AddressSanitizer.cpp revision 9db5b5ffa9fccd5c7f1f39a3e9aa66cc4a5eedc1 
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/Function.h"
20#include "llvm/IRBuilder.h"
21#include "llvm/IntrinsicInst.h"
22#include "llvm/LLVMContext.h"
23#include "llvm/Module.h"
24#include "llvm/Type.h"
25#include "llvm/ADT/ArrayRef.h"
26#include "llvm/ADT/OwningPtr.h"
27#include "llvm/ADT/SmallSet.h"
28#include "llvm/ADT/SmallString.h"
29#include "llvm/ADT/SmallVector.h"
30#include "llvm/ADT/StringExtras.h"
31#include "llvm/ADT/Triple.h"
32#include "llvm/Support/CommandLine.h"
33#include "llvm/Support/DataTypes.h"
34#include "llvm/Support/Debug.h"
35#include "llvm/Support/raw_ostream.h"
36#include "llvm/Support/system_error.h"
37#include "llvm/Target/TargetData.h"
38#include "llvm/Target/TargetMachine.h"
39#include "llvm/Transforms/Instrumentation.h"
40#include "llvm/Transforms/Utils/BasicBlockUtils.h"
41#include "llvm/Transforms/Utils/ModuleUtils.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));
82static cl::opt<bool> ClInstrumentAtomics("asan-instrument-atomics",
83       cl::desc("instrument atomic instructions (rmw, cmpxchg)"),
84       cl::Hidden, cl::init(true));
85// This flags limits the number of instructions to be instrumented
86// in any given BB. Normally, this should be set to unlimited (INT_MAX),
87// but due to http://llvm.org/bugs/show_bug.cgi?id=12652 we temporary
88// set it to 10000.
89static cl::opt<int> ClMaxInsnsToInstrumentPerBB("asan-max-ins-per-bb",
90       cl::init(10000),
91       cl::desc("maximal number of instructions to instrument in any given BB"),
92       cl::Hidden);
93// This flag may need to be replaced with -f[no]asan-stack.
94static cl::opt<bool> ClStack("asan-stack",
95       cl::desc("Handle stack memory"), cl::Hidden, cl::init(true));
96// This flag may need to be replaced with -f[no]asan-use-after-return.
97static cl::opt<bool> ClUseAfterReturn("asan-use-after-return",
98       cl::desc("Check return-after-free"), cl::Hidden, cl::init(false));
99// This flag may need to be replaced with -f[no]asan-globals.
100static cl::opt<bool> ClGlobals("asan-globals",
101       cl::desc("Handle global objects"), cl::Hidden, cl::init(true));
102static cl::opt<bool> ClMemIntrin("asan-memintrin",
103       cl::desc("Handle memset/memcpy/memmove"), cl::Hidden, cl::init(true));
104// This flag may need to be replaced with -fasan-blacklist.
105static cl::opt<std::string>  ClBlackListFile("asan-blacklist",
106       cl::desc("File containing the list of functions to ignore "
107                "during instrumentation"), cl::Hidden);
108
109// These flags allow to change the shadow mapping.
110// The shadow mapping looks like
111//    Shadow = (Mem >> scale) + (1 << offset_log)
112static cl::opt<int> ClMappingScale("asan-mapping-scale",
113       cl::desc("scale of asan shadow mapping"), cl::Hidden, cl::init(0));
114static cl::opt<int> ClMappingOffsetLog("asan-mapping-offset-log",
115       cl::desc("offset of asan shadow mapping"), cl::Hidden, cl::init(-1));
116
117// Optimization flags. Not user visible, used mostly for testing
118// and benchmarking the tool.
119static cl::opt<bool> ClOpt("asan-opt",
120       cl::desc("Optimize instrumentation"), cl::Hidden, cl::init(true));
121static cl::opt<bool> ClOptSameTemp("asan-opt-same-temp",
122       cl::desc("Instrument the same temp just once"), cl::Hidden,
123       cl::init(true));
124static cl::opt<bool> ClOptGlobals("asan-opt-globals",
125       cl::desc("Don't instrument scalar globals"), cl::Hidden, cl::init(true));
126
127// Debug flags.
128static cl::opt<int> ClDebug("asan-debug", cl::desc("debug"), cl::Hidden,
129                            cl::init(0));
130static cl::opt<int> ClDebugStack("asan-debug-stack", cl::desc("debug stack"),
131                                 cl::Hidden, cl::init(0));
132static cl::opt<std::string> ClDebugFunc("asan-debug-func",
133                                        cl::Hidden, cl::desc("Debug func"));
134static cl::opt<int> ClDebugMin("asan-debug-min", cl::desc("Debug min inst"),
135                               cl::Hidden, cl::init(-1));
136static cl::opt<int> ClDebugMax("asan-debug-max", cl::desc("Debug man inst"),
137                               cl::Hidden, cl::init(-1));
138
139namespace {
140
141/// AddressSanitizer: instrument the code in module to find memory bugs.
