static Function *checkInterfaceFunction(Constant *FuncOrBitcast) {
740  if (isa<Function>(FuncOrBitcast)) return cast<Function>(FuncOrBitcast);
741  FuncOrBitcast->dump();
742  report_fatal_error("trying to redefine an AddressSanitizer "
743                     "interface function");
744}
745
746Instruction *AddressSanitizer::generateCrashCode(
747    Instruction *InsertBefore, Value *Addr,
748    bool IsWrite, size_t AccessSizeIndex, Value *SizeArgument) {
749  IRBuilder<> IRB(InsertBefore);
750  CallInst *Call = SizeArgument
751    ? IRB.CreateCall2(AsanErrorCallbackSized[IsWrite], Addr, SizeArgument)
752    : IRB.CreateCall(AsanErrorCallback[IsWrite][AccessSizeIndex], Addr);
753
754  // We don't do Call->setDoesNotReturn() because the BB already has
755  // UnreachableInst at the end.
756  // This EmptyAsm is required to avoid callback merge.
757  IRB.CreateCall(EmptyAsm);
758  return Call;
759}
760
761Value *AddressSanitizer::createSlowPathCmp(IRBuilder<> &IRB, Value *AddrLong,
762                                            Value *ShadowValue,
763                                            uint32_t TypeSize) {
764  size_t Granularity = 1 << Mapping.Scale;
765  // Addr & (Granularity - 1)
766  Value *LastAccessedByte = IRB.CreateAnd(
767      AddrLong, ConstantInt::get(IntptrTy, Granularity - 1));
768  // (Addr & (Granularity - 1)) + size - 1
769  if (TypeSize / 8 > 1)
770    LastAccessedByte = IRB.CreateAdd(
771        LastAccessedByte, ConstantInt::get(IntptrTy, TypeSize / 8 - 1));
772  // (uint8_t) ((Addr & (Granularity-1)) + size - 1)
773  LastAccessedByte = IRB.CreateIntCast(
774      LastAccessedByte, ShadowValue->getType(), false);
775  // ((uint8_t) ((Addr & (Granularity-1)) + size - 1)) >= ShadowValue
776  return IRB.CreateICmpSGE(LastAccessedByte, ShadowValue);
777}
778
779void AddressSanitizer::instrumentAddress(Instruction *OrigIns,
780                                         Instruction *InsertBefore,
781                                         Value *Addr, uint32_t TypeSize,
782                                         bool IsWrite, Value *SizeArgument) {
783  IRBuilder<> IRB(InsertBefore);
784  Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
785
786  Type *ShadowTy  = IntegerType::get(
787      *C, std::max(8U, TypeSize >> Mapping.Scale));
788  Type *ShadowPtrTy = PointerType::get(ShadowTy, 0);
789  Value *ShadowPtr = memToShadow(AddrLong, IRB);
790  Value *CmpVal = Constant::getNullValue(ShadowTy);
791  Value *ShadowValue = IRB.CreateLoad(
792      IRB.CreateIntToPtr(ShadowPtr, ShadowPtrTy));
793
794  Value *Cmp = IRB.CreateICmpNE(ShadowValue, CmpVal);
795  size_t AccessSizeIndex = TypeSizeToSizeIndex(TypeSize);
796  size_t Granularity = 1 << Mapping.Scale;
797  TerminatorInst *CrashTerm = 0;
798
799  if (ClAlwaysSlowPath || (TypeSize < 8 * Granularity)) {
800    TerminatorInst *CheckTerm =
801        SplitBlockAndInsertIfThen(cast<Instruction>(Cmp), false);
802    assert(dyn_cast<BranchInst>(CheckTerm)->isUnconditional());
803    BasicBlock *NextBB = CheckTerm->getSuccessor(0);
804    IRB.SetInsertPoint(CheckTerm);
805    Value *Cmp2 = createSlowPathCmp(IRB, AddrLong, ShadowValue, TypeSize);
806    BasicBlock *CrashBlock =
807        BasicBlock::Create(*C, "", NextBB->getParent(), NextBB);
808    CrashTerm = new UnreachableInst(*C, CrashBlock);
809    BranchInst *NewTerm = BranchInst::Create(CrashBlock, NextBB, Cmp2);
810    ReplaceInstWithInst(CheckTerm, NewTerm);
811  } else {
812    CrashTerm = SplitBlockAndInsertIfThen(cast<Instruction>(Cmp), true);
813  }
814
815  Instruction *Crash = generateCrashCode(
816      CrashTerm, AddrLong, IsWrite, AccessSizeIndex, SizeArgument);
817  Crash->setDebugLoc(OrigIns->getDebugLoc());
818}
819
820void AddressSanitizerModule::createInitializerPoisonCalls(
821    Module &M, GlobalValue *ModuleName) {
822  // We do all of our poisoning and unpoisoning within _GLOBAL__I_a.
823  Function *GlobalInit = M.getFunction("_GLOBAL__I_a");
824  // If that function is not present, this TU contains no globals, or they have
825  // all been optimized away
826  if (!GlobalInit)
827    return;
828
829  // Set up the arguments to our poison/unpoison functions.
830  IRBuilder<> IRB(GlobalInit->begin()->getFirstInsertionPt());
831
832  // Add a call to poison all external globals before the given function starts.
833  Value *ModuleNameAddr = ConstantExpr::getPointerCast(ModuleName, IntptrTy);
834  IRB.CreateCall(AsanPoisonGlobals, ModuleNameAddr);
835
836  // Add calls to unpoison all globals before each return instruction.
837  for (Function::iterator I = GlobalInit->begin(), E = GlobalInit->end();
838      I != E; ++I) {
839    if (ReturnInst *RI = dyn_cast<ReturnInst>(I->getTerminator())) {
840      CallInst::Create(AsanUnpoisonGlobals, "", RI);
841    }
842  }
843}
844
845bool AddressSanitizerModule::ShouldInstrumentGlobal(GlobalVariable *G) {
846  Type *Ty = cast<PointerType>(G->getType())->getElementType();
847  DEBUG(dbgs() << "GLOBAL: " << *G << "\n");
848
849  if (BL->isIn(*G)) return false;
850  if (!Ty->isSized()) return false;
851  if (!G->hasInitializer()) return false;
852  if (GlobalWasGeneratedByAsan(G)) return false;  // Our own global.
853  // Touch only those globals that will not be defined in other modules.
