AddressSanitizer.cpp revision a1c45044099cf7fb4a072f1326f164706d5bb2e2 
1//===-- AddressSanitizer.cpp - memory error detector ------------*- C++ -*-===//
2//
3//                     The LLVM Compiler Infrastructure
4//
5// This file is distributed under the University of Illinois Open Source
6// License. See LICENSE.TXT for details.
7//
8//===----------------------------------------------------------------------===//
9//
10// This file is a part of AddressSanitizer, an address sanity checker.
11// Details of the algorithm:
12//  http://code.google.com/p/address-sanitizer/wiki/AddressSanitizerAlgorithm
13//
14//===----------------------------------------------------------------------===//
15
16#define DEBUG_TYPE "asan"
17
18#include "FunctionBlackList.h"
19#include "llvm/ADT/ArrayRef.h"
20#include "llvm/ADT/OwningPtr.h"
21#include "llvm/ADT/SmallSet.h"
22#include "llvm/ADT/SmallString.h"
23#include "llvm/ADT/SmallVector.h"
24#include "llvm/ADT/StringExtras.h"
25#include "llvm/Function.h"
26#include "llvm/IntrinsicInst.h"
27#include "llvm/LLVMContext.h"
28#include "llvm/Module.h"
29#include "llvm/Support/CommandLine.h"
30#include "llvm/Support/DataTypes.h"
31#include "llvm/Support/Debug.h"
32#include "llvm/Support/IRBuilder.h"
33#include "llvm/Support/raw_ostream.h"
34#include "llvm/Support/system_error.h"
35#include "llvm/Target/TargetData.h"
36#include "llvm/Target/TargetMachine.h"
37#include "llvm/Transforms/Instrumentation.h"
38#include "llvm/Transforms/Utils/BasicBlockUtils.h"
39#include "llvm/Transforms/Utils/ModuleUtils.h"
40#include "llvm/Type.h"
41
42#include <string>
43#include <algorithm>
44
45using namespace llvm;
46
47static const uint64_t kDefaultShadowScale = 3;
48static const uint64_t kDefaultShadowOffset32 = 1ULL << 29;
49static const uint64_t kDefaultShadowOffset64 = 1ULL << 44;
50
51static const size_t kMaxStackMallocSize = 1 << 16;  // 64K
52static const uintptr_t kCurrentStackFrameMagic = 0x41B58AB3;
53static const uintptr_t kRetiredStackFrameMagic = 0x45E0360E;
54
55static const char *kAsanModuleCtorName = "asan.module_ctor";
56static const char *kAsanModuleDtorName = "asan.module_dtor";
57static const int   kAsanCtorAndCtorPriority = 1;
58static const char *kAsanReportErrorTemplate = "__asan_report_";
59static const char *kAsanRegisterGlobalsName = "__asan_register_globals";
60static const char *kAsanUnregisterGlobalsName = "__asan_unregister_globals";
61static const char *kAsanInitName = "__asan_init";
62static const char *kAsanHandleNoReturnName = "__asan_handle_no_return";
63static const char *kAsanMappingOffsetName = "__asan_mapping_offset";
64static const char *kAsanMappingScaleName = "__asan_mapping_scale";
65static const char *kAsanStackMallocName = "__asan_stack_malloc";
66static const char *kAsanStackFreeName = "__asan_stack_free";
67
68static const int kAsanStackLeftRedzoneMagic = 0xf1;
69static const int kAsanStackMidRedzoneMagic = 0xf2;
70static const int kAsanStackRightRedzoneMagic = 0xf3;
71static const int kAsanStackPartialRedzoneMagic = 0xf4;
72
73// Command-line flags.
74
75// This flag may need to be replaced with -f[no-]asan-reads.
76static cl::opt<bool> ClInstrumentReads("asan-instrument-reads",
77       cl::desc("instrument read instructions"), cl::Hidden, cl::init(true));
78static cl::opt<bool> ClInstrumentWrites("asan-instrument-writes",
79       cl::desc("instrument write instructions"), cl::Hidden, cl::init(true));
80// This flag may need to be replaced with -f[no]asan-stack.
81static cl::opt<bool> ClStack("asan-stack",
82       cl::desc("Handle stack memory"), cl::Hidden, cl::init(true));
83// This flag may need to be replaced with -f[no]asan-use-after-return.
84static cl::opt<bool> ClUseAfterReturn("asan-use-after-return",
85       cl::desc("Check return-after-free"), cl::Hidden, cl::init(false));
86// This flag may need to be replaced with -f[no]asan-globals.
87static cl::opt<bool> ClGlobals("asan-globals",
88       cl::desc("Handle global objects"), cl::Hidden, cl::init(true));
89static cl::opt<bool> ClMemIntrin("asan-memintrin",
90       cl::desc("Handle memset/memcpy/memmove"), cl::Hidden, cl::init(true));
91// This flag may need to be replaced with -fasan-blacklist.
92static cl::opt<std::string>  ClBlackListFile("asan-blacklist",
93       cl::desc("File containing the list of functions to ignore "
94                "during instrumentation"), cl::Hidden);
95
96// These flags allow to change the shadow mapping.
97// The shadow mapping looks like
98//    Shadow = (Mem >> scale) + (1 << offset_log)
99static cl::opt<int> ClMappingScale("asan-mapping-scale",
100       cl::desc("scale of asan shadow mapping"), cl::Hidden, cl::init(0));
101static cl::opt<int> ClMappingOffsetLog("asan-mapping-offset-log",
102       cl::desc("offset of asan shadow mapping"), cl::Hidden, cl::init(-1));
103
104// Optimization flags. Not user visible, used mostly for testing
105// and benchmarking the tool.