142struct AddressSanitizer : public ModulePass {
143  AddressSanitizer();
144  virtual const char *getPassName() const;
145  void instrumentMop(Instruction *I);
146  void instrumentAddress(Instruction *OrigIns, IRBuilder<> &IRB,
147                         Value *Addr, uint32_t TypeSize, bool IsWrite);
148  Instruction *generateCrashCode(IRBuilder<> &IRB, Value *Addr,
149                                 bool IsWrite, uint32_t TypeSize);
150  bool instrumentMemIntrinsic(MemIntrinsic *MI);
151  void instrumentMemIntrinsicParam(Instruction *OrigIns, Value *Addr,
152                                  Value *Size,
153                                   Instruction *InsertBefore, bool IsWrite);
154  Value *memToShadow(Value *Shadow, IRBuilder<> &IRB);
155  bool handleFunction(Module &M, Function &F);
156  bool maybeInsertAsanInitAtFunctionEntry(Function &F);
157  bool poisonStackInFunction(Module &M, Function &F);
158  virtual bool runOnModule(Module &M);
159  bool insertGlobalRedzones(Module &M);
160  static char ID;  // Pass identification, replacement for typeid
161
162 private:
163
164  uint64_t getAllocaSizeInBytes(AllocaInst *AI) {
165    Type *Ty = AI->getAllocatedType();
166    uint64_t SizeInBytes = TD->getTypeAllocSize(Ty);
167    return SizeInBytes;
168  }
169  uint64_t getAlignedSize(uint64_t SizeInBytes) {
170    return ((SizeInBytes + RedzoneSize - 1)
171            / RedzoneSize) * RedzoneSize;
172  }
173  uint64_t getAlignedAllocaSize(AllocaInst *AI) {
174    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
175    return getAlignedSize(SizeInBytes);
176  }
177
178  Function *checkInterfaceFunction(Constant *FuncOrBitcast);
179  void PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec, IRBuilder<> IRB,
180                   Value *ShadowBase, bool DoPoison);
181  bool LooksLikeCodeInBug11395(Instruction *I);
182
183  LLVMContext *C;
184  TargetData *TD;
185  uint64_t MappingOffset;
186  int MappingScale;
187  size_t RedzoneSize;
188  int LongSize;
189  Type *IntptrTy;
190  Type *IntptrPtrTy;
191  Function *AsanCtorFunction;
192  Function *AsanInitFunction;
193  Instruction *CtorInsertBefore;
194  OwningPtr<FunctionBlackList> BL;
195  // Accesses sizes are powers of two: 1, 2, 4, 8, 16.
196  static const size_t kNumberOfAccessSizes = 5;
197  // This array is indexed by AccessIsWrite and log2(AccessSize).
198  Function *AsanErrorCallback[2][kNumberOfAccessSizes];
199};
200}  // namespace
201
202char AddressSanitizer::ID = 0;
203INITIALIZE_PASS(AddressSanitizer, "asan",
204    "AddressSanitizer: detects use-after-free and out-of-bounds bugs.",
205    false, false)
206AddressSanitizer::AddressSanitizer() : ModulePass(ID) { }
207ModulePass *llvm::createAddressSanitizerPass() {
208  return new AddressSanitizer();
209}
210
211const char *AddressSanitizer::getPassName() const {
212  return "AddressSanitizer";
213}
214
215// Create a constant for Str so that we can pass it to the run-time lib.
216static GlobalVariable *createPrivateGlobalForString(Module &M, StringRef Str) {
217  Constant *StrConst = ConstantDataArray::getString(M.getContext(), Str);
218  return new GlobalVariable(M, StrConst->getType(), true,
219                            GlobalValue::PrivateLinkage, StrConst, "");
220}
221
222// Split the basic block and insert an if-then code.
223// Before:
224//   Head
225//   Cmp
226//   Tail
227// After:
228//   Head
229//   if (Cmp)
230//     ThenBlock
231//   Tail
232//
233// Returns the ThenBlock's terminator.