854  // Don't handle ODR type linkages since other modules may be built w/o asan.
855  if (G->getLinkage() != GlobalVariable::ExternalLinkage &&
856      G->getLinkage() != GlobalVariable::PrivateLinkage &&
857      G->getLinkage() != GlobalVariable::InternalLinkage)
858    return false;
859  // Two problems with thread-locals:
860  //   - The address of the main thread's copy can't be computed at link-time.
861  //   - Need to poison all copies, not just the main thread's one.
862  if (G->isThreadLocal())
863    return false;
864  // For now, just ignore this Alloca if the alignment is large.
865  if (G->getAlignment() > RedzoneSize()) return false;
866
867  // Ignore all the globals with the names starting with "\01L_OBJC_".
868  // Many of those are put into the .cstring section. The linker compresses
869  // that section by removing the spare \0s after the string terminator, so
870  // our redzones get broken.
871  if ((G->getName().find("\01L_OBJC_") == 0) ||
872      (G->getName().find("\01l_OBJC_") == 0)) {
873    DEBUG(dbgs() << "Ignoring \\01L_OBJC_* global: " << *G);
874    return false;
875  }
876
877  if (G->hasSection()) {
878    StringRef Section(G->getSection());
879    // Ignore the globals from the __OBJC section. The ObjC runtime assumes
880    // those conform to /usr/lib/objc/runtime.h, so we can't add redzones to
881    // them.
882    if ((Section.find("__OBJC,") == 0) ||
883        (Section.find("__DATA, __objc_") == 0)) {
884      DEBUG(dbgs() << "Ignoring ObjC runtime global: " << *G);
885      return false;
886    }
887    // See http://code.google.com/p/address-sanitizer/issues/detail?id=32
888    // Constant CFString instances are compiled in the following way:
889    //  -- the string buffer is emitted into
890    //     __TEXT,__cstring,cstring_literals
891    //  -- the constant NSConstantString structure referencing that buffer
892    //     is placed into __DATA,__cfstring
893    // Therefore there's no point in placing redzones into __DATA,__cfstring.
894    // Moreover, it causes the linker to crash on OS X 10.7
895    if (Section.find("__DATA,__cfstring") == 0) {
896      DEBUG(dbgs() << "Ignoring CFString: " << *G);
897      return false;
898    }
899  }
900
901  return true;
902}
903
904void AddressSanitizerModule::initializeCallbacks(Module &M) {
905  IRBuilder<> IRB(*C);
906  // Declare our poisoning and unpoisoning functions.
907  AsanPoisonGlobals = checkInterfaceFunction(M.getOrInsertFunction(
908      kAsanPoisonGlobalsName, IRB.getVoidTy(), IntptrTy, NULL));
909  AsanPoisonGlobals->setLinkage(Function::ExternalLinkage);
910  AsanUnpoisonGlobals = checkInterfaceFunction(M.getOrInsertFunction(
911      kAsanUnpoisonGlobalsName, IRB.getVoidTy(), NULL));
912  AsanUnpoisonGlobals->setLinkage(Function::ExternalLinkage);
913  // Declare functions that register/unregister globals.
914  AsanRegisterGlobals = checkInterfaceFunction(M.getOrInsertFunction(
915      kAsanRegisterGlobalsName, IRB.getVoidTy(),
916      IntptrTy, IntptrTy, NULL));
917  AsanRegisterGlobals->setLinkage(Function::ExternalLinkage);
918  AsanUnregisterGlobals = checkInterfaceFunction(M.getOrInsertFunction(
919      kAsanUnregisterGlobalsName,
920      IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
921  AsanUnregisterGlobals->setLinkage(Function::ExternalLinkage);
922}
923
924// This function replaces all global variables with new variables that have
925// trailing redzones. It also creates a function that poisons
926// redzones and inserts this function into llvm.global_ctors.
927bool AddressSanitizerModule::runOnModule(Module &M) {
928  if (!ClGlobals) return false;
929  TD = getAnalysisIfAvailable<DataLayout>();
930  if (!TD)
931    return false;
932  BL.reset(SpecialCaseList::createOrDie(BlacklistFile));
933  if (BL->isIn(M)) return false;
934  C = &(M.getContext());
935  int LongSize = TD->getPointerSizeInBits();
936  IntptrTy = Type::getIntNTy(*C, LongSize);
937  Mapping = getShadowMapping(M, LongSize, ZeroBaseShadow);
938  initializeCallbacks(M);
939  DynamicallyInitializedGlobals.Init(M);
940
941  SmallVector<GlobalVariable *, 16> GlobalsToChange;
942
943  for (Module::GlobalListType::iterator G = M.global_begin(),
944       E = M.global_end(); G != E; ++G) {
945    if (ShouldInstrumentGlobal(G))
946      GlobalsToChange.push_back(G);
947  }
948
949  size_t n = GlobalsToChange.size();
950  if (n == 0) return false;
951
952  // A global is described by a structure
953  //   size_t beg;
954  //   size_t size;
955  //   size_t size_with_redzone;
956  //   const char *name;
957  //   const char *module_name;
958  //   size_t has_dynamic_init;
959  // We initialize an array of such structures and pass it to a run-time call.
960  StructType *GlobalStructTy = StructType::get(IntptrTy, IntptrTy,
961                                               IntptrTy, IntptrTy,
962                                               IntptrTy, IntptrTy, NULL);
963  SmallVector<Constant *, 16> Initializers(n);
964
965  Function *CtorFunc = M.getFunction(kAsanModuleCtorName);
966  assert(CtorFunc);
967  IRBuilder<> IRB(CtorFunc->getEntryBlock().getTerminator());
968
969  bool HasDynamicallyInitializedGlobals = false;
970
971  GlobalVariable *ModuleName = createPrivateGlobalForString(
972      M, M.getModuleIdentifier());
973  // We shouldn't merge same module names, as this string serves as unique
974  // module ID in runtime.
975  ModuleName->setUnnamedAddr(false);
976
977  for (size_t i = 0; i < n; i++) {
978    static const uint64_t kMaxGlobalRedzone = 1 << 18;
979    GlobalVariable *G = GlobalsToChange[i];
980    PointerType *PtrTy = cast<PointerType>(G->getType());
981    Type *Ty = PtrTy->getElementType();
982    uint64_t SizeInBytes = TD->getTypeAllocSize(Ty);
983    uint64_t MinRZ = RedzoneSize();
984    // MinRZ <= RZ <= kMaxGlobalRedzone
985    // and trying to make RZ to be ~ 1/4 of SizeInBytes.