106static cl::opt<bool> ClOpt("asan-opt",
107       cl::desc("Optimize instrumentation"), cl::Hidden, cl::init(true));
108static cl::opt<bool> ClOptSameTemp("asan-opt-same-temp",
109       cl::desc("Instrument the same temp just once"), cl::Hidden,
110       cl::init(true));
111static cl::opt<bool> ClOptGlobals("asan-opt-globals",
112       cl::desc("Don't instrument scalar globals"), cl::Hidden, cl::init(true));
113
114// Debug flags.
115static cl::opt<int> ClDebug("asan-debug", cl::desc("debug"), cl::Hidden,
116                            cl::init(0));
117static cl::opt<int> ClDebugStack("asan-debug-stack", cl::desc("debug stack"),
118                                 cl::Hidden, cl::init(0));
119static cl::opt<std::string> ClDebugFunc("asan-debug-func",
120                                        cl::Hidden, cl::desc("Debug func"));
121static cl::opt<int> ClDebugMin("asan-debug-min", cl::desc("Debug min inst"),
122                               cl::Hidden, cl::init(-1));
123static cl::opt<int> ClDebugMax("asan-debug-max", cl::desc("Debug man inst"),
124                               cl::Hidden, cl::init(-1));
125
126namespace {
127
128/// AddressSanitizer: instrument the code in module to find memory bugs.
129struct AddressSanitizer : public ModulePass {
130  AddressSanitizer();
131  virtual const char *getPassName() const;
132  void instrumentMop(Instruction *I);
133  void instrumentAddress(Instruction *OrigIns, IRBuilder<> &IRB,
134                         Value *Addr, uint32_t TypeSize, bool IsWrite);
135  Instruction *generateCrashCode(IRBuilder<> &IRB, Value *Addr,
136                                 bool IsWrite, uint32_t TypeSize);
137  bool instrumentMemIntrinsic(MemIntrinsic *MI);
138  void instrumentMemIntrinsicParam(Instruction *OrigIns, Value *Addr,
139                                  Value *Size,
140                                   Instruction *InsertBefore, bool IsWrite);
141  Value *memToShadow(Value *Shadow, IRBuilder<> &IRB);
142  bool handleFunction(Module &M, Function &F);
143  bool maybeInsertAsanInitAtFunctionEntry(Function &F);
144  bool poisonStackInFunction(Module &M, Function &F);
145  virtual bool runOnModule(Module &M);
146  bool insertGlobalRedzones(Module &M);
147  BranchInst *splitBlockAndInsertIfThen(Instruction *SplitBefore, Value *Cmp);
148  static char ID;  // Pass identification, replacement for typeid
149
150 private:
151
152  uint64_t getAllocaSizeInBytes(AllocaInst *AI) {
153    Type *Ty = AI->getAllocatedType();
154    uint64_t SizeInBytes = TD->getTypeAllocSize(Ty);
155    return SizeInBytes;
156  }
157  uint64_t getAlignedSize(uint64_t SizeInBytes) {
158    return ((SizeInBytes + RedzoneSize - 1)
159            / RedzoneSize) * RedzoneSize;
160  }
161  uint64_t getAlignedAllocaSize(AllocaInst *AI) {
162    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
163    return getAlignedSize(SizeInBytes);
164  }
165
166  void PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec, IRBuilder<> IRB,
167                   Value *ShadowBase, bool DoPoison);
168  bool LooksLikeCodeInBug11395(Instruction *I);
169
170  Module      *CurrentModule;
171  LLVMContext *C;
172  TargetData *TD;
173  uint64_t MappingOffset;
174  int MappingScale;
175  size_t RedzoneSize;
176  int LongSize;
177  Type *IntptrTy;
178  Type *IntptrPtrTy;
179  Function *AsanCtorFunction;
180  Function *AsanInitFunction;
181  Instruction *CtorInsertBefore;
182  OwningPtr<FunctionBlackList> BL;
183};
184}  // namespace
185
186char AddressSanitizer::ID = 0;
187INITIALIZE_PASS(AddressSanitizer, "asan",
188    "AddressSanitizer: detects use-after-free and out-of-bounds bugs.",
189    false, false)
190AddressSanitizer::AddressSanitizer() : ModulePass(ID) { }
191ModulePass *llvm::createAddressSanitizerPass() {
192  return new AddressSanitizer();
193}
194
195const char *AddressSanitizer::getPassName() const {
196  return "AddressSanitizer";
197}
198
199// Create a constant for Str so that we can pass it to the run-time lib.
200static GlobalVariable *createPrivateGlobalForString(Module &M, StringRef Str) {
201  Constant *StrConst = ConstantDataArray::getString(M.getContext(), Str);
202  return new GlobalVariable(M, StrConst->getType(), true,
203                            GlobalValue::PrivateLinkage, StrConst, "");
204}
205
206// Split the basic block and insert an if-then code.
207// Before:
208//   Head
209//   SplitBefore
210//   Tail
211// After:
212//   Head
213//   if (Cmp)
214//     NewBasicBlock
215//   SplitBefore
216//   Tail
217//
218// Returns the NewBasicBlock's terminator.