234static BranchInst *splitBlockAndInsertIfThen(Value *Cmp) {
235  Instruction *SplitBefore = cast<Instruction>(Cmp)->getNextNode();
236  BasicBlock *Head = SplitBefore->getParent();
237  BasicBlock *Tail = Head->splitBasicBlock(SplitBefore);
238  TerminatorInst *HeadOldTerm = Head->getTerminator();
239  LLVMContext &C = Head->getParent()->getParent()->getContext();
240  BasicBlock *ThenBlock = BasicBlock::Create(C, "", Head->getParent());
241  BranchInst *HeadNewTerm =
242    BranchInst::Create(/*ifTrue*/ThenBlock, /*ifFalse*/Tail, Cmp);
243  ReplaceInstWithInst(HeadOldTerm, HeadNewTerm);
244
245  BranchInst *CheckTerm = BranchInst::Create(Tail, ThenBlock);
246  return CheckTerm;
247}
248
249Value *AddressSanitizer::memToShadow(Value *Shadow, IRBuilder<> &IRB) {
250  // Shadow >> scale
251  Shadow = IRB.CreateLShr(Shadow, MappingScale);
252  if (MappingOffset == 0)
253    return Shadow;
254  // (Shadow >> scale) | offset
255  return IRB.CreateOr(Shadow, ConstantInt::get(IntptrTy,
256                                               MappingOffset));
257}
258
259void AddressSanitizer::instrumentMemIntrinsicParam(Instruction *OrigIns,
260    Value *Addr, Value *Size, Instruction *InsertBefore, bool IsWrite) {
261  // Check the first byte.
262  {
263    IRBuilder<> IRB(InsertBefore);
264    instrumentAddress(OrigIns, IRB, Addr, 8, IsWrite);
265  }
266  // Check the last byte.
267  {
268    IRBuilder<> IRB(InsertBefore);
269    Value *SizeMinusOne = IRB.CreateSub(
270        Size, ConstantInt::get(Size->getType(), 1));
271    SizeMinusOne = IRB.CreateIntCast(SizeMinusOne, IntptrTy, false);
272    Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
273    Value *AddrPlusSizeMinisOne = IRB.CreateAdd(AddrLong, SizeMinusOne);
274    instrumentAddress(OrigIns, IRB, AddrPlusSizeMinisOne, 8, IsWrite);
275  }
276}
277
278// Instrument memset/memmove/memcpy
279bool AddressSanitizer::instrumentMemIntrinsic(MemIntrinsic *MI) {
280  Value *Dst = MI->getDest();
281  MemTransferInst *MemTran = dyn_cast<MemTransferInst>(MI);
282  Value *Src = MemTran ? MemTran->getSource() : NULL;
283  Value *Length = MI->getLength();
284
285  Constant *ConstLength = dyn_cast<Constant>(Length);
286  Instruction *InsertBefore = MI;
287  if (ConstLength) {
288    if (ConstLength->isNullValue()) return false;
289  } else {
290    // The size is not a constant so it could be zero -- check at run-time.
291    IRBuilder<> IRB(InsertBefore);
292
293    Value *Cmp = IRB.CreateICmpNE(Length,
294                                  Constant::getNullValue(Length->getType()));
295    InsertBefore = splitBlockAndInsertIfThen(Cmp);
296  }
297
298  instrumentMemIntrinsicParam(MI, Dst, Length, InsertBefore, true);
299  if (Src)
300    instrumentMemIntrinsicParam(MI, Src, Length, InsertBefore, false);
301  return true;
302}
303
304// If I is an interesting memory access, return the PointerOperand
305// and set IsWrite. Otherwise return NULL.
306static Value *isInterestingMemoryAccess(Instruction *I, bool *IsWrite) {
307  if (LoadInst *LI = dyn_cast<LoadInst>(I)) {
308    if (!ClInstrumentReads) return NULL;
309    *IsWrite = false;
310    return LI->getPointerOperand();
311  }
312  if (StoreInst *SI = dyn_cast<StoreInst>(I)) {
313    if (!ClInstrumentWrites) return NULL;
314    *IsWrite = true;
315    return SI->getPointerOperand();
316  }
317  if (AtomicRMWInst *RMW = dyn_cast<AtomicRMWInst>(I)) {
318    if (!ClInstrumentAtomics) return NULL;
319    *IsWrite = true;
320    return RMW->getPointerOperand();
321  }
322  if (AtomicCmpXchgInst *XCHG = dyn_cast<AtomicCmpXchgInst>(I)) {
323    if (!ClInstrumentAtomics) return NULL;
324    *IsWrite = true;
325    return XCHG->getPointerOperand();
326  }
327  return NULL;
328}
329
330void AddressSanitizer::instrumentMop(Instruction *I) {
331  bool IsWrite;
332  Value *Addr = isInterestingMemoryAccess(I, &IsWrite);
333  assert(Addr);
334  if (ClOpt && ClOptGlobals && isa<GlobalVariable>(Addr)) {
335    // We are accessing a global scalar variable. Nothing to catch here.