986    uint64_t RZ = std::max(MinRZ,
987                         std::min(kMaxGlobalRedzone,
988                                  (SizeInBytes / MinRZ / 4) * MinRZ));
989    uint64_t RightRedzoneSize = RZ;
990    // Round up to MinRZ
991    if (SizeInBytes % MinRZ)
992      RightRedzoneSize += MinRZ - (SizeInBytes % MinRZ);
993    assert(((RightRedzoneSize + SizeInBytes) % MinRZ) == 0);
994    Type *RightRedZoneTy = ArrayType::get(IRB.getInt8Ty(), RightRedzoneSize);
995    // Determine whether this global should be poisoned in initialization.
996    bool GlobalHasDynamicInitializer =
997        DynamicallyInitializedGlobals.Contains(G);
998    // Don't check initialization order if this global is blacklisted.
999    GlobalHasDynamicInitializer &= !BL->isIn(*G, "init");
1000
1001    StructType *NewTy = StructType::get(Ty, RightRedZoneTy, NULL);
1002    Constant *NewInitializer = ConstantStruct::get(
1003        NewTy, G->getInitializer(),
1004        Constant::getNullValue(RightRedZoneTy), NULL);
1005
1006    GlobalVariable *Name = createPrivateGlobalForString(M, G->getName());
1007
1008    // Create a new global variable with enough space for a redzone.
1009    GlobalValue::LinkageTypes Linkage = G->getLinkage();
1010    if (G->isConstant() && Linkage == GlobalValue::PrivateLinkage)
1011      Linkage = GlobalValue::InternalLinkage;
1012    GlobalVariable *NewGlobal = new GlobalVariable(
1013        M, NewTy, G->isConstant(), Linkage,
1014        NewInitializer, "", G, G->getThreadLocalMode());
1015    NewGlobal->copyAttributesFrom(G);
1016    NewGlobal->setAlignment(MinRZ);
1017
1018    Value *Indices2[2];
1019    Indices2[0] = IRB.getInt32(0);
1020    Indices2[1] = IRB.getInt32(0);
1021
1022    G->replaceAllUsesWith(
1023        ConstantExpr::getGetElementPtr(NewGlobal, Indices2, true));
1024    NewGlobal->takeName(G);
1025    G->eraseFromParent();
1026
1027    Initializers[i] = ConstantStruct::get(
1028        GlobalStructTy,
1029        ConstantExpr::getPointerCast(NewGlobal, IntptrTy),
1030        ConstantInt::get(IntptrTy, SizeInBytes),
1031        ConstantInt::get(IntptrTy, SizeInBytes + RightRedzoneSize),
1032        ConstantExpr::getPointerCast(Name, IntptrTy),
1033        ConstantExpr::getPointerCast(ModuleName, IntptrTy),
1034        ConstantInt::get(IntptrTy, GlobalHasDynamicInitializer),
1035        NULL);
1036
1037    // Populate the first and last globals declared in this TU.
1038    if (CheckInitOrder && GlobalHasDynamicInitializer)
1039      HasDynamicallyInitializedGlobals = true;
1040
1041    DEBUG(dbgs() << "NEW GLOBAL: " << *NewGlobal << "\n");
1042  }
1043
1044  ArrayType *ArrayOfGlobalStructTy = ArrayType::get(GlobalStructTy, n);
1045  GlobalVariable *AllGlobals = new GlobalVariable(
1046      M, ArrayOfGlobalStructTy, false, GlobalVariable::InternalLinkage,
1047      ConstantArray::get(ArrayOfGlobalStructTy, Initializers), "");
1048
1049  // Create calls for poisoning before initializers run and unpoisoning after.
1050  if (CheckInitOrder && HasDynamicallyInitializedGlobals)
1051    createInitializerPoisonCalls(M, ModuleName);
1052  IRB.CreateCall2(AsanRegisterGlobals,
1053                  IRB.CreatePointerCast(AllGlobals, IntptrTy),
1054                  ConstantInt::get(IntptrTy, n));
1055
1056  // We also need to unregister globals at the end, e.g. when a shared library
1057  // gets closed.
1058  Function *AsanDtorFunction = Function::Create(
1059      FunctionType::get(Type::getVoidTy(*C), false),
1060      GlobalValue::InternalLinkage, kAsanModuleDtorName, &M);
1061  BasicBlock *AsanDtorBB = BasicBlock::Create(*C, "", AsanDtorFunction);
1062  IRBuilder<> IRB_Dtor(ReturnInst::Create(*C, AsanDtorBB));
1063  IRB_Dtor.CreateCall2(AsanUnregisterGlobals,
1064                       IRB.CreatePointerCast(AllGlobals, IntptrTy),
1065                       ConstantInt::get(IntptrTy, n));
1066  appendToGlobalDtors(M, AsanDtorFunction, kAsanCtorAndCtorPriority);
1067
1068  DEBUG(dbgs() << M);
1069  return true;
1070}
1071
1072void AddressSanitizer::initializeCallbacks(Module &M) {
1073  IRBuilder<> IRB(*C);
1074  // Create __asan_report* callbacks.
1075  for (size_t AccessIsWrite = 0; AccessIsWrite <= 1; AccessIsWrite++) {
1076    for (size_t AccessSizeIndex = 0; AccessSizeIndex < kNumberOfAccessSizes;
1077         AccessSizeIndex++) {
1078      // IsWrite and TypeSize are encoded in the function name.
1079      std::string FunctionName = std::string(kAsanReportErrorTemplate) +
1080          (AccessIsWrite ? "store" : "load") + itostr(1 << AccessSizeIndex);
1081      // If we are merging crash callbacks, they have two parameters.