219BranchInst *AddressSanitizer::splitBlockAndInsertIfThen(
220    Instruction *SplitBefore, Value *Cmp) {
221  BasicBlock *Head = SplitBefore->getParent();
222  BasicBlock *Tail = Head->splitBasicBlock(SplitBefore);
223  TerminatorInst *HeadOldTerm = Head->getTerminator();
224  BasicBlock *NewBasicBlock =
225      BasicBlock::Create(*C, "", Head->getParent());
226  BranchInst *HeadNewTerm = BranchInst::Create(/*ifTrue*/NewBasicBlock,
227                                               /*ifFalse*/Tail,
228                                               Cmp);
229  ReplaceInstWithInst(HeadOldTerm, HeadNewTerm);
230
231  BranchInst *CheckTerm = BranchInst::Create(Tail, NewBasicBlock);
232  return CheckTerm;
233}
234
235Value *AddressSanitizer::memToShadow(Value *Shadow, IRBuilder<> &IRB) {
236  // Shadow >> scale
237  Shadow = IRB.CreateLShr(Shadow, MappingScale);
238  if (MappingOffset == 0)
239    return Shadow;
240  // (Shadow >> scale) | offset
241  return IRB.CreateOr(Shadow, ConstantInt::get(IntptrTy,
242                                               MappingOffset));
243}
244
245void AddressSanitizer::instrumentMemIntrinsicParam(Instruction *OrigIns,
246    Value *Addr, Value *Size, Instruction *InsertBefore, bool IsWrite) {
247  // Check the first byte.
248  {
249    IRBuilder<> IRB(InsertBefore);
250    instrumentAddress(OrigIns, IRB, Addr, 8, IsWrite);
251  }
252  // Check the last byte.
253  {
254    IRBuilder<> IRB(InsertBefore);
255    Value *SizeMinusOne = IRB.CreateSub(
256        Size, ConstantInt::get(Size->getType(), 1));
257    SizeMinusOne = IRB.CreateIntCast(SizeMinusOne, IntptrTy, false);
258    Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
259    Value *AddrPlusSizeMinisOne = IRB.CreateAdd(AddrLong, SizeMinusOne);
260    instrumentAddress(OrigIns, IRB, AddrPlusSizeMinisOne, 8, IsWrite);
261  }
262}
263
264// Instrument memset/memmove/memcpy
265bool AddressSanitizer::instrumentMemIntrinsic(MemIntrinsic *MI) {
266  Value *Dst = MI->getDest();
267  MemTransferInst *MemTran = dyn_cast<MemTransferInst>(MI);
268  Value *Src = MemTran ? MemTran->getSource() : NULL;
269  Value *Length = MI->getLength();
270
271  Constant *ConstLength = dyn_cast<Constant>(Length);
272  Instruction *InsertBefore = MI;
273  if (ConstLength) {
274    if (ConstLength->isNullValue()) return false;
275  } else {
276    // The size is not a constant so it could be zero -- check at run-time.
277    IRBuilder<> IRB(InsertBefore);
278
279    Value *Cmp = IRB.CreateICmpNE(Length,
280                                   Constant::getNullValue(Length->getType()));
281    InsertBefore = splitBlockAndInsertIfThen(InsertBefore, Cmp);
282  }
283
284  instrumentMemIntrinsicParam(MI, Dst, Length, InsertBefore, true);
285  if (Src)
286    instrumentMemIntrinsicParam(MI, Src, Length, InsertBefore, false);
287  return true;
288}
289
290static Value *getLDSTOperand(Instruction *I) {
291  if (LoadInst *LI = dyn_cast<LoadInst>(I)) {
292    return LI->getPointerOperand();
293  }
294  return cast<StoreInst>(*I).getPointerOperand();
295}
296
297void AddressSanitizer::instrumentMop(Instruction *I) {
298  int IsWrite = isa<StoreInst>(*I);
299  Value *Addr = getLDSTOperand(I);
300  if (ClOpt && ClOptGlobals && isa<GlobalVariable>(Addr)) {
301    // We are accessing a global scalar variable. Nothing to catch here.
302    return;
303  }
304  Type *OrigPtrTy = Addr->getType();
305  Type *OrigTy = cast<PointerType>(OrigPtrTy)->getElementType();
306
307  assert(OrigTy->isSized());
308  uint32_t TypeSize = TD->getTypeStoreSizeInBits(OrigTy);
309
310  if (TypeSize != 8  && TypeSize != 16 &&
311      TypeSize != 32 && TypeSize != 64 && TypeSize != 128) {
312    // Ignore all unusual sizes.
313    return;
314  }
315
316  IRBuilder<> IRB(I);
317  instrumentAddress(I, IRB, Addr, TypeSize, IsWrite);
318}
319
320Instruction *AddressSanitizer::generateCrashCode(
321    IRBuilder<> &IRB, Value *Addr, bool IsWrite, uint32_t TypeSize) {
322  // IsWrite and TypeSize are encoded in the function name.
323  std::string FunctionName = std::string(kAsanReportErrorTemplate) +
324      (IsWrite ? "store" : "load") + itostr(TypeSize / 8);
325  Value *ReportWarningFunc = CurrentModule->getOrInsertFunction(
326      FunctionName, IRB.getVoidTy(), IntptrTy, NULL);
327  CallInst *Call = IRB.CreateCall(ReportWarningFunc, Addr);
328  Call->setDoesNotReturn();
329  return Call;
330}
331
332void AddressSanitizer::instrumentAddress(Instruction *OrigIns,
333                                         IRBuilder<> &IRB, Value *Addr,
334                                         uint32_t TypeSize, bool IsWrite) {
335  Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
336
337  Type *ShadowTy  = IntegerType::get(
338      *C, std::max(8U, TypeSize >> MappingScale));
339  Type *ShadowPtrTy = PointerType::get(ShadowTy, 0);
340  Value *ShadowPtr = memToShadow(AddrLong, IRB);
341  Value *CmpVal = Constant::getNullValue(ShadowTy);
342  Value *ShadowValue = IRB.CreateLoad(
343      IRB.CreateIntToPtr(ShadowPtr, ShadowPtrTy));
344
345  Value *Cmp = IRB.CreateICmpNE(ShadowValue, CmpVal);
346
347  Instruction *CheckTerm = splitBlockAndInsertIfThen(
348      cast<Instruction>(Cmp)->getNextNode(), Cmp);
349  IRBuilder<> IRB2(CheckTerm);
350
351  size_t Granularity = 1 << MappingScale;
352  if (TypeSize < 8 * Granularity) {
353    // Addr & (Granularity - 1)
354    Value *Lower3Bits = IRB2.CreateAnd(
355        AddrLong, ConstantInt::get(IntptrTy, Granularity - 1));
356    // (Addr & (Granularity - 1)) + size - 1
357    Value *LastAccessedByte = IRB2.CreateAdd(
358        Lower3Bits, ConstantInt::get(IntptrTy, TypeSize / 8 - 1));
359    // (uint8_t) ((Addr & (Granularity-1)) + size - 1)
360    LastAccessedByte = IRB2.CreateIntCast(
361        LastAccessedByte, IRB.getInt8Ty(), false);
362    // ((uint8_t) ((Addr & (Granularity-1)) + size - 1)) >= ShadowValue
363    Value *Cmp2 = IRB2.CreateICmpSGE(LastAccessedByte, ShadowValue);
364
365    CheckTerm = splitBlockAndInsertIfThen(CheckTerm, Cmp2);
366  }
367
368  IRBuilder<> IRB1(CheckTerm);
369  Instruction *Crash = generateCrashCode(IRB1, AddrLong, IsWrite, TypeSize);
370  Crash->setDebugLoc(OrigIns->getDebugLoc());
371  ReplaceInstWithInst(CheckTerm, new UnreachableInst(*C));
372}
373
374// This function replaces all global variables with new variables that have
375// trailing redzones. It also creates a function that poisons
376// redzones and inserts this function into llvm.global_ctors.
377bool AddressSanitizer::insertGlobalRedzones(Module &M) {
378  SmallVector<GlobalVariable *, 16> GlobalsToChange;
379
380  for (Module::GlobalListType::iterator G = M.getGlobalList().begin(),
381       E = M.getGlobalList().end(); G != E; ++G) {
382    Type *Ty = cast<PointerType>(G->getType())->getElementType();
383    DEBUG(dbgs() << "GLOBAL: " << *G);
384
385    if (!Ty->isSized()) continue;
386    if (!G->hasInitializer()) continue;
387    // Touch only those globals that will not be defined in other modules.
388    // Don't handle ODR type linkages since other modules may be built w/o asan.
389    if (G->getLinkage() != GlobalVariable::ExternalLinkage &&
390        G->getLinkage() != GlobalVariable::PrivateLinkage &&
391        G->getLinkage() != GlobalVariable::InternalLinkage)
392      continue;
393    // Two problems with thread-locals:
394    //   - The address of the main thread's copy can't be computed at link-time.
395    //   - Need to poison all copies, not just the main thread's one.
396    if (G->isThreadLocal())
397      continue;
398    // For now, just ignore this Alloca if the alignment is large.
399    if (G->getAlignment() > RedzoneSize) continue;
400
401    // Ignore all the globals with the names starting with "\01L_OBJC_".
402    // Many of those are put into the .cstring section. The linker compresses
403    // that section by removing the spare \0s after the string terminator, so
404    // our redzones get broken.
405    if ((G->getName().find("\01L_OBJC_") == 0) ||
406        (G->getName().find("\01l_OBJC_") == 0)) {
407      DEBUG(dbgs() << "Ignoring \\01L_OBJC_* global: " << *G);
408      continue;
409    }
410
411    if (G->hasSection()) {
412      StringRef Section(G->getSection());
413      // Ignore the globals from the __OBJC section. The ObjC runtime assumes
414      // those conform to /usr/lib/objc/runtime.h, so we can't add redzones to
415      // them.
416      if ((Section.find("__OBJC,") == 0) ||
417          (Section.find("__DATA, __objc_") == 0)) {
418        DEBUG(dbgs() << "Ignoring ObjC runtime global: " << *G);
419        continue;
420      }
421      // See http://code.google.com/p/address-sanitizer/issues/detail?id=32
422      // Constant CFString instances are compiled in the following way:
423      //  -- the string buffer is emitted into
424      //     __TEXT,__cstring,cstring_literals
425      //  -- the constant NSConstantString structure referencing that buffer
426      //     is placed into __DATA,__cfstring
427      // Therefore there's no point in placing redzones into __DATA,__cfstring.
428      // Moreover, it causes the linker to crash on OS X 10.7
429      if (Section.find("__DATA,__cfstring") == 0) {
430        DEBUG(dbgs() << "Ignoring CFString: " << *G);
431        continue;
432      }
433    }
434
435    GlobalsToChange.push_back(G);
436  }
437
438  size_t n = GlobalsToChange.size();
439  if (n == 0) return false;
440
441  // A global is described by a structure
442  //   size_t beg;
443  //   size_t size;
444  //   size_t size_with_redzone;
445  //   const char *name;
446  // We initialize an array of such structures and pass it to a run-time call.
447  StructType *GlobalStructTy = StructType::get(IntptrTy, IntptrTy,
448                                               IntptrTy, IntptrTy, NULL);
449  SmallVector<Constant *, 16> Initializers(n);
450
451  IRBuilder<> IRB(CtorInsertBefore);
452
453  for (size_t i = 0; i < n; i++) {
454    GlobalVariable *G = GlobalsToChange[i];
455    PointerType *PtrTy = cast<PointerType>(G->getType());
456    Type *Ty = PtrTy->getElementType();
457    uint64_t SizeInBytes = TD->getTypeStoreSizeInBits(Ty) / 8;
458    uint64_t RightRedzoneSize = RedzoneSize +
459        (RedzoneSize - (SizeInBytes % RedzoneSize));
460    Type *RightRedZoneTy = ArrayType::get(IRB.getInt8Ty(), RightRedzoneSize);
461
462    StructType *NewTy = StructType::get(Ty, RightRedZoneTy, NULL);
463    Constant *NewInitializer = ConstantStruct::get(
464        NewTy, G->getInitializer(),
465        Constant::getNullValue(RightRedZoneTy), NULL);
466
467    SmallString<2048> DescriptionOfGlobal = G->getName();
468    DescriptionOfGlobal += " (";
469    DescriptionOfGlobal += M.getModuleIdentifier();
470    DescriptionOfGlobal += ")";
471    GlobalVariable *Name = createPrivateGlobalForString(M, DescriptionOfGlobal);
472
473    // Create a new global variable with enough space for a redzone.
474    GlobalVariable *NewGlobal = new GlobalVariable(
475        M, NewTy, G->isConstant(), G->getLinkage(),
476        NewInitializer, "", G, G->isThreadLocal());
477    NewGlobal->copyAttributesFrom(G);
478    NewGlobal->setAlignment(RedzoneSize);
479
480    Value *Indices2[2];
481    Indices2[0] = IRB.getInt32(0);
482    Indices2[1] = IRB.getInt32(0);
483
484    G->replaceAllUsesWith(
485        ConstantExpr::getGetElementPtr(NewGlobal, Indices2, true));
486    NewGlobal->takeName(G);
487    G->eraseFromParent();
488
489    Initializers[i] = ConstantStruct::get(
490        GlobalStructTy,
491        ConstantExpr::getPointerCast(NewGlobal, IntptrTy),
492        ConstantInt::get(IntptrTy, SizeInBytes),
493        ConstantInt::get(IntptrTy, SizeInBytes + RightRedzoneSize),
494        ConstantExpr::getPointerCast(Name, IntptrTy),
495        NULL);
496    DEBUG(dbgs() << "NEW GLOBAL:\n" << *NewGlobal);
497  }
498
499  ArrayType *ArrayOfGlobalStructTy = ArrayType::get(GlobalStructTy, n);
500  GlobalVariable *AllGlobals = new GlobalVariable(
501      M, ArrayOfGlobalStructTy, false, GlobalVariable::PrivateLinkage,
502      ConstantArray::get(ArrayOfGlobalStructTy, Initializers), "");
503
504  Function *AsanRegisterGlobals = cast<Function>(M.getOrInsertFunction(
505      kAsanRegisterGlobalsName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
506  AsanRegisterGlobals->setLinkage(Function::ExternalLinkage);
507
508  IRB.CreateCall2(AsanRegisterGlobals,
509                  IRB.CreatePointerCast(AllGlobals, IntptrTy),
510                  ConstantInt::get(IntptrTy, n));
511
512  // We also need to unregister globals at the end, e.g. when a shared library
513  // gets closed.
514  Function *AsanDtorFunction = Function::Create(
515      FunctionType::get(Type::getVoidTy(*C), false),
516      GlobalValue::InternalLinkage, kAsanModuleDtorName, &M);
517  BasicBlock *AsanDtorBB = BasicBlock::Create(*C, "", AsanDtorFunction);
518  IRBuilder<> IRB_Dtor(ReturnInst::Create(*C, AsanDtorBB));
519  Function *AsanUnregisterGlobals = cast<Function>(M.getOrInsertFunction(
520      kAsanUnregisterGlobalsName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
521  AsanUnregisterGlobals->setLinkage(Function::ExternalLinkage);
522
523  IRB_Dtor.CreateCall2(AsanUnregisterGlobals,
524                       IRB.CreatePointerCast(AllGlobals, IntptrTy),
525                       ConstantInt::get(IntptrTy, n));
526  appendToGlobalDtors(M, AsanDtorFunction, kAsanCtorAndCtorPriority);
527
528  DEBUG(dbgs() << M);
529  return true;
530}
531
532// virtual
533bool AddressSanitizer::runOnModule(Module &M) {
534  // Initialize the private fields. No one has accessed them before.
535  TD = getAnalysisIfAvailable<TargetData>();
536  if (!TD)
537    return false;
538  BL.reset(new FunctionBlackList(ClBlackListFile));
539
540  CurrentModule = &M;
541  C = &(M.getContext());
542  LongSize = TD->getPointerSizeInBits();
543  IntptrTy = Type::getIntNTy(*C, LongSize);
544  IntptrPtrTy = PointerType::get(IntptrTy, 0);
545
546  AsanCtorFunction = Function::Create(
547      FunctionType::get(Type::getVoidTy(*C), false),
548      GlobalValue::InternalLinkage, kAsanModuleCtorName, &M);
549  BasicBlock *AsanCtorBB = BasicBlock::Create(*C, "", AsanCtorFunction);
550  CtorInsertBefore = ReturnInst::Create(*C, AsanCtorBB);
551
552  // call __asan_init in the module ctor.
553  IRBuilder<> IRB(CtorInsertBefore);
554  AsanInitFunction = cast<Function>(
555      M.getOrInsertFunction(kAsanInitName, IRB.getVoidTy(), NULL));
556  AsanInitFunction->setLinkage(Function::ExternalLinkage);
557  IRB.CreateCall(AsanInitFunction);
558
559  MappingOffset = LongSize == 32
560      ? kDefaultShadowOffset32 : kDefaultShadowOffset64;
561  if (ClMappingOffsetLog >= 0) {
562    if (ClMappingOffsetLog == 0) {
563      // special case
564      MappingOffset = 0;
565    } else {
566      MappingOffset = 1ULL << ClMappingOffsetLog;
567    }
568  }
569  MappingScale = kDefaultShadowScale;
570  if (ClMappingScale) {
571    MappingScale = ClMappingScale;
572  }
573  // Redzone used for stack and globals is at least 32 bytes.
574  // For scales 6 and 7, the redzone has to be 64 and 128 bytes respectively.
575  RedzoneSize = std::max(32, (int)(1 << MappingScale));
576
577  bool Res = false;
578
579  if (ClGlobals)
580    Res |= insertGlobalRedzones(M);
581
582  // Tell the run-time the current values of mapping offset and scale.
583  GlobalValue *asan_mapping_offset =
584      new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
585                     ConstantInt::get(IntptrTy, MappingOffset),
586                     kAsanMappingOffsetName);
587  GlobalValue *asan_mapping_scale =
588      new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
589                         ConstantInt::get(IntptrTy, MappingScale),
590                         kAsanMappingScaleName);
591  // Read these globals, otherwise they may be optimized away.
592  IRB.CreateLoad(asan_mapping_scale, true);
593  IRB.CreateLoad(asan_mapping_offset, true);
594
595
596  for (Module::iterator F = M.begin(), E = M.end(); F != E; ++F) {
597    if (F->isDeclaration()) continue;
598    Res |= handleFunction(M, *F);
599  }
600
601  appendToGlobalCtors(M, AsanCtorFunction, kAsanCtorAndCtorPriority);
602
603  return Res;
604}
605
606bool AddressSanitizer::maybeInsertAsanInitAtFunctionEntry(Function &F) {
607  // For each NSObject descendant having a +load method, this method is invoked
608  // by the ObjC runtime before any of the static constructors is called.
609  // Therefore we need to instrument such methods with a call to __asan_init
610  // at the beginning in order to initialize our runtime before any access to
611  // the shadow memory.
612  // We cannot just ignore these methods, because they may call other
613  // instrumented functions.
614  if (F.getName().find(" load]") != std::string::npos) {
615    IRBuilder<> IRB(F.begin()->begin());
616    IRB.CreateCall(AsanInitFunction);
617    return true;
618  }
619  return false;
620}
621
622bool AddressSanitizer::handleFunction(Module &M, Function &F) {
623  if (BL->isIn(F)) return false;
624  if (&F == AsanCtorFunction) return false;
625
626  // If needed, insert __asan_init before checking for AddressSafety attr.
627  maybeInsertAsanInitAtFunctionEntry(F);
628
629  if (!F.hasFnAttr(Attribute::AddressSafety)) return false;
630
631  if (!ClDebugFunc.empty() && ClDebugFunc != F.getName())
632    return false;
633  // We want to instrument every address only once per basic block
634  // (unless there are calls between uses).
635  SmallSet<Value*, 16> TempsToInstrument;
636  SmallVector<Instruction*, 16> ToInstrument;
637  SmallVector<Instruction*, 8> NoReturnCalls;
638
639  // Fill the set of memory operations to instrument.
640  for (Function::iterator FI = F.begin(), FE = F.end();
641       FI != FE; ++FI) {
642    TempsToInstrument.clear();
643    for (BasicBlock::iterator BI = FI->begin(), BE = FI->end();
644         BI != BE; ++BI) {
645      if (LooksLikeCodeInBug11395(BI)) return false;
646      if ((isa<LoadInst>(BI) && ClInstrumentReads) ||
647          (isa<StoreInst>(BI) && ClInstrumentWrites)) {
648        Value *Addr = getLDSTOperand(BI);
649        if (ClOpt && ClOptSameTemp) {
650          if (!TempsToInstrument.insert(Addr))
651            continue;  // We've seen this temp in the current BB.
652        }
653      } else if (isa<MemIntrinsic>(BI) && ClMemIntrin) {
654        // ok, take it.
655      } else {
656        if (CallInst *CI = dyn_cast<CallInst>(BI)) {
657          // A call inside BB.
658          TempsToInstrument.clear();
659          if (CI->doesNotReturn()) {
660            NoReturnCalls.push_back(CI);
661          }
662        }
663        continue;
664      }
665      ToInstrument.push_back(BI);
666    }
667  }
668
669  // Instrument.
670  int NumInstrumented = 0;
671  for (size_t i = 0, n = ToInstrument.size(); i != n; i++) {
672    Instruction *Inst = ToInstrument[i];
673    if (ClDebugMin < 0 || ClDebugMax < 0 ||
674        (NumInstrumented >= ClDebugMin && NumInstrumented <= ClDebugMax)) {
675      if (isa<StoreInst>(Inst) || isa<LoadInst>(Inst))
676        instrumentMop(Inst);
677      else
678        instrumentMemIntrinsic(cast<MemIntrinsic>(Inst));
679    }
680    NumInstrumented++;
681  }
682
683  DEBUG(dbgs() << F);
684
685  bool ChangedStack = poisonStackInFunction(M, F);
686
687  // We must unpoison the stack before every NoReturn call (throw, _exit, etc).
688  // See e.g. http://code.google.com/p/address-sanitizer/issues/detail?id=37
689  for (size_t i = 0, n = NoReturnCalls.size(); i != n; i++) {
690    Instruction *CI = NoReturnCalls[i];
691    IRBuilder<> IRB(CI);
692    IRB.CreateCall(M.getOrInsertFunction(kAsanHandleNoReturnName,
693                                         IRB.getVoidTy(), NULL));
694  }
695
696  return NumInstrumented > 0 || ChangedStack || !NoReturnCalls.empty();
697}
698
699static uint64_t ValueForPoison(uint64_t PoisonByte, size_t ShadowRedzoneSize) {
700  if (ShadowRedzoneSize == 1) return PoisonByte;
701  if (ShadowRedzoneSize == 2) return (PoisonByte << 8) + PoisonByte;
702  if (ShadowRedzoneSize == 4)
703    return (PoisonByte << 24) + (PoisonByte << 16) +
704        (PoisonByte << 8) + (PoisonByte);
705  llvm_unreachable("ShadowRedzoneSize is either 1, 2 or 4");
706}
707
708static void PoisonShadowPartialRightRedzone(uint8_t *Shadow,
709                                            size_t Size,
710                                            size_t RedzoneSize,
711                                            size_t ShadowGranularity,
712                                            uint8_t Magic) {
713  for (size_t i = 0; i < RedzoneSize;
714       i+= ShadowGranularity, Shadow++) {
715    if (i + ShadowGranularity <= Size) {
716      *Shadow = 0;  // fully addressable
717    } else if (i >= Size) {
718      *Shadow = Magic;  // unaddressable
719    } else {
720      *Shadow = Size - i;  // first Size-i bytes are addressable
721    }
722  }
723}
724
725void AddressSanitizer::PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec,
726                                   IRBuilder<> IRB,
727                                   Value *ShadowBase, bool DoPoison) {
728  size_t ShadowRZSize = RedzoneSize >> MappingScale;
729  assert(ShadowRZSize >= 1 && ShadowRZSize <= 4);
730  Type *RZTy = Type::getIntNTy(*C, ShadowRZSize * 8);
731  Type *RZPtrTy = PointerType::get(RZTy, 0);
732
733  Value *PoisonLeft  = ConstantInt::get(RZTy,
734    ValueForPoison(DoPoison ? kAsanStackLeftRedzoneMagic : 0LL, ShadowRZSize));
735  Value *PoisonMid   = ConstantInt::get(RZTy,
736    ValueForPoison(DoPoison ? kAsanStackMidRedzoneMagic : 0LL, ShadowRZSize));
737  Value *PoisonRight = ConstantInt::get(RZTy,
738    ValueForPoison(DoPoison ? kAsanStackRightRedzoneMagic : 0LL, ShadowRZSize));
739
740  // poison the first red zone.
741  IRB.CreateStore(PoisonLeft, IRB.CreateIntToPtr(ShadowBase, RZPtrTy));
742
743  // poison all other red zones.
744  uint64_t Pos = RedzoneSize;
745  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
746    AllocaInst *AI = AllocaVec[i];
747    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
748    uint64_t AlignedSize = getAlignedAllocaSize(AI);
749    assert(AlignedSize - SizeInBytes < RedzoneSize);
750    Value *Ptr = NULL;
751
752    Pos += AlignedSize;
753
754    assert(ShadowBase->getType() == IntptrTy);
755    if (SizeInBytes < AlignedSize) {
756      // Poison the partial redzone at right
757      Ptr = IRB.CreateAdd(
758          ShadowBase, ConstantInt::get(IntptrTy,
759                                       (Pos >> MappingScale) - ShadowRZSize));
760      size_t AddressableBytes = RedzoneSize - (AlignedSize - SizeInBytes);
761      uint32_t Poison = 0;
762      if (DoPoison) {
763        PoisonShadowPartialRightRedzone((uint8_t*)&Poison, AddressableBytes,
764                                        RedzoneSize,
765                                        1ULL << MappingScale,
766                                        kAsanStackPartialRedzoneMagic);
767      }
768      Value *PartialPoison = ConstantInt::get(RZTy, Poison);
769      IRB.CreateStore(PartialPoison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
770    }
771
772    // Poison the full redzone at right.
773    Ptr = IRB.CreateAdd(ShadowBase,
774                        ConstantInt::get(IntptrTy, Pos >> MappingScale));
775    Value *Poison = i == AllocaVec.size() - 1 ? PoisonRight : PoisonMid;
776    IRB.CreateStore(Poison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
777
778    Pos += RedzoneSize;
779  }
780}
781
782// Workaround for bug 11395: we don't want to instrument stack in functions
783// with large assembly blobs (32-bit only), otherwise reg alloc may crash.
784// FIXME: remove once the bug 11395 is fixed.
785bool AddressSanitizer::LooksLikeCodeInBug11395(Instruction *I) {
786  if (LongSize != 32) return false;
787  CallInst *CI = dyn_cast<CallInst>(I);
788  if (!CI || !CI->isInlineAsm()) return false;
789  if (CI->getNumArgOperands() <= 5) return false;
790  // We have inline assembly with quite a few arguments.
791  return true;
792}
793
794// Find all static Alloca instructions and put
795// poisoned red zones around all of them.
796// Then unpoison everything back before the function returns.
797//
798// Stack poisoning does not play well with exception handling.
799// When an exception is thrown, we essentially bypass the code
800// that unpoisones the stack. This is why the run-time library has
801// to intercept __cxa_throw (as well as longjmp, etc) and unpoison the entire
802// stack in the interceptor. This however does not work inside the
803// actual function which catches the exception. Most likely because the
804// compiler hoists the load of the shadow value somewhere too high.
805// This causes asan to report a non-existing bug on 453.povray.
806// It sounds like an LLVM bug.
807bool AddressSanitizer::poisonStackInFunction(Module &M, Function &F) {
808  if (!ClStack) return false;
809  SmallVector<AllocaInst*, 16> AllocaVec;
810  SmallVector<Instruction*, 8> RetVec;
811  uint64_t TotalSize = 0;
812
813  // Filter out Alloca instructions we want (and can) handle.
814  // Collect Ret instructions.
815  for (Function::iterator FI = F.begin(), FE = F.end();
816       FI != FE; ++FI) {
817    BasicBlock &BB = *FI;
818    for (BasicBlock::iterator BI = BB.begin(), BE = BB.end();
819         BI != BE; ++BI) {
820      if (isa<ReturnInst>(BI)) {
821          RetVec.push_back(BI);
822          continue;
823      }
824
825      AllocaInst *AI = dyn_cast<AllocaInst>(BI);
826      if (!AI) continue;
827      if (AI->isArrayAllocation()) continue;
828      if (!AI->isStaticAlloca()) continue;
829      if (!AI->getAllocatedType()->isSized()) continue;
830      if (AI->getAlignment() > RedzoneSize) continue;
831      AllocaVec.push_back(AI);
832      uint64_t AlignedSize =  getAlignedAllocaSize(AI);
833      TotalSize += AlignedSize;
834    }
835  }
836
837  if (AllocaVec.empty()) return false;
838
839  uint64_t LocalStackSize = TotalSize + (AllocaVec.size() + 1) * RedzoneSize;
840
841  bool DoStackMalloc = ClUseAfterReturn
842      && LocalStackSize <= kMaxStackMallocSize;
843
844  Instruction *InsBefore = AllocaVec[0];
845  IRBuilder<> IRB(InsBefore);
846
847
848  Type *ByteArrayTy = ArrayType::get(IRB.getInt8Ty(), LocalStackSize);
849  AllocaInst *MyAlloca =
850      new AllocaInst(ByteArrayTy, "MyAlloca", InsBefore);
851  MyAlloca->setAlignment(RedzoneSize);
852  assert(MyAlloca->isStaticAlloca());
853  Value *OrigStackBase = IRB.CreatePointerCast(MyAlloca, IntptrTy);
854  Value *LocalStackBase = OrigStackBase;
855
856  if (DoStackMalloc) {
857    Value *AsanStackMallocFunc = M.getOrInsertFunction(
858        kAsanStackMallocName, IntptrTy, IntptrTy, IntptrTy, NULL);
859    LocalStackBase = IRB.CreateCall2(AsanStackMallocFunc,
860        ConstantInt::get(IntptrTy, LocalStackSize), OrigStackBase);
861  }
862
863  // This string will be parsed by the run-time (DescribeStackAddress).
864  SmallString<2048> StackDescriptionStorage;
865  raw_svector_ostream StackDescription(StackDescriptionStorage);
866  StackDescription << F.getName() << " " << AllocaVec.size() << " ";
867
868  uint64_t Pos = RedzoneSize;
869  // Replace Alloca instructions with base+offset.
870  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
871    AllocaInst *AI = AllocaVec[i];
872    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
873    StringRef Name = AI->getName();
874    StackDescription << Pos << " " << SizeInBytes << " "
875                     << Name.size() << " " << Name << " ";
876    uint64_t AlignedSize = getAlignedAllocaSize(AI);
877    assert((AlignedSize % RedzoneSize) == 0);
878    AI->replaceAllUsesWith(
879        IRB.CreateIntToPtr(
880            IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy, Pos)),
881            AI->getType()));
882    Pos += AlignedSize + RedzoneSize;
883  }
884  assert(Pos == LocalStackSize);
885
886  // Write the Magic value and the frame description constant to the redzone.
887  Value *BasePlus0 = IRB.CreateIntToPtr(LocalStackBase, IntptrPtrTy);
888  IRB.CreateStore(ConstantInt::get(IntptrTy, kCurrentStackFrameMagic),
889                  BasePlus0);
890  Value *BasePlus1 = IRB.CreateAdd(LocalStackBase,
891                                   ConstantInt::get(IntptrTy, LongSize/8));
892  BasePlus1 = IRB.CreateIntToPtr(BasePlus1, IntptrPtrTy);
893  Value *Description = IRB.CreatePointerCast(
894      createPrivateGlobalForString(M, StackDescription.str()),
895      IntptrTy);
896  IRB.CreateStore(Description, BasePlus1);
897
898  // Poison the stack redzones at the entry.
899  Value *ShadowBase = memToShadow(LocalStackBase, IRB);
900  PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRB, ShadowBase, true);
901
902  Value *AsanStackFreeFunc = NULL;
903  if (DoStackMalloc) {
904    AsanStackFreeFunc = M.getOrInsertFunction(
905        kAsanStackFreeName, IRB.getVoidTy(),
906        IntptrTy, IntptrTy, IntptrTy, NULL);
907  }
908
909  // Unpoison the stack before all ret instructions.
910  for (size_t i = 0, n = RetVec.size(); i < n; i++) {
911    Instruction *Ret = RetVec[i];
912    IRBuilder<> IRBRet(Ret);
913
914    // Mark the current frame as retired.
915    IRBRet.CreateStore(ConstantInt::get(IntptrTy, kRetiredStackFrameMagic),
916                       BasePlus0);
917    // Unpoison the stack.
918    PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRBRet, ShadowBase, false);
919
920    if (DoStackMalloc) {
921      IRBRet.CreateCall3(AsanStackFreeFunc, LocalStackBase,
922                         ConstantInt::get(IntptrTy, LocalStackSize),
923                         OrigStackBase);
924    }
925  }
926
927  if (ClDebugStack) {
928    DEBUG(dbgs() << F);
929  }
930
931  return true;
932}
933