336    return;
337  }
338  Type *OrigPtrTy = Addr->getType();
339  Type *OrigTy = cast<PointerType>(OrigPtrTy)->getElementType();
340
341  assert(OrigTy->isSized());
342  uint32_t TypeSize = TD->getTypeStoreSizeInBits(OrigTy);
343
344  if (TypeSize != 8  && TypeSize != 16 &&
345      TypeSize != 32 && TypeSize != 64 && TypeSize != 128) {
346    // Ignore all unusual sizes.
347    return;
348  }
349
350  IRBuilder<> IRB(I);
351  instrumentAddress(I, IRB, Addr, TypeSize, IsWrite);
352}
353
354// Validate the result of Module::getOrInsertFunction called for an interface
355// function of AddressSanitizer. If the instrumented module defines a function
356// with the same name, their prototypes must match, otherwise
357// getOrInsertFunction returns a bitcast.
358Function *AddressSanitizer::checkInterfaceFunction(Constant *FuncOrBitcast) {
359  if (isa<Function>(FuncOrBitcast)) return cast<Function>(FuncOrBitcast);
360  FuncOrBitcast->dump();
361  report_fatal_error("trying to redefine an AddressSanitizer "
362                     "interface function");
363}
364
365Instruction *AddressSanitizer::generateCrashCode(
366    IRBuilder<> &IRB, Value *Addr, bool IsWrite, uint32_t TypeSize) {
367  size_t AccessSizeIndex = CountTrailingZeros_32(TypeSize / 8);
368  assert(AccessSizeIndex < kNumberOfAccessSizes);
369  CallInst *Call = IRB.CreateCall(AsanErrorCallback[IsWrite][AccessSizeIndex],
370                                  Addr);
371  Call->setDoesNotReturn();
372  return Call;
373}
374
375void AddressSanitizer::instrumentAddress(Instruction *OrigIns,
376                                         IRBuilder<> &IRB, Value *Addr,
377                                         uint32_t TypeSize, bool IsWrite) {
378  Value *AddrLong = IRB.CreatePointerCast(Addr, IntptrTy);
379
380  Type *ShadowTy  = IntegerType::get(
381      *C, std::max(8U, TypeSize >> MappingScale));
382  Type *ShadowPtrTy = PointerType::get(ShadowTy, 0);
383  Value *ShadowPtr = memToShadow(AddrLong, IRB);
384  Value *CmpVal = Constant::getNullValue(ShadowTy);
385  Value *ShadowValue = IRB.CreateLoad(
386      IRB.CreateIntToPtr(ShadowPtr, ShadowPtrTy));
387
388  Value *Cmp = IRB.CreateICmpNE(ShadowValue, CmpVal);
389
390  Instruction *CheckTerm = splitBlockAndInsertIfThen(Cmp);
391  IRBuilder<> IRB2(CheckTerm);
392
393  size_t Granularity = 1 << MappingScale;
394  if (TypeSize < 8 * Granularity) {
395    // Addr & (Granularity - 1)
396    Value *LastAccessedByte = IRB2.CreateAnd(
397        AddrLong, ConstantInt::get(IntptrTy, Granularity - 1));
398    // (Addr & (Granularity - 1)) + size - 1
399    if (TypeSize / 8 > 1)
400      LastAccessedByte = IRB2.CreateAdd(
401          LastAccessedByte, ConstantInt::get(IntptrTy, TypeSize / 8 - 1));
402    // (uint8_t) ((Addr & (Granularity-1)) + size - 1)
403    LastAccessedByte = IRB2.CreateIntCast(
404        LastAccessedByte, IRB.getInt8Ty(), false);
405    // ((uint8_t) ((Addr & (Granularity-1)) + size - 1)) >= ShadowValue
406    Value *Cmp2 = IRB2.CreateICmpSGE(LastAccessedByte, ShadowValue);
407
408    CheckTerm = splitBlockAndInsertIfThen(Cmp2);
409  }
410
411  IRBuilder<> IRB1(CheckTerm);
412  Instruction *Crash = generateCrashCode(IRB1, AddrLong, IsWrite, TypeSize);
413  Crash->setDebugLoc(OrigIns->getDebugLoc());
414  ReplaceInstWithInst(CheckTerm, new UnreachableInst(*C));
415}
416
417// This function replaces all global variables with new variables that have
418// trailing redzones. It also creates a function that poisons
419// redzones and inserts this function into llvm.global_ctors.
420bool AddressSanitizer::insertGlobalRedzones(Module &M) {
421  SmallVector<GlobalVariable *, 16> GlobalsToChange;
422
423  for (Module::GlobalListType::iterator G = M.getGlobalList().begin(),
424       E = M.getGlobalList().end(); G != E; ++G) {
425    Type *Ty = cast<PointerType>(G->getType())->getElementType();
426    DEBUG(dbgs() << "GLOBAL: " << *G);
427
428    if (!Ty->isSized()) continue;
429    if (!G->hasInitializer()) continue;
430    // Touch only those globals that will not be defined in other modules.
431    // Don't handle ODR type linkages since other modules may be built w/o asan.
432    if (G->getLinkage() != GlobalVariable::ExternalLinkage &&
433        G->getLinkage() != GlobalVariable::PrivateLinkage &&
434        G->getLinkage() != GlobalVariable::InternalLinkage)
435      continue;
436    // Two problems with thread-locals:
437    //   - The address of the main thread's copy can't be computed at link-time.
438    //   - Need to poison all copies, not just the main thread's one.
439    if (G->isThreadLocal())
440      continue;
441    // For now, just ignore this Alloca if the alignment is large.
442    if (G->getAlignment() > RedzoneSize) continue;
443
444    // Ignore all the globals with the names starting with "\01L_OBJC_".
445    // Many of those are put into the .cstring section. The linker compresses
446    // that section by removing the spare \0s after the string terminator, so
447    // our redzones get broken.
448    if ((G->getName().find("\01L_OBJC_") == 0) ||
449        (G->getName().find("\01l_OBJC_") == 0)) {
450      DEBUG(dbgs() << "Ignoring \\01L_OBJC_* global: " << *G);
451      continue;
452    }
453
454    if (G->hasSection()) {
455      StringRef Section(G->getSection());
456      // Ignore the globals from the __OBJC section. The ObjC runtime assumes
457      // those conform to /usr/lib/objc/runtime.h, so we can't add redzones to
458      // them.
459      if ((Section.find("__OBJC,") == 0) ||
460          (Section.find("__DATA, __objc_") == 0)) {
461        DEBUG(dbgs() << "Ignoring ObjC runtime global: " << *G);
462        continue;
463      }
464      // See http://code.google.com/p/address-sanitizer/issues/detail?id=32
465      // Constant CFString instances are compiled in the following way:
466      //  -- the string buffer is emitted into
467      //     __TEXT,__cstring,cstring_literals
468      //  -- the constant NSConstantString structure referencing that buffer
469      //     is placed into __DATA,__cfstring
470      // Therefore there's no point in placing redzones into __DATA,__cfstring.
471      // Moreover, it causes the linker to crash on OS X 10.7
472      if (Section.find("__DATA,__cfstring") == 0) {
473        DEBUG(dbgs() << "Ignoring CFString: " << *G);
474        continue;
475      }
476    }
477
478    GlobalsToChange.push_back(G);
479  }
480
481  size_t n = GlobalsToChange.size();
482  if (n == 0) return false;
483
484  // A global is described by a structure
485  //   size_t beg;
486  //   size_t size;
487  //   size_t size_with_redzone;
488  //   const char *name;
489  // We initialize an array of such structures and pass it to a run-time call.
490  StructType *GlobalStructTy = StructType::get(IntptrTy, IntptrTy,
491                                               IntptrTy, IntptrTy, NULL);
492  SmallVector<Constant *, 16> Initializers(n);
493
494  IRBuilder<> IRB(CtorInsertBefore);
495
496  for (size_t i = 0; i < n; i++) {
497    GlobalVariable *G = GlobalsToChange[i];
498    PointerType *PtrTy = cast<PointerType>(G->getType());
499    Type *Ty = PtrTy->getElementType();
500    uint64_t SizeInBytes = TD->getTypeAllocSize(Ty);
501    uint64_t RightRedzoneSize = RedzoneSize +
502        (RedzoneSize - (SizeInBytes % RedzoneSize));
503    Type *RightRedZoneTy = ArrayType::get(IRB.getInt8Ty(), RightRedzoneSize);
504
505    StructType *NewTy = StructType::get(Ty, RightRedZoneTy, NULL);
506    Constant *NewInitializer = ConstantStruct::get(
507        NewTy, G->getInitializer(),
508        Constant::getNullValue(RightRedZoneTy), NULL);
509
510    SmallString<2048> DescriptionOfGlobal = G->getName();
511    DescriptionOfGlobal += " (";
512    DescriptionOfGlobal += M.getModuleIdentