AddressSanitizer.cpp revision 06fdbaa9145f01a291d4981ca5120b7bdcad44c6 
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/ADT/Triple.h"
26#include "llvm/Function.h"
27#include "llvm/IntrinsicInst.h"
28#include "llvm/LLVMContext.h"
29#include "llvm/Module.h"
30#include "llvm/Support/CommandLine.h"
31#include "llvm/Support/DataTypes.h"
32#include "llvm/Support/Debug.h"
33#include "llvm/Support/IRBuilder.h"
34#include "llvm/Support/raw_ostream.h"
35#include "llvm/Support/system_error.h"
36#include "llvm/Target/TargetData.h"
37#include "llvm/Target/TargetMachine.h"
38#include "llvm/Transforms/Instrumentation.h"
39#include "llvm/Transforms/Utils/BasicBlockUtils.h"
40#include "llvm/Transforms/Utils/ModuleUtils.h"
41#include "llvm/Type.h"
42
43#include <string>
44#include <algorithm>
45
46using namespace llvm;
47
48static const uint64_t kDefaultShadowScale = 3;
49static const uint64_t kDefaultShadowOffset32 = 1ULL << 29;
50static const uint64_t kDefaultShadowOffset64 = 1ULL << 44;
51static const uint64_t kDefaultShadowOffsetAndroid = 0;
52
53static const size_t kMaxStackMallocSize = 1 << 16;  // 64K
54static const uintptr_t kCurrentStackFrameMagic = 0x41B58AB3;
55static const uintptr_t kRetiredStackFrameMagic = 0x45E0360E;
56
57static const char *kAsanModuleCtorName = "asan.module_ctor";
58static const char *kAsanModuleDtorName = "asan.module_dtor";
59static const int   kAsanCtorAndCtorPriority = 1;
60static const char *kAsanReportErrorTemplate = "__asan_report_";
61static const char *kAsanRegisterGlobalsName = "__asan_register_globals";
62static const char *kAsanUnregisterGlobalsName = "__asan_unregister_globals";
63static const char *kAsanInitName = "__asan_init";
64static const char *kAsanHandleNoReturnName = "__asan_handle_no_return";
65static const char *kAsanMappingOffsetName = "__asan_mapping_offset";
66static const char *kAsanMappingScaleName = "__asan_mapping_scale";
67static const char *kAsanStackMallocName = "__asan_stack_malloc";
68static const char *kAsanStackFreeName = "__asan_stack_free";
69
70static const int kAsanStackLeftRedzoneMagic = 0xf1;
71static const int kAsanStackMidRedzoneMagic = 0xf2;
72static const int kAsanStackRightRedzoneMagic = 0xf3;
73static const int kAsanStackPartialRedzoneMagic = 0xf4;
74
75// Command-line flags.
76
77// This flag may need to be replaced with -f[no-]asan-reads.
78static cl::opt<bool> ClInstrumentReads("asan-instrument-reads",
79       cl::desc("instrument read instructions"), cl::Hidden, cl::init(true));
80static cl::opt<bool> ClInstrumentWrites("asan-instrument-writes",
81       cl::desc("instrument write instructions"), cl::Hidden, cl::init(true));
82// This flag may need to be replaced with -f[no]asan-stack.
83static cl::opt<bool> ClStack("asan-stack",
84       cl::desc("Handle stack memory"), cl::Hidden, cl::init(true));
85// This flag may need to be replaced with -f[no]asan-use-after-return.
86static cl::opt<bool> ClUseAfterReturn("asan-use-after-return",
87       cl::desc("Check return-after-free"), cl::Hidden, cl::init(false));
88// This flag may need to be replaced with -f[no]asan-globals.
89static cl::opt<bool> ClGlobals("asan-globals",
90       cl::desc("Handle global objects"), cl::Hidden, cl::init(true));
91static cl::opt<bool> ClMemIntrin("asan-memintrin",
92       cl::desc("Handle memset/memcpy/memmove"), cl::Hidden, cl::init(true));
93// This flag may need to be replaced with -fasan-blacklist.
94static cl::opt<std::string>  ClBlackListFile("asan-blacklist",
95       cl::desc("File containing the list of functions to ignore "
96                "during instrumentation"), cl::Hidden);
97
98// These flags allow to change the shadow mapping.
99// The shadow mapping looks like
100//    Shadow = (Mem >> scale) + (1 << offset_log)
101static cl::opt<int> ClMappingScale("asan-mapping-scale",
102       cl::desc("scale of asan shadow mapping"), cl::Hidden, cl::init(0));
103static cl::opt<int> ClMappingOffsetLog("asan-mapping-offset-log",
104       cl::desc("offset of asan shadow mapping"), cl::Hidden, cl::init(-1));
105
106// Optimization flags. Not user visible, used mostly for testing
107// and benchmarking the tool.
108static cl::opt<bool> ClOpt("asan-opt",
109       cl::desc("Optimize instrumentation"), cl::Hidden, cl::init(true));
110static cl::opt<bool> ClOptSameTemp("asan-opt-same-temp",
111       cl::desc("Instrument the same temp just once"), cl::Hidden,
112       cl::init(true));
113static cl::opt<bool> ClOptGlobals("asan-opt-globals",
114       cl::desc("Don't instrument scalar globals"), cl::Hidden, cl::init(true));
115
116// Debug flags.
117static cl::opt<int> ClDebug("asan-debug", cl::desc("debug"), cl::Hidden,
118                            cl::init(0));
119static cl::opt<int> ClDebugStack("asan-debug-stack", cl::desc("debug stack"),
120                                 cl::Hidden, cl::init(0));
121static cl::opt<std::string> ClDebugFunc("asan-debug-func",
122                                        cl::Hidden, cl::desc("Debug func"));
123static cl::opt<int> ClDebugMin("asan-debug-min", cl::desc("Debug min inst"),
124                               cl::Hidden, cl::init(-1));
125static cl::opt<int> ClDebugMax("asan-debug-max", cl::desc("Debug man inst"),
126                               cl::Hidden, cl::init(-1));
127
128namespace {
129
130/// AddressSanitizer: instrument the code in module to find memory bugs.
131struct AddressSanitizer : public ModulePass {
132  AddressSanitizer();
133  virtual const char *getPassName() const;
134  void instrumentMop(Instruction *I);
135  void instrumentAddress(Instruction *OrigIns, IRBuilder<> &IRB,
136                         Value *Addr, uint32_t TypeSize, bool IsWrite);
137  Instruction *generateCrashCode(IRBuilder<> &IRB, Value *Addr,
138                                 bool IsWrite, uint32_t TypeSize);
139  bool instrumentMemIntrinsic(MemIntrinsic *MI);
140  void instrumentMemIntrinsicParam(Instruction *OrigIns, Value *Addr,
141                                  Value *Size,
142                                   Instruction *InsertBefore, bool IsWrite);
143  Value *memToShadow(Value *Shadow, IRBuilder<> &IRB);
144  bool handleFunction(Module &M, Function &F);
145  bool maybeInsertAsanInitAtFunctionEntry(Function &F);
146  bool poisonStackInFunction(Module &M, Function &F);
147  virtual bool runOnModule(Module &M);
148  bool insertGlobalRedzones(Module &M);
149  BranchInst *splitBlockAndInsertIfThen(Instruction *SplitBefore, Value *Cmp);
150  static char ID;  // Pass identification, replacement for typeid
151
152 private:
153
154  uint64_t getAllocaSizeInBytes(AllocaInst *AI) {
155    Type *Ty = AI->getAllocatedType();
156    uint64_t SizeInBytes = TD->getTypeAllocSize(Ty);
157    return SizeInBytes;
158  }
159  uint64_t getAlignedSize(uint64_t SizeInBytes) {
160    return ((SizeInBytes + RedzoneSize - 1)
161            / RedzoneSize) * RedzoneSize;
162  }
163  uint64_t getAlignedAllocaSize(AllocaInst *AI) {
164    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
165    return getAlignedSize(SizeInBytes);
166  }
167
168  Function *checkInterfaceFunction(Constant *FuncOrBitcast);
169  void PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec, IRBuilder<> IRB,
170                   Value *ShadowBase, bool DoPoison);
171  bool LooksLikeCodeInBug11395(Instruction *I);
172
173  Module      *CurrentModule;
174  LLVMContext *C;
175  TargetData *TD;
176  uint64_t MappingOffset;
177  int MappingScale;
178  size_t RedzoneSize;
179  int LongSize;
180  Type *IntptrTy;
181  Type *IntptrPtrTy;
182  Function *AsanCtorFunction;
183  Function *AsanInitFunction;
184  Instruction *CtorInsertBefore;
185  OwningPtr<FunctionBlackList> BL;
186};
187}  // namespace
188
189char AddressSanitizer::ID = 0;
190INITIALIZE_PASS(AddressSanitizer, "asan",
191    "AddressSanitizer: detects use-after-free and out-of-bounds bugs.",
192    false, false)
193AddressSanitizer::AddressSanitizer() : ModulePass(ID) { }
194ModulePass *llvm::createAddressSanitizerPass() {
195  return new AddressSanitizer();
196}
197
198const char *AddressSanitizer::getPassName() const {
199  return "AddressSanitizer";
200}
201
202// Create a constant for Str so that we can pass it to the run-time lib.
203static GlobalVariable *createPrivateGlobalForString(Module &M, StringRef Str) {
204  Constant *StrConst = ConstantDataArray::getString(M.getContext(), Str);
205  return new GlobalVariable(M, StrConst->getType(), true,
206                            GlobalValue::PrivateLinkage, StrConst, "");
207}
208
209// Split the basic block and insert an if-then code.
210// Before:
211//   Head
212//   SplitBefore
213//   Tail
214// After:
215//   Head
216//   if (Cmp)
217//     NewBasicBlock
218//   SplitBefore
219//   Tail
220//
221// Returns the NewBasicBlock's terminator.
222BranchInst *AddressSanitizer::splitBlockAndInsertIfThen(
223    Instruction *SplitBefore, Value *Cmp) {
224  BasicBlock *Head = SplitBefore->getParent();
225  BasicBlock *Tail = Head->splitBasicBlock(SplitBefore);
226  TerminatorInst *HeadOldTerm = Head->getTerminator();
227  BasicBlock *NewBasicBlock =
228      BasicBlock::Create(*C, "", Head->getParent());
229  BranchInst *HeadNewTerm = BranchInst::Create(/*ifTrue*/NewBasicBlock,
230                                               /*ifFalse*/Tail,
231                                               Cmp);
232  ReplaceInstWithInst(HeadOldTerm, HeadNewTerm);
233
234  BranchInst *CheckTerm = BranchInst::Create(Tail, NewBasicBlock);
235  return CheckTerm;
236}
237
238Value *AddressSanitizer::memToShadow(Value *Shadow, IRBuilder<> &IRB) {
239  // Shadow >> scale
240  Shadow = IRB.CreateLShr(Shadow, MappingScale);
241  if (MappingOffset == 0)
242    return Shadow;
243  // (Shadow >> scale) | offset
244  return IRB.CreateOr(Shadow, ConstantInt::get(IntptrTy,
245                                               MappingOffset));
246}
247
248void AddressSanitizer::instrumentMemIntrinsicParam(Instruction *OrigIns,
249    Value *Addr, Value *Size, Instruction *InsertBefore, bool IsWrite) {
250  // Check the first byte.
251  {
252    IRBuilder<> IRB(InsertBefore);
253    instrumentAddress(OrigIns, IRB, Addr, 8, IsWrite);
254  }
255  // Check the last byte.
256  {
257    IRBuilder<> IRB(InsertBefore);
258    Value *SizeMinusOne = IRB.CreateSub(
259        Size, ConstantInt::get(Size->getType(), 1));
260    SizeMinusOne = IRB.CreateIntCast(SizeMinusOne, IntptrTy, false);
261    Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
262    Value *AddrPlusSizeMinisOne = IRB.CreateAdd(AddrLong, SizeMinusOne);
263    instrumentAddress(OrigIns, IRB, AddrPlusSizeMinisOne, 8, IsWrite);
264  }
265}
266
267// Instrument memset/memmove/memcpy
268bool AddressSanitizer::instrumentMemIntrinsic(MemIntrinsic *MI) {
269  Value *Dst = MI->getDest();
270  MemTransferInst *MemTran = dyn_cast<MemTransferInst>(MI);
271  Value *Src = MemTran ? MemTran->getSource() : NULL;
272  Value *Length = MI->getLength();
273
274  Constant *ConstLength = dyn_cast<Constant>(Length);
275  Instruction *InsertBefore = MI;
276  if (ConstLength) {
277    if (ConstLength->isNullValue()) return false;
278  } else {
279    // The size is not a constant so it could be zero -- check at run-time.
280    IRBuilder<> IRB(InsertBefore);
281
282    Value *Cmp = IRB.CreateICmpNE(Length,
283                                   Constant::getNullValue(Length->getType()));
284    InsertBefore = splitBlockAndInsertIfThen(InsertBefore, Cmp);
285  }
286
287  instrumentMemIntrinsicParam(MI, Dst, Length, InsertBefore, true);
288  if (Src)
289    instrumentMemIntrinsicParam(MI, Src, Length, InsertBefore, false);
290  return true;
291}
292
293static Value *getLDSTOperand(Instruction *I) {
294  if (LoadInst *LI = dyn_cast<LoadInst>(I)) {
295    return LI->getPointerOperand();
296  }
297  return cast<StoreInst>(*I).getPointerOperand();
298}
299
300void AddressSanitizer::instrumentMop(Instruction *I) {
301  int IsWrite = isa<StoreInst>(*I);
302  Value *Addr = getLDSTOperand(I);
303  if (ClOpt && ClOptGlobals && isa<GlobalVariable>(Addr)) {
304    // We are accessing a global scalar variable. Nothing to catch here.
305    return;
306  }
307  Type *OrigPtrTy = Addr->getType();
308  Type *OrigTy = cast<PointerType>(OrigPtrTy)->getElementType();
309
310  assert(OrigTy->isSized());
311  uint32_t TypeSize = TD->getTypeStoreSizeInBits(OrigTy);
312
313  if (TypeSize != 8  && TypeSize != 16 &&
314      TypeSize != 32 && TypeSize != 64 && TypeSize != 128) {
315    // Ignore all unusual sizes.
316    return;
317  }
318
319  IRBuilder<> IRB(I);
320  instrumentAddress(I, IRB, Addr, TypeSize, IsWrite);
321}
322
323// Validate the result of Module::getOrInsertFunction called for an interface
324// function of AddressSanitizer. If the instrumented module defines a function
325// with the same name, their prototypes must match, otherwise
326// getOrInsertFunction returns a bitcast.
327Function *AddressSanitizer::checkInterfaceFunction(Constant *FuncOrBitcast) {
328  if (isa<Function>(FuncOrBitcast)) return cast<Function>(FuncOrBitcast);
329  FuncOrBitcast->dump();
330  report_fatal_error("trying to redefine an AddressSanitizer "
331                     "interface function");
332}
333
334Instruction *AddressSanitizer::generateCrashCode(
335    IRBuilder<> &IRB, Value *Addr, bool IsWrite, uint32_t TypeSize) {
336  // IsWrite and TypeSize are encoded in the function name.