1082      AsanErrorCallback[AccessIsWrite][AccessSizeIndex] =
1083          checkInterfaceFunction(M.getOrInsertFunction(
1084              FunctionName, IRB.getVoidTy(), IntptrTy, NULL));
1085    }
1086  }
1087  AsanErrorCallbackSized[0] = checkInterfaceFunction(M.getOrInsertFunction(
1088              kAsanReportLoadN, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
1089  AsanErrorCallbackSized[1] = checkInterfaceFunction(M.getOrInsertFunction(
1090              kAsanReportStoreN, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
1091
1092  AsanHandleNoReturnFunc = checkInterfaceFunction(M.getOrInsertFunction(
1093      kAsanHandleNoReturnName, IRB.getVoidTy(), NULL));
1094  AsanCovFunction = checkInterfaceFunction(M.getOrInsertFunction(
1095      kAsanCovName, IRB.getVoidTy(), IntptrTy, NULL));
1096  // We insert an empty inline asm after __asan_report* to avoid callback merge.
1097  EmptyAsm = InlineAsm::get(FunctionType::get(IRB.getVoidTy(), false),
1098                            StringRef(""), StringRef(""),
1099                            /*hasSideEffects=*/true);
1100}
1101
1102void AddressSanitizer::emitShadowMapping(Module &M, IRBuilder<> &IRB) const {
1103  // Tell the values of mapping offset and scale to the run-time.
1104  GlobalValue *asan_mapping_offset =
1105      new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
1106                     ConstantInt::get(IntptrTy, Mapping.Offset),
1107                     kAsanMappingOffsetName);
1108  // Read the global, otherwise it may be optimized away.
1109  IRB.CreateLoad(asan_mapping_offset, true);
1110
1111  GlobalValue *asan_mapping_scale =
1112      new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
1113                         ConstantInt::get(IntptrTy, Mapping.Scale),
1114                         kAsanMappingScaleName);
1115  // Read the global, otherwise it may be optimized away.
1116  IRB.CreateLoad(asan_mapping_scale, true);
1117}
1118
1119// virtual
1120bool AddressSanitizer::doInitialization(Module &M) {
1121  // Initialize the private fields. No one has accessed them before.
1122  TD = getAnalysisIfAvailable<DataLayout>();
1123
1124  if (!TD)
1125    return false;
1126  BL.reset(SpecialCaseList::createOrDie(BlacklistFile));
1127  DynamicallyInitializedGlobals.Init(M);
1128
1129  C = &(M.getContext());
1130  LongSize = TD->getPointerSizeInBits();
1131  IntptrTy = Type::getIntNTy(*C, LongSize);
1132
1133  AsanCtorFunction = Function::Create(
1134      FunctionType::get(Type::getVoidTy(*C), false),
1135      GlobalValue::InternalLinkage, kAsanModuleCtorName, &M);
1136  BasicBlock *AsanCtorBB = BasicBlock::Create(*C, "", AsanCtorFunction);
1137  // call __asan_init in the module ctor.
1138  IRBuilder<> IRB(ReturnInst::Create(*C, AsanCtorBB));
1139  AsanInitFunction = checkInterfaceFunction(
1140      M.getOrInsertFunction(kAsanInitName, IRB.getVoidTy(), NULL));
1141  AsanInitFunction->setLinkage(Function::ExternalLinkage);
1142  IRB.CreateCall(AsanInitFunction);
1143
1144  Mapping = getShadowMapping(M, LongSize, ZeroBaseShadow);
1145  emitShadowMapping(M, IRB);
1146
1147  appendToGlobalCtors(M, AsanCtorFunction, kAsanCtorAndCtorPriority);
1148  return true;
1149}
1150
1151bool AddressSanitizer::maybeInsertAsanInitAtFunctionEntry(Function &F) {
1152  // For each NSObject descendant having a +load method, this method is invoked
1153  // by the ObjC runtime before any of the static constructors is called.
1154  // Therefore we need to instrument such methods with a call to __asan_init
1155  // at the beginning in order to initialize our runtime before any access to
1156  // the shadow memory.
1157  // We cannot just ignore these methods, because they may call other
1158  // instrumented functions.
1159  if (F.getName().find(" load]") != std::string::npos) {
1160    IRBuilder<> IRB(F.begin()->begin());
1161    IRB.CreateCall(AsanInitFunction);
1162    return true;
1163  }
1164  return false;
1165}
1166
1167// Poor man's coverage that works with ASan.
1168// We create a Guard boolean variable with the same linkage
1169// as the function and inject this code into the entry block:
1170// if (*Guard) {
1171//    __sanitizer_cov(&F);
1172//    *Guard = 1;
1173// }
1174// The accesses to Guard are atomic. The rest of the logic is
1175// in __sanitizer_cov (it's fine to call it more than once).
1176//
1177// This coverage implementation provides very limited data:
1178// it only tells if a given function was ever executed.
1179// No counters, no per-basic-block or per-edge data.
1180// But for many use cases this is what we need and the added slowdown
1181// is negligible. This simple implementation will probably be obsoleted
1182// by the upcoming Clang-based coverage implementation.
1183// By having it here and now we hope to
1184//  a) get the functionality to users earlier and
1185//  b) collect usage statistics to help improve Clang coverage design.
1186bool AddressSanitizer::InjectCoverage(Function &F) {
1187  if (!ClCoverage) return false;
1188  IRBuilder<> IRB(F.getEntryBlock().getFirstInsertionPt());
1189  Type *Int8Ty = IRB.getInt8Ty();
1190  GlobalVariable *Guard = new GlobalVariable(
1191      *F.getParent(), Int8Ty, false, GlobalValue::PrivateLinkage,
1192      Constant::getNullValue(Int8Ty), "__asan_gen_cov_" + F.getName());
1193  LoadInst *Load = IRB.CreateLoad(Guard);
1194  Load->setAtomic(Monotonic);
1195  Load->setAlignment(1);
1196  Value *Cmp = IRB.CreateICmpEQ(Constant::getNullValue(Int8Ty), Load);
1197  Instruction *Ins = SplitBlockAndInsertIfThen(cast<Instruction>(Cmp), false);
1198  IRB.SetInsertPoint(Ins);
1199  // We pass &F to __sanitizer_cov. We could avoid this and rely on
1200  // GET_CALLER_PC, but having the PC of the first instruction is just nice.
1201  IRB.CreateCall(AsanCovFunction, IRB.CreatePointerCast(&F, IntptrTy));
1202  StoreInst *Store = IRB.CreateStore(ConstantInt::get(Int8Ty, 1), Guard);
1203  Store->setAtomic(Monotonic);
1204  Store->setAlignment(1);
1205  return true;
1206}
1207
1208bool AddressSanitizer::runOnFunction(Function &F) {
1209  if (BL->isIn(F)) return false;
1210  if (&F == AsanCtorFunction) return false;
1211  if (F.getLinkage() == GlobalValue::AvailableExternallyLinkage) return false;
1212  DEBUG(dbgs() << "ASAN instrumenting:\n" << F << "\n");
1213  initializeCallbacks(*F.getParent());
1214
1215  // If needed, insert __asan_init before checking for SanitizeAddress attr.
1216  maybeInsertAsanInitAtFunctionEntry(F);
1217
1218  if (!F.hasFnAttribute(Attribute::SanitizeAddress))
1219    return false;
1220
1221  if (!ClDebugFunc.empty() && ClDebugFunc != F.getName())
1222    return false;
1223
1224  // We want to instrument every address only once per basic block (unless there
1225  // are calls between uses).
1226  SmallSet<Value*, 16> TempsToInstrument;
1227  SmallVector<Instruction*, 16> ToInstrument;
1228  SmallVector<Instruction*, 8> NoReturnCalls;
1229  int NumAllocas = 0;
1230  bool IsWrite;
1231
1232  // Fill the set of memory operations to instrument.
1233  for (Function::iterator FI = F.begin(), FE = F.end();
1234       FI != FE; ++FI) {
1235    TempsToInstrument.clear();
1236    int NumInsnsPerBB = 0;
1237    for (BasicBlock::iterator BI = FI->begin(), BE = FI->end();
1238         BI != BE; ++BI) {
1239      if (LooksLikeCodeInBug11395(BI)) return false;
1240      if (Value *Addr = isInterestingMemoryAccess(BI, &IsWrite)) {
1241        if (ClOpt && ClOptSameTemp) {
1242          if (!TempsToInstrument.insert(Addr))
1243            continue;  // We've seen this temp in the current BB.
1244        }
1245      } else if (isa<MemIntrinsic>(BI) && ClMemIntrin) {
1246        // ok, take it.
1247      } else {
1248        if (isa<AllocaInst>(BI))
1249          NumAllocas++;
1250        CallSite CS(BI);
1251        if (CS) {
1252          // A call inside BB.
1253          TempsToInstrument.clear();
1254          if (CS.doesNotReturn())
1255            NoReturnCalls.push_back(CS.getInstruction());
1256        }
1257        continue;
1258      }
1259      ToInstrument.push_back(BI);
1260      NumInsnsPerBB++;
1261      if (NumInsnsPerBB >= ClMaxInsnsToInstrumentPerBB)
1262        break;
1263    }
1264  }
1265
1266  Function *UninstrumentedDuplicate = 0;
1267  bool LikelyToInstrument =
1268      !NoReturnCalls.empty() || !ToInstrument.empty() || (NumAllocas > 0);
1269  if (ClKeepUninstrumented && LikelyToInstrument) {
1270    ValueToValueMapTy VMap;
1271    UninstrumentedDuplicate = CloneFunction(&F, VMap, false);
1272    UninstrumentedDuplicate->removeFnAttr(Attribute::SanitizeAddress);
1273    UninstrumentedDuplicate->setName("NOASAN_" + F.getName());
1274    F.getParent()->getFunctionList().push_back(UninstrumentedDuplicate);
1275  }
1276
1277  // Instrument.
1278  int NumInstrumented = 0;
1279  for (size_t i = 0, n = ToInstrument.size(); i != n; i++) {
1280    Instruction *Inst = ToInstrument[i];
1281    if (ClDebugMin < 0 || ClDebugMax < 0 ||
1282        (NumInstrumented >= ClDebugMin && NumInstrumented <= ClDebugMax)) {
1283      if (isInterestingMemoryAccess(Inst, &IsWrite))
1284        instrumentMop(Inst);
1285      else
1286        instrumentMemIntrinsic(cast<MemIntrinsic>(Inst));
1287    }
1288    NumInstrumented++;
1289  }
1290
1291  FunctionStackPoisoner FSP(F, *this);
1292  bool ChangedStack = FSP.runOnFunction();
1293
1294  // We must unpoison the stack before every NoReturn call (throw, _exit, etc).
1295  // See e.g. http://code.google.com/p/address-sanitizer/issues/detail?id=37
1296  for (size_t i = 0, n = NoReturnCalls.size(); i != n; i++) {
1297    Instruction *CI = NoReturnCalls[i];
1298    IRBuilder<> IRB(CI);
1299    IRB.CreateCall(AsanHandleNoReturnFunc);
1300  }
1301
1302  bool res = NumInstrumented > 0 || ChangedStack || !NoReturnCalls.empty();
1303
1304  if (InjectCoverage(F))
1305    res = true;
1306
1307  DEBUG(dbgs() << "ASAN done instrumenting: " << res << " " << F << "\n");
1308
1309  if (ClKeepUninstrumented) {
1310    if (!res) {
1311      // No instrumentation is done, no need for the duplicate.
1312      if (UninstrumentedDuplicate)
1313        UninstrumentedDuplicate->eraseFromParent();
1314    } else {
1315      // The function was instrumented. We must have the duplicate.
1316      assert(UninstrumentedDuplicate);
1317      UninstrumentedDuplicate->setSection("NOASAN");
1318      assert(!F.hasSection());
1319      F.setSection("ASAN");
1320    }
1321  }
1322
1323  return res;
1324}
1325
1326static uint64_t ValueForPoison(uint64_t PoisonByte, size_t ShadowRedzoneSize) {
1327  if (ShadowRedzoneSize == 1) return PoisonByte;
1328  if (ShadowRedzoneSize == 2) return (PoisonByte << 8) + PoisonByte;
1329  if (ShadowRedzoneSize == 4)
1330    return (PoisonByte << 24) + (PoisonByte << 16) +
1331        (PoisonByte << 8) + (PoisonByte);
1332  llvm_unreachable("ShadowRedzoneSize is either 1, 2 or 4");
1333}
1334
1335static void PoisonShadowPartialRightRedzone(uint8_t *Shadow,
1336                                            size_t Size,
1337                                            size_t RZSize,
1338                                            size_t ShadowGranularity,
1339                                            uint8_t Magic) {
1340  for (size_t i = 0; i < RZSize;
1341       i+= ShadowGranularity, Shadow++) {
1342    if (i + ShadowGranularity <= Size) {
1343      *Shadow = 0;  // fully addressable
1344    } else if (i >= Size) {
1345      *Shadow = Magic;  // unaddressable
1346    } else {
1347      *Shadow = Size - i;  // first Size-i bytes are addressable
1348    }
1349  }
1350}
1351
1352// Workaround for bug 11395: we don't want to instrument stack in functions
1353// with large assembly blobs (32-bit only), otherwise reg alloc may crash.
1354// FIXME: remove once the bug 11395 is fixed.
1355bool AddressSanitizer::LooksLikeCodeInBug11395(Instruction *I) {
1356  if (LongSize != 32) return false;
1357  CallInst *CI = dyn_cast<CallInst>(I);
1358  if (!CI || !CI->isInlineAsm()) return false;
1359  if (CI->getNumArgOperands() <= 5) return false;
1360  // We have inline assembly with quite a few arguments.
1361  return true;
1362}
1363
1364void FunctionStackPoisoner::initializeCallbacks(Module &M) {
1365  IRBuilder<> IRB(*C);
1366  for (int i = 0; i <= kMaxAsanStackMallocSizeClass; i++) {
1367    std::string Suffix = itostr(i);
1368    AsanStackMallocFunc[i] = checkInterfaceFunction(
1369        M.getOrInsertFunction(kAsanStackMallocNameTemplate + Suffix, IntptrTy,
1370                              IntptrTy, IntptrTy, NULL));
1371    AsanStackFreeFunc[i] = checkInterfaceFunction(M.getOrInsertFunction(
1372        kAsanStackFreeNameTemplate + Suffix, IRB.getVoidTy(), IntptrTy,
1373        IntptrTy, IntptrTy, NULL));
1374  }
1375  AsanPoisonStackMemoryFunc = checkInterfaceFunction(M.getOrInsertFunction(
1376      kAsanPoisonStackMemoryName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
1377  AsanUnpoisonStackMemoryFunc = checkInterfaceFunction(M.getOrInsertFunction(
1378      kAsanUnpoisonStackMemoryName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
1379}
1380
1381void FunctionStackPoisoner::poisonRedZones(
1382  const ArrayRef<AllocaInst*> &AllocaVec, IRBuilder<> &IRB, Value *ShadowBase,
1383  bool DoPoison) {
1384  size_t ShadowRZSize = RedzoneSize() >> Mapping.Scale;
1385  assert(ShadowRZSize >= 1 && ShadowRZSize <= 4);
1386  Type *RZTy = Type::getIntNTy(*C, ShadowRZSize * 8);
1387  Type *RZPtrTy = PointerType::get(RZTy, 0);
1388
1389  Value *PoisonLeft  = ConstantInt::get(RZTy,
1390    ValueForPoison(DoPoison ? kAsanStackLeftRedzoneMagic : 0LL, ShadowRZSize));
1391  Value *PoisonMid   = ConstantInt::get(RZTy,
1392    ValueForPoison(DoPoison ? kAsanStackMidRedzoneMagic : 0LL, ShadowRZSize));
1393  Value *PoisonRight = ConstantInt::get(RZTy,
1394    ValueForPoison(DoPoison ? kAsanStackRightRedzoneMagic : 0LL, ShadowRZSize));
1395
1396  // poison the first red zone.
1397  IRB.CreateStore(PoisonLeft, IRB.CreateIntToPtr(ShadowBase, RZPtrTy));
1398
1399  // poison all other red zones.
1400  uint64_t Pos = RedzoneSize();
1401  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
1402    AllocaInst *AI = AllocaVec[i];
1403    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
1404    uint64_t AlignedSize = getAlignedAllocaSize(AI);
1405    assert(AlignedSize - SizeInBytes < RedzoneSize());
1406    Value *Ptr = NULL;
1407
1408    Pos += AlignedSize;
1409
1410    assert(ShadowBase->getType() == IntptrTy);
1411    if (SizeInBytes < AlignedSize) {
1412      // Poison the partial redzone at right
1413      Ptr = IRB.CreateAdd(
1414          ShadowBase, ConstantInt::get(IntptrTy,
1415                                       (Pos >> Mapping.Scale) - ShadowRZSize));
1416      size_t AddressableBytes = RedzoneSize() - (AlignedSize - SizeInBytes);
1417      uint32_t Poison = 0;
1418      if (DoPoison) {
1419        PoisonShadowPartialRightRedzone((uint8_t*)&Poison, AddressableBytes,
1420                                        RedzoneSize(),
1421                                        1ULL << Mapping.Scale,
1422                                        kAsanStackPartialRedzoneMagic);
1423        Poison =
1424            ASan.TD->isLittleEndian()
1425                ? support::endian::byte_swap<uint32_t, support::little>(Poison)
1426                : support::endian::byte_swap<uint32_t, support::big>(Poison);
1427      }
1428      Value *PartialPoison = ConstantInt::get(RZTy, Poison);
1429      IRB.CreateStore(PartialPoison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
1430    }
1431
1432    // Poison the full redzone at right.
1433    Ptr = IRB.CreateAdd(ShadowBase,
1434                        ConstantInt::get(IntptrTy, Pos >> Mapping.Scale));
1435    bool LastAlloca = (i == AllocaVec.size() - 1);
1436    Value *Poison = LastAlloca ? PoisonRight : PoisonMid;
1437    IRB.CreateStore(Poison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
1438
1439    Pos += RedzoneSize();
1440  }
1441}
1442
1443// Fake stack allocator (asan_fake_stack.h) has 11 size classes
1444// for every power of 2 from kMinStackMallocSize to kMaxAsanStackMallocSizeClass
1445static int StackMallocSizeClass(uint64_t LocalStackSize) {
1446  assert(LocalStackSize <= kMaxStackMallocSize);
1447  uint64_t MaxSize = kMinStackMallocSize;
1448  for (int i = 0; ; i++, MaxSize *= 2)
1449    if (LocalStackSize <= MaxSize)
1450      return i;
1451  llvm_unreachable("impossible LocalStackSize");
1452}
1453
1454// Set Size bytes starting from ShadowBase to kAsanStackAfterReturnMagic.
1455// We can not use MemSet intrinsic because it may end up calling the actual
1456// memset. Size is a multiple of 8.
1457// Currently this generates 8-byte stores on x86_64; it may be better to
1458// generate wider stores.
1459void FunctionStackPoisoner::SetShadowToStackAfterReturnInlined(
1460    IRBuilder<> &IRB, Value *ShadowBase, int Size) {
1461  assert(!(Size % 8));
1462  assert(kAsanStackAfterReturnMagic == 0xf5);
1463  for (int i = 0; i < Size; i += 8) {
1464    Value *p = IRB.CreateAdd(ShadowBase, ConstantInt::get(IntptrTy, i));
1465    IRB.CreateStore(ConstantInt::get(IRB.getInt64Ty(), 0xf5f5f5f5f5f5f5f5ULL),
1466                    IRB.CreateIntToPtr(p, IRB.getInt64Ty()->getPointerTo()));
1467  }
1468}
1469
1470void FunctionStackPoisoner::poisonStack() {
1471  uint64_t LocalStackSize = TotalStackSize +
1472                            (AllocaVec.size() + 1) * RedzoneSize();
1473
1474  bool DoStackMalloc = ASan.CheckUseAfterReturn
1475      && LocalStackSize <= kMaxStackMallocSize;
1476  int StackMallocIdx = -1;
1477
1478  assert(AllocaVec.size() > 0);
1479  Instruction *InsBefore = AllocaVec[0];
1480  IRBuilder<> IRB(InsBefore);
1481
1482
1483  Type *ByteArrayTy = ArrayType::get(IRB.getInt8Ty(), LocalStackSize);
1484  AllocaInst *MyAlloca =
1485      new AllocaInst(ByteArrayTy, "MyAlloca", InsBefore);
1486  if (ClRealignStack && StackAlignment < RedzoneSize())
1487    StackAlignment = RedzoneSize();
1488  MyAlloca->setAlignment(StackAlignment);
1489  assert(MyAlloca->isStaticAlloca());
1490  Value *OrigStackBase = IRB.CreatePointerCast(MyAlloca, IntptrTy);
1491  Value *LocalStackBase = OrigStackBase;
1492
1493  if (DoStackMalloc) {
1494    // LocalStackBase = OrigStackBase
1495    // if (__asan_option_detect_stack_use_after_return)
1496    //   LocalStackBase = __asan_stack_malloc_N(LocalStackBase, OrigStackBase);
1497    StackMallocIdx = StackMallocSizeClass(LocalStackSize);
1498    assert(StackMallocIdx <= kMaxAsanStackMallocSizeClass);
1499    Constant *OptionDetectUAR = F.getParent()->getOrInsertGlobal(
1500        kAsanOptionDetectUAR, IRB.getInt32Ty());
1501    Value *Cmp = IRB.CreateICmpNE(IRB.CreateLoad(OptionDetectUAR),
1502                                  Constant::getNullValue(IRB.getInt32Ty()));
1503    Instruction *Term =
1504        SplitBlockAndInsertIfThen(cast<Instruction>(Cmp), false);
1505    BasicBlock *CmpBlock = cast<Instruction>(Cmp)->getParent();
1506    IRBuilder<> IRBIf(Term);
1507    LocalStackBase = IRBIf.CreateCall2(
1508        AsanStackMallocFunc[StackMallocIdx],
1509        ConstantInt::get(IntptrTy, LocalStackSize), OrigStackBase);
1510    BasicBlock *SetBlock = cast<Instruction>(LocalStackBase)->getParent();
1511    IRB.SetInsertPoint(InsBefore);
1512    PHINode *Phi = IRB.CreatePHI(IntptrTy, 2);
1513    Phi->addIncoming(OrigStackBase, CmpBlock);
1514    Phi->addIncoming(LocalStackBase, SetBlock);
1515    LocalStackBase = Phi;
1516  }
1517
1518  // This string will be parsed by the run-time (DescribeAddressIfStack).
1519  SmallString<2048> StackDescriptionStorage;
1520  raw_svector_ostream StackDescription(StackDescriptionStorage);
1521  StackDescription << AllocaVec.size() << " ";
1522
1523  // Insert poison calls for lifetime intrinsics for alloca.
1524  bool HavePoisonedAllocas = false;
1525  for (size_t i = 0, n = AllocaPoisonCallVec.size(); i < n; i++) {
1526    const AllocaPoisonCall &APC = AllocaPoisonCallVec[i];
1527    assert(APC.InsBefore);
1528    assert(APC.AI);
1529    IRBuilder<> IRB(APC.InsBefore);
1530    poisonAlloca(APC.AI, APC.Size, IRB, APC.DoPoison);
1531    HavePoisonedAllocas |= APC.DoPoison;
1532  }
1533
1534  uint64_t Pos = RedzoneSize();
1535  // Replace Alloca instructions with base+offset.
1536  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
1537    AllocaInst *AI = AllocaVec[i];
1538    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
1539    StringRef Name = AI->getName();
1540    StackDescription << Pos << " " << SizeInBytes << " "
1541                     << Name.size() << " " << Name << " ";
1542    uint64_t AlignedSize = getAlignedAllocaSize(AI);
1543    assert((AlignedSize % RedzoneSize()) == 0);
1544    Value *NewAllocaPtr = IRB.CreateIntToPtr(
1545            IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy, Pos)),
1546            AI->getType());
1547    replaceDbgDeclareForAlloca(AI, NewAllocaPtr, DIB);
1548    AI->replaceAllUsesWith(NewAllocaPtr);
1549    Pos += AlignedSize + RedzoneSize();
1550  }
1551  assert(Pos == LocalStackSize);
1552
1553  // The left-most redzone has enough space for at least 4 pointers.
1554  // Write the Magic value to redzone[0].
1555  Value *BasePlus0 = IRB.CreateIntToPtr(LocalStackBase, IntptrPtrTy);
1556  IRB.CreateStore(ConstantInt::get(IntptrTy, kCurrentStackFrameMagic),
1557                  BasePlus0);
1558  // Write the frame description constant to redzone[1].
1559  Value *BasePlus1 = IRB.CreateIntToPtr(
1560    IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy, ASan.LongSize/8)),
1561    IntptrPtrTy);
1562  GlobalVariable *StackDescriptionGlobal =
1563      createPrivateGlobalForString(*F.getParent(), StackDescription.str());
1564  Value *Description = IRB.CreatePointerCast(StackDescriptionGlobal,
1565                                             IntptrTy);
1566  IRB.CreateStore(Description, BasePlus1);
1567  // Write the PC to redzone[2].
1568  Value *BasePlus2 = IRB.CreateIntToPtr(
1569    IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy,
1570                                                   2 * ASan.LongSize/8)),
1571    IntptrPtrTy);
1572  IRB.CreateStore(IRB.CreatePointerCast(&F, IntptrTy), BasePlus2);
1573
1574  // Poison the stack redzones at the entry.
1575  Value *ShadowBase = ASan.memToShadow(LocalStackBase, IRB);
1576  poisonRedZones(AllocaVec, IRB, ShadowBase, true);
1577
1578  // Unpoison the stack before all ret instructions.
1579  for (size_t i = 0, n = RetVec.size(); i < n; i++) {
1580    Instruction *Ret = RetVec[i];
1581    IRBuilder<> IRBRet(Ret);
1582    // Mark the current frame as retired.
1583    IRBRet.CreateStore(ConstantInt::get(IntptrTy, kRetiredStackFrameMagic),
1584                       BasePlus0);
1585    // Unpoison the stack.
1586    poisonRedZones(AllocaVec, IRBRet, ShadowBase, false);
1587    if (DoStackMalloc) {
1588      assert(StackMallocIdx >= 0);
1589      // In use-after-return mode, mark the whole stack frame unaddressable.
1590      if (StackMallocIdx <= 4) {
1591        // For small sizes inline the whole thing:
1592        // if LocalStackBase != OrigStackBase:
1593        //     memset(ShadowBase, kAsanStackAfterReturnMagic, ShadowSize);
1594        //     **SavedFlagPtr(LocalStackBase) = 0
1595        // FIXME: if LocalStackBase != OrigStackBase don't call poisonRedZones.
1596        Value *Cmp = IRBRet.CreateICmpNE(LocalStackBase, OrigStackBase);
1597        TerminatorInst *PoisonTerm =
1598            SplitBlockAndInsertIfThen(cast<Instruction>(Cmp), false);
1599        IRBuilder<> IRBPoison(PoisonTerm);
1600        int ClassSize = kMinStackMallocSize << StackMallocIdx;
1601        SetShadowToStackAfterReturnInlined(IRBPoison, ShadowBase,
1602                                           ClassSize >> Mapping.Scale);
1603        Value *SavedFlagPtrPtr = IRBPoison.CreateAdd(
1604            LocalStackBase,
1605            ConstantInt::get(IntptrTy, ClassSize - ASan.LongSize / 8));
1606        Value *SavedFlagPtr = IRBPoison.CreateLoad(
1607            IRBPoison.CreateIntToPtr(SavedFlagPtrPtr, IntptrPtrTy));
1608        IRBPoison.CreateStore(
1609            Constant::getNullValue(IRBPoison.getInt8Ty()),
1610            IRBPoison.CreateIntToPtr(SavedFlagPtr, IRBPoison.getInt8PtrTy()));
1611      } else {
1612        // For larger frames call __asan_stack_free_*.
1613        IRBRet.CreateCall3(AsanStackFreeFunc[StackMallocIdx], LocalStackBase,
1614                           ConstantInt::get(IntptrTy, LocalStackSize),
1615                           OrigStackBase);
1616      }
1617    } else if (HavePoisonedAllocas) {
1618      // If we poisoned some allocas in llvm.lifetime analysis,
1619      // unpoison whole stack frame now.
1620      assert(LocalStackBase == OrigStackBase);
1621      poisonAlloca(LocalStackBase, LocalStackSize, IRBRet, false);
1622    }
1623  }
1624
1625  // We are done. Remove the old unused alloca instructions.
1626  for (size_t i = 0, n = AllocaVec.size(); i < n; i++)
1627    AllocaVec[i]->eraseFromParent();
1628}
1629
1630void FunctionStackPoisoner::poisonAlloca(Value *V, uint64_t Size,
1631                                         IRBuilder<> &IRB, bool DoPoison) {
1632  // For now just insert the call to ASan runtime.
1633  Value *AddrArg = IRB.CreatePointerCast(V, IntptrTy);
1634  Value *SizeArg = ConstantInt::get(IntptrTy, Size);
1635  IRB.CreateCall2(DoPoison ? AsanPoisonStackMemoryFunc
1636                           : AsanUnpoisonStackMemoryFunc,
1637                  AddrArg, SizeArg);
1638}
1639
1640// Handling llvm.lifetime intrinsics for a given %alloca:
1641// (1) collect all llvm.lifetime.xxx(%size, %value) describing the alloca.
1642// (2) if %size is constant, poison memory for llvm.lifetime.end (to detect
1643//     invalid accesses) and unpoison it for llvm.lifetime.start (the memory
1644//     could be poisoned by previous llvm.lifetime.end instruction, as the
1645//     variable may go in and out of scope several times, e.g. in loops).
1646// (3) if we poisoned at least one %alloca in a function,
1647//     unpoison the whole stack frame at function exit.
1648
1649AllocaInst *FunctionStackPoisoner::findAllocaForValue(Value *V) {
1650  if (AllocaInst *AI = dyn_cast<AllocaInst>(V))
1651    // We're intested only in allocas we can handle.
1652    return isInterestingAlloca(*AI) ? AI : 0;
1653  // See if we've already calculated (or started to calculate) alloca for a
1654  // given value.
1655  AllocaForValueMapTy::iterator I = AllocaForValue.find(V);
1656  if (I != AllocaForValue.end())
1657    return I->second;
1658  // Store 0 while we're calculating alloca for value V to avoid
1659  // infinite recursion if the value references itself.
1660  AllocaForValue[V] = 0;
1661  AllocaInst *Res = 0;
1662  if (CastInst *CI = dyn_cast<CastInst>(V))
1663    Res = findAllocaForValue(CI->getOperand(0));
1664  else if (PHINode *PN = dyn_cast<PHINode>(V)) {
1665    for (unsigned i = 0, e = PN->getNumIncomingValues(); i != e; ++i) {
1666      Value *IncValue = PN->getIncomingValue(i);
1667      // Allow self-referencing phi-nodes.
1668      if (IncValue == PN) continue;
1669      AllocaInst *IncValueAI = findAllocaForValue(IncValue);
1670      // AI for incoming values should exist and should all be equal.
1671      if (IncValueAI == 0 || (Res != 0 && IncValueAI != Res))
1672        return 0;
1673      Res = IncValueAI;
1674    }
1675  }
1676  if (Res != 0)
1677    AllocaForValue[V] = Res;
1678  return Res;
1679}
1680