337  std::string FunctionName = std::string(kAsanReportErrorTemplate) +
338      (IsWrite ? "store" : "load") + itostr(TypeSize / 8);
339  Value *ReportWarningFunc = CurrentModule->getOrInsertFunction(
340      FunctionName, IRB.getVoidTy(), IntptrTy, NULL);
341  CallInst *Call = IRB.CreateCall(ReportWarningFunc, Addr);
342  Call->setDoesNotReturn();
343  return Call;
344}
345
346void AddressSanitizer::instrumentAddress(Instruction *OrigIns,
347                                         IRBuilder<> &IRB, Value *Addr,
348                                         uint32_t TypeSize, bool IsWrite) {
349  Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
350
351  Type *ShadowTy  = IntegerType::get(
352      *C, std::max(8U, TypeSize >> MappingScale));
353  Type *ShadowPtrTy = PointerType::get(ShadowTy, 0);
354  Value *ShadowPtr = memToShadow(AddrLong, IRB);
355  Value *CmpVal = Constant::getNullValue(ShadowTy);
356  Value *ShadowValue = IRB.CreateLoad(
357      IRB.CreateIntToPtr(ShadowPtr, ShadowPtrTy));
358
359  Value *Cmp = IRB.CreateICmpNE(ShadowValue, CmpVal);
360
361  Instruction *CheckTerm = splitBlockAndInsertIfThen(
362      cast<Instruction>(Cmp)->getNextNode(), Cmp);
363  IRBuilder<> IRB2(CheckTerm);
364
365  size_t Granularity = 1 << MappingScale;
366  if (TypeSize < 8 * Granularity) {
367    // Addr & (Granularity - 1)
368    Value *LastAccessedByte = IRB2.CreateAnd(
369        AddrLong, ConstantInt::get(IntptrTy, Granularity - 1));
370    // (Addr & (Granularity - 1)) + size - 1
371    if (TypeSize / 8 > 1)
372      LastAccessedByte = IRB2.CreateAdd(
373          LastAccessedByte, ConstantInt::get(IntptrTy, TypeSize / 8 - 1));
374    // (uint8_t) ((Addr & (Granularity-1)) + size - 1)
375    LastAccessedByte = IRB2.CreateIntCast(
376        LastAccessedByte, IRB.getInt8Ty(), false);
377    // ((uint8_t) ((Addr & (Granularity-1)) + size - 1)) >= ShadowValue
378    Value *Cmp2 = IRB2.CreateICmpSGE(LastAccessedByte, ShadowValue);
379
380    CheckTerm = splitBlockAndInsertIfThen(CheckTerm, Cmp2);
381  }
382
383  IRBuilder<> IRB1(CheckTerm);
384  Instruction *Crash = generateCrashCode(IRB1, AddrLong, IsWrite, TypeSize);
385  Crash->setDebugLoc(OrigIns->getDebugLoc());
386  ReplaceInstWithInst(CheckTerm, new UnreachableInst(*C));
387}
388
389// This function replaces all global variables with new variables that have
390// trailing redzones. It also creates a function that poisons
391// redzones and inserts this function into llvm.global_ctors.
392bool AddressSanitizer::insertGlobalRedzones(Module &M) {
393  SmallVector<GlobalVariable *, 16> GlobalsToChange;
394
395  for (Module::GlobalListType::iterator G = M.getGlobalList().begin(),
396       E = M.getGlobalList().end(); G != E; ++G) {
397    Type *Ty = cast<PointerType>(G->getType())->getElementType();
398    DEBUG(dbgs() << "GLOBAL: " << *G);
399
400    if (!Ty->isSized()) continue;
401    if (!G->hasInitializer()) continue;
402    // Touch only those globals that will not be defined in other modules.
403    // Don't handle ODR type linkages since other modules may be built w/o asan.
404    if (G->getLinkage() != GlobalVariable::ExternalLinkage &&
405        G->getLinkage() != GlobalVariable::PrivateLinkage &&
406        G->getLinkage() != GlobalVariable::InternalLinkage)
407      continue;
408    // Two problems with thread-locals:
409    //   - The address of the main thread's copy can't be computed at link-time.
410    //   - Need to poison all copies, not just the main thread's one.
411    if (G->isThreadLocal())
412      continue;
413    // For now, just ignore this Alloca if the alignment is large.
414    if (G->getAlignment() > RedzoneSize) continue;
415
416    // Ignore all the globals with the names starting with "\01L_OBJC_".
417    // Many of those are put into the .cstring section. The linker compresses
418    // that section by removing the spare \0s after the string terminator, so
419    // our redzones get broken.
420    if ((G->getName().find("\01L_OBJC_") == 0) ||
421        (G->getName().find("\01l_OBJC_") == 0)) {
422      DEBUG(dbgs() << "Ignoring \\01L_OBJC_* global: " << *G);
423      continue;
424    }
425
426    if (G->hasSection()) {
427      StringRef Section(G->getSection());
428      // Ignore the globals from the __OBJC section. The ObjC runtime assumes
429      // those conform to /usr/lib/objc/runtime.h, so we can't add redzones to
430      // them.
431      if ((Section.find("__OBJC,") == 0) ||
432          (Section.find("__DATA, __objc_") == 0)) {
433        DEBUG(dbgs() << "Ignoring ObjC runtime global: " << *G);
434        continue;
435      }
436      // See http://code.google.com/p/address-sanitizer/issues/detail?id=32
437      // Constant CFString instances are compiled in the following way:
438      //  -- the string buffer is emitted into
439      //     __TEXT,__cstring,cstring_literals
440      //  -- the constant NSConstantString structure referencing that buffer
441      //     is placed into __DATA,__cfstring
442      // Therefore there's no point in placing redzones into __DATA,__cfstring.
443      // Moreover, it causes the linker to crash on OS X 10.7
444      if (Section.find("__DATA,__cfstring") == 0) {
445        DEBUG(dbgs() << "Ignoring CFString: " << *G);
446        continue;
447      }
448    }
449
450    GlobalsToChange.push_back(G);
451  }
452
453  size_t n = GlobalsToChange.size();
454  if (n == 0) return false;
455
456  // A global is described by a structure
457  //   size_t beg;
458  //   size_t size;
459  //   size_t size_with_redzone;
460  //   const char *name;
461  // We initialize an array of such structures and pass it to a run-time call.
462  StructType *GlobalStructTy = StructType::get(IntptrTy, IntptrTy,
463                                               IntptrTy, IntptrTy, NULL);
464  SmallVector<Constant *, 16> Initializers(n);
465
466  IRBuilder<> IRB(CtorInsertBefore);
467
468  for (size_t i = 0; i < n; i++) {
469    GlobalVariable *G = GlobalsToChange[i];
470    PointerType *PtrTy = cast<PointerType>(G->getType());
471    Type *Ty = PtrTy->getElementType();
472    uint64_t SizeInBytes = TD->getTypeAllocSize(Ty);
473    uint64_t RightRedzoneSize = RedzoneSize +
474        (RedzoneSize - (SizeInBytes % RedzoneSize));
475    Type *RightRedZoneTy = ArrayType::get(IRB.getInt8Ty(), RightRedzoneSize);
476
477    StructType *NewTy = StructType::get(Ty, RightRedZoneTy, NULL);
478    Constant *NewInitializer = ConstantStruct::get(
479        NewTy, G->getInitializer(),
480        Constant::getNullValue(RightRedZoneTy), NULL);
481
482    SmallString<2048> DescriptionOfGlobal = G->getName();
483    DescriptionOfGlobal += " (";
484    DescriptionOfGlobal += M.getModuleIdentifier();
485    DescriptionOfGlobal += ")";
486    GlobalVariable *Name = createPrivateGlobalForString(M, DescriptionOfGlobal);
487
488    // Create a new global variable with enough space for a redzone.
489    GlobalVariable *NewGlobal = new GlobalVariable(
490        M, NewTy, G->isConstant(), G->getLinkage(),
491        NewInitializer, "", G, G->isThreadLocal());
492    NewGlobal->copyAttributesFrom(G);
493    NewGlobal->setAlignment(RedzoneSize);
494
495    Value *Indices2[2];
496    Indices2[0] = IRB.getInt32(0);
497    Indices2[1] = IRB.getInt32(0);
498
499    G->replaceAllUsesWith(
500        ConstantExpr::getGetElementPtr(NewGlobal, Indices2, true));
501    NewGlobal->takeName(G);
502    G->eraseFromParent();
503
504    Initializers[i] = ConstantStruct::get(
505        GlobalStructTy,
506        ConstantExpr::getPointerCast(NewGlobal, IntptrTy),
507        ConstantInt::get(IntptrTy, SizeInBytes),
508        ConstantInt::get(IntptrTy, SizeInBytes + RightRedzoneSize),
509        ConstantExpr::getPointerCast(Name, IntptrTy),
510        NULL);
511    DEBUG(dbgs() << "NEW GLOBAL:\n" << *NewGlobal);
512  }
513
514  ArrayType *ArrayOfGlobalStructTy = ArrayType::get(GlobalStructTy, n);
515  GlobalVariable *AllGlobals = new GlobalVariable(
516      M, ArrayOfGlobalStructTy, false, GlobalVariable::PrivateLinkage,
517      ConstantArray::get(ArrayOfGlobalStructTy, Initializers), "");
518
519  Function *AsanRegisterGlobals = checkInterfaceFunction(M.getOrInsertFunction(
520      kAsanRegisterGlobalsName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
521  AsanRegisterGlobals->setLinkage(Function::ExternalLinkage);
522
523  IRB.CreateCall2(AsanRegisterGlobals,
524                  IRB.CreatePointerCast(AllGlobals, IntptrTy),
525                  ConstantInt::get(IntptrTy, n));
526
527  // We also need to unregister globals at the end, e.g. when a shared library
528  // gets closed.
529  Function *AsanDtorFunction = Function::Create(
530      FunctionType::get(Type::getVoidTy(*C), false),
531      GlobalValue::InternalLinkage, kAsanModuleDtorName, &M);
532  BasicBlock *AsanDtorBB = BasicBlock::Create(*C, "", AsanDtorFunction);
533  IRBuilder<> IRB_Dtor(ReturnInst::Create(*C, AsanDtorBB));
534  Function *AsanUnregisterGlobals =
535      checkInterfaceFunction(M.getOrInsertFunction(
536          kAsanUnregisterGlobalsName,
537          IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
538  AsanUnregisterGlobals->setLinkage(Function::ExternalLinkage);
539
540  IRB_Dtor.CreateCall2(AsanUnregisterGlobals,
541                       IRB.CreatePointerCast(AllGlobals, IntptrTy),
542                       ConstantInt::get(IntptrTy, n));
543  appendToGlobalDtors(M, AsanDtorFunction, kAsanCtorAndCtorPriority);
544
545  DEBUG(dbgs() << M);
546  return true;
547}
548
549// virtual
550bool AddressSanitizer::runOnModule(Module &M) {
551  // Initialize the private fields. No one has accessed them before.
552  TD = getAnalysisIfAvailable<TargetData>();
553  if (!TD)
554    return false;
555  BL.reset(new FunctionBlackList(ClBlackListFile));
556
557  CurrentModule = &M;
558  C = &(M.getContext());
559  LongSize = TD->getPointerSizeInBits();
560  IntptrTy = Type::getIntNTy(*C, LongSize);
561  IntptrPtrTy = PointerType::get(IntptrTy, 0);
562
563  AsanCtorFunction = Function::Create(
564      FunctionType::get(Type::getVoidTy(*C), false),
565      GlobalValue::InternalLinkage, kAsanModuleCtorName, &M);
566  BasicBlock *AsanCtorBB = BasicBlock::Create(*C, "", AsanCtorFunction);
567  CtorInsertBefore = ReturnInst::Create(*C, AsanCtorBB);
568
569  // call __asan_init in the module ctor.
570  IRBuilder<> IRB(CtorInsertBefore);
571  AsanInitFunction = checkInterfaceFunction(
572      M.getOrInsertFunction(kAsanInitName, IRB.getVoidTy(), NULL));
573  AsanInitFunction->setLinkage(Function::ExternalLinkage);
574  IRB.CreateCall(AsanInitFunction);
575
576  llvm::Triple targetTriple(M.getTargetTriple());
577  bool isAndroid = targetTriple.getEnvironment() == llvm::Triple::ANDROIDEABI;
578
579  MappingOffset = isAndroid ? kDefaultShadowOffsetAndroid :
580    (LongSize == 32 ? kDefaultShadowOffset32 : kDefaultShadowOffset64);
581  if (ClMappingOffsetLog >= 0) {
582    if (ClMappingOffsetLog == 0) {
583      // special case
584      MappingOffset = 0;
585    } else {
586      MappingOffset = 1ULL << ClMappingOffsetLog;
587    }
588  }
589  MappingScale = kDefaultShadowScale;
590  if (ClMappingScale) {
591    MappingScale = ClMappingScale;
592  }
593  // Redzone used for stack and globals is at least 32 bytes.
594  // For scales 6 and 7, the redzone has to be 64 and 128 bytes respectively.
595  RedzoneSize = std::max(32, (int)(1 << MappingScale));
596
597  bool Res = false;
598
599  if (ClGlobals)
600    Res |= insertGlobalRedzones(M);
601
602  if (ClMappingOffsetLog >= 0) {
603    // Tell the run-time the current values of mapping offset and scale.
604    GlobalValue *asan_mapping_offset =
605        new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
606                       ConstantInt::get(IntptrTy, MappingOffset),
607                       kAsanMappingOffsetName);
608    // Read the global, otherwise it may be optimized away.
609    IRB.CreateLoad(asan_mapping_offset, true);
610  }
611  if (ClMappingScale) {
612    GlobalValue *asan_mapping_scale =
613        new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
614                           ConstantInt::get(IntptrTy, MappingScale),
615                           kAsanMappingScaleName);
616    // Read the global, otherwise it may be optimized away.
617    IRB.CreateLoad(asan_mapping_scale, true);
618  }
619
620
621  for (Module::iterator F = M.begin(), E = M.end(); F != E; ++F) {
622    if (F->isDeclaration()) continue;
623    Res |= handleFunction(M, *F);
624  }
625
626  appendToGlobalCtors(M, AsanCtorFunction, kAsanCtorAndCtorPriority);
627
628  return Res;
629}
630
631bool AddressSanitizer::maybeInsertAsanInitAtFunctionEntry(Function &F) {
632  // For each NSObject descendant having a +load method, this method is invoked
633  // by the ObjC runtime before any of the static constructors is called.
634  // Therefore we need to instrument such methods with a call to __asan_init
635  // at the beginning in order to initialize our runtime before any access to
636  // the shadow memory.
637  // We cannot just ignore these methods, because they may call other
638  // instrumented functions.
639  if (F.getName().find(" load]") != std::string::npos) {
640    IRBuilder<> IRB(F.begin()->begin());
641    IRB.CreateCall(AsanInitFunction);
642    return true;
643  }
644  return false;
645}
646
647bool AddressSanitizer::handleFunction(Module &M, Function &F) {
648  if (BL->isIn(F)) return false;
649  if (&F == AsanCtorFunction) return false;
650
651  // If needed, insert __asan_init before checking for AddressSafety attr.
652  maybeInsertAsanInitAtFunctionEntry(F);
653
654  if (!F.hasFnAttr(Attribute::AddressSafety)) return false;
655
656  if (!ClDebugFunc.empty() && ClDebugFunc != F.getName())
657    return false;
658  // We want to instrument every address only once per basic block
659  // (unless there are calls between uses).
660  SmallSet<Value*, 16> TempsToInstrument;
661  SmallVector<Instruction*, 16> ToInstrument;
662  SmallVector<Instruction*, 8> NoReturnCalls;
663
664  // Fill the set of memory operations to instrument.
665  for (Function::iterator FI = F.begin(), FE = F.end();
666       FI != FE; ++FI) {
667    TempsToInstrument.clear();
668    for (BasicBlock::iterator BI = FI->begin(), BE = FI->end();
669         BI != BE; ++BI) {
670      if (LooksLikeCodeInBug11395(BI)) return false;
671      if ((isa<LoadInst>(BI) && ClInstrumentReads) ||
672          (isa<StoreInst>(BI) && ClInstrumentWrites)) {
673        Value *Addr = getLDSTOperand(BI);
674        if (ClOpt && ClOptSameTemp) {
675          if (!TempsToInstrument.insert(Addr))
676            continue;  // We've seen this temp in the current BB.
677        }
678      } else if (isa<MemIntrinsic>(BI) && ClMemIntrin) {
679        // ok, take it.
680      } else {
681        if (CallInst *CI = dyn_cast<CallInst>(BI)) {
682          // A call inside BB.
683          TempsToInstrument.clear();
684          if (CI->doesNotReturn()) {
685            NoReturnCalls.push_back(CI);
686          }
687        }
688        continue;
689      }
690      ToInstrument.push_back(BI);
691    }
692  }
693
694  // Instrument.
695  int NumInstrumented = 0;
696  for (size_t i = 0, n = ToInstrument.size(); i != n; i++) {
697    Instruction *Inst = ToInstrument[i];
698    if (ClDebugMin < 0 || ClDebugMax < 0 ||
699        (NumInstrumented >= ClDebugMin && NumInstrumented <= ClDebugMax)) {
700      if (isa<StoreInst>(Inst) || isa<LoadInst>(Inst))
701        instrumentMop(Inst);
702      else
703        instrumentMemIntrinsic(cast<MemIntrinsic>(Inst));
704    }
705    NumInstrumented++;
706  }
707
708  DEBUG(dbgs() << F);
709
710  bool ChangedStack = poisonStackInFunction(M, F);
711
712  // We must unpoison the stack before every NoReturn call (throw, _exit, etc).
713  // See e.g. http://code.google.com/p/address-sanitizer/issues/detail?id=37
714  for (size_t i = 0, n = NoReturnCalls.size(); i != n; i++) {
715    Instruction *CI = NoReturnCalls[i];
716    IRBuilder<> IRB(CI);
717    IRB.CreateCall(M.getOrInsertFunction(kAsanHandleNoReturnName,
718                                         IRB.getVoidTy(), NULL));
719  }
720
721  return NumInstrumented > 0 || ChangedStack || !NoReturnCalls.empty();
722}
723
724static uint64_t ValueForPoison(uint64_t PoisonByte, size_t ShadowRedzoneSize) {
725  if (ShadowRedzoneSize == 1) return PoisonByte;
726  if (ShadowRedzoneSize == 2) return (PoisonByte << 8) + PoisonByte;
727  if (ShadowRedzoneSize == 4)
728    return (PoisonByte << 24) + (PoisonByte << 16) +
729        (PoisonByte << 8) + (PoisonByte);
730  llvm_unreachable("ShadowRedzoneSize is either 1, 2 or 4");
731}
732
733static void PoisonShadowPartialRightRedzone(uint8_t *Shadow,
734                                            size_t Size,
735                                            size_t RedzoneSize,
736                                            size_t ShadowGranularity,
737                                            uint8_t Magic) {
738  for (size_t i = 0; i < RedzoneSize;
739       i+= ShadowGranularity, Shadow++) {
740    if (i + ShadowGranularity <= Size) {
741      *Shadow = 0;  // fully addressable
742    } else if (i >= Size) {
743      *Shadow = Magic;  // unaddressable
744    } else {
745      *Shadow = Size - i;  // first Size-i bytes are addressable
746    }
747  }
748}
749
750void AddressSanitizer::PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec,
751                                   IRBuilder<> IRB,
752                                   Value *ShadowBase, bool DoPoison) {
753  size_t ShadowRZSize = RedzoneSize >> MappingScale;
754  assert(ShadowRZSize >= 1 && ShadowRZSize <= 4);
755  Type *RZTy = Type::getIntNTy(*C, ShadowRZSize * 8);
756  Type *RZPtrTy = PointerType::get(RZTy, 0);
757
758  Value *PoisonLeft  = ConstantInt::get(RZTy,
759    ValueForPoison(DoPoison ? kAsanStackLeftRedzoneMagic : 0LL, ShadowRZSize));
760  Value *PoisonMid   = ConstantInt::get(RZTy,
761    ValueForPoison(DoPoison ? kAsanStackMidRedzoneMagic : 0LL, ShadowRZSize));
762  Value *PoisonRight = ConstantInt::get(RZTy,
763    ValueForPoison(DoPoison ? kAsanStackRightRedzoneMagic : 0LL, ShadowRZSize));
764
765  // poison the first red zone.
766  IRB.CreateStore(PoisonLeft, IRB.CreateIntToPtr(ShadowBase, RZPtrTy));
767
768  // poison all other red zones.
769  uint64_t Pos = RedzoneSize;
770  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
771    AllocaInst *AI = AllocaVec[i];
772    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
773    uint64_t AlignedSize = getAlignedAllocaSize(AI);
774    assert(AlignedSize - SizeInBytes < RedzoneSize);
775    Value *Ptr = NULL;
776
777    Pos += AlignedSize;
778
779    assert(ShadowBase->getType() == IntptrTy);
780    if (SizeInBytes < AlignedSize) {
781      // Poison the partial redzone at right
782      Ptr = IRB.CreateAdd(
783          ShadowBase, ConstantInt::get(IntptrTy,
784                                       (Pos >> MappingScale) - ShadowRZSize));
785      size_t AddressableBytes = RedzoneSize - (AlignedSize - SizeInBytes);
786      uint32_t Poison = 0;
787      if (DoPoison) {
788        PoisonShadowPartialRightRedzone((uint8_t*)&Poison, AddressableBytes,
789                                        RedzoneSize,
790                                        1ULL << MappingScale,
791                                        kAsanStackPartialRedzoneMagic);
792      }
793      Value *PartialPoison = ConstantInt::get(RZTy, Poison);
794      IRB.CreateStore(PartialPoison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
795    }
796
797    // Poison the full redzone at right.
798    Ptr = IRB.CreateAdd(ShadowBase,
799                        ConstantInt::get(IntptrTy, Pos >> MappingScale));
800    Value *Poison = i == AllocaVec.size() - 1 ? PoisonRight : PoisonMid;
801    IRB.CreateStore(Poison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
802
803    Pos += RedzoneSize;
804  }
805}
806
807// Workaround for bug 11395: we don't want to instrument stack in functions
808// with large assembly blobs (32-bit only), otherwise reg alloc may crash.
809// FIXME: remove once the bug 11395 is fixed.
810bool AddressSanitizer::LooksLikeCodeInBug11395(Instruction *I) {
811  if (LongSize != 32) return false;
812  CallInst *CI = dyn_cast<CallInst>(I);
813  if (!CI || !CI->isInlineAsm()) return false;
814  if (CI->getNumArgOperands() <= 5) return false;
815  // We have inline assembly with quite a few arguments.
816  return true;
817}
818
819// Find all static Alloca instructions and put
820// poisoned red zones around all of them.
821// Then unpoison everything back before the function returns.
822//
823// Stack poisoning does not play well with exception handling.
824// When an exception is thrown, we essentially bypass the code
825// that unpoisones the stack. This is why the run-time library has
826// to intercept __cxa_throw (as well as longjmp, etc) and unpoison the entire
827// stack in the interceptor. This however does not work inside the
828// actual function which catches the exception. Most likely because the
829// compiler hoists the load of the shadow value somewhere too high.
830// This causes asan to report a non-existing bug on 453.povray.
831// It sounds like an LLVM bug.
832bool AddressSanitizer::poisonStackInFunction(Module &M, Function &F) {
833  if (!ClStack) return false;
834  SmallVector<AllocaInst*, 16> AllocaVec;
835  SmallVector<Instruction*, 8> RetVec;
836  uint64_t TotalSize = 0;
837
838  // Filter out Alloca instructions we want (and can) handle.
839  // Collect Ret instructions.
840  for (Function::iterator FI = F.begin(), FE = F.end();
841       FI != FE; ++FI) {
842    BasicBlock &BB = *FI;
843    for (BasicBlock::iterator BI = BB.begin(), BE = BB.end();
844         BI != BE; ++BI) {
845      if (isa<ReturnInst>(BI)) {
846          RetVec.push_back(BI);
847          continue;
848      }
849
850      AllocaInst *AI = dyn_cast<AllocaInst>(BI);
851      if (!AI) continue;
852      if (AI->isArrayAllocation()) continue;
853      if (!AI->isStaticAlloca()) continue;
854      if (!AI->getAllocatedType()->isSized()) continue;
855      if (AI->getAlignment() > RedzoneSize) continue;
856      AllocaVec.push_back(AI);
857      uint64_t AlignedSize =  getAlignedAllocaSize(AI);
858      TotalSize += AlignedSize;
859    }
860  }
861
862  if (AllocaVec.empty()) return false;
863
864  uint64_t LocalStackSize = TotalSize + (AllocaVec.size() + 1) * RedzoneSize;
865
866  bool DoStackMalloc = ClUseAfterReturn
867      && LocalStackSize <= kMaxStackMallocSize;
868
869  Instruction *InsBefore = AllocaVec[0];
870  IRBuilder<> IRB(InsBefore);
871
872
873  Type *ByteArrayTy = ArrayType::get(IRB.getInt8Ty(), LocalStackSize);
874  AllocaInst *MyAlloca =
875      new AllocaInst(ByteArrayTy, "MyAlloca", InsBefore);
876  MyAlloca->setAlignment(RedzoneSize);
877  assert(MyAlloca->isStaticAlloca());
878  Value *OrigStackBase = IRB.CreatePointerCast(MyAlloca, IntptrTy);
879  Value *LocalStackBase = OrigStackBase;
880
881  if (DoStackMalloc) {
882    Value *AsanStackMallocFunc = M.getOrInsertFunction(
883        kAsanStackMallocName, IntptrTy, IntptrTy, IntptrTy, NULL);
884    LocalStackBase = IRB.CreateCall2(AsanStackMallocFunc,
885        ConstantInt::get(IntptrTy, LocalStackSize), OrigStackBase);
886  }
887
888  // This string will be parsed by the run-time (DescribeStackAddress).
889  SmallString<2048> StackDescriptionStorage;
890  raw_svector_ostream StackDescription(StackDescriptionStorage);
891  StackDescription << F.getName() << " " << AllocaVec.size() << " ";
892
893  uint64_t Pos = RedzoneSize;
894  // Replace Alloca instructions with base+offset.
895  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
896    AllocaInst *AI = AllocaVec[i];
897    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
898    StringRef Name = AI->getName();
899    StackDescription << Pos << " " << SizeInBytes << " "
900                     << Name.size() << " " << Name << " ";
901    uint64_t AlignedSize = getAlignedAllocaSize(AI);
902    assert((AlignedSize % RedzoneSize) == 0);
903    AI->replaceAllUsesWith(
904        IRB.CreateIntToPtr(
905            IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy, Pos)),
906            AI->getType()));
907    Pos += AlignedSize + RedzoneSize;
908  }
909  assert(Pos == LocalStackSize);
910
911  // Write the Magic value and the frame description constant to the redzone.
912  Value *BasePlus0 = IRB.CreateIntToPtr(LocalStackBase, IntptrPtrTy);
913  IRB.CreateStore(ConstantInt::get(IntptrTy, kCurrentStackFrameMagic),
914                  BasePlus0);
915  Value *BasePlus1 = IRB.CreateAdd(LocalStackBase,
916                                   ConstantInt::get(IntptrTy, LongSize/8));
917  BasePlus1 = IRB.CreateIntToPtr(BasePlus1, IntptrPtrTy);
918  Value *Description = IRB.CreatePointerCast(
919      createPrivateGlobalForString(M, StackDescription.str()),
920      IntptrTy);
921  IRB.CreateStore(Description, BasePlus1);
922
923  // Poison the stack redzones at the entry.
924  Value *ShadowBase = memToShadow(LocalStackBase, IRB);
925  PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRB, ShadowBase, true);
926
927  Value *AsanStackFreeFunc = NULL;
928  if (DoStackMalloc) {
929    AsanStackFreeFunc = M.getOrInsertFunction(
930        kAsanStackFreeName, IRB.getVoidTy(),
931        IntptrTy, IntptrTy, IntptrTy, NULL);
932  }
933
934  // Unpoison the stack before all ret instructions.
935  for (size_t i = 0, n = RetVec.size(); i < n; i++) {
936    Instruction *Ret = RetVec[i];
937    IRBuilder<> IRBRet(Ret);
938
939    // Mark the current frame as retired.
940    IRBRet.CreateStore(ConstantInt::get(IntptrTy, kRetiredStackFrameMagic),
941                       BasePlus0);
942    // Unpoison the stack.
943    PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRBRet, ShadowBase, false);
944
945    if (DoStackMalloc) {
946      IRBRet.CreateCall3(AsanStackFreeFunc, LocalStackBase,
947                         ConstantInt::get(IntptrTy, LocalStackSize),
948                         OrigStackBase);
949    }
950  }
951
952  if (ClDebugStack) {
953    DEBUG(dbgs() << F);
954  }
955
956  return true;
957}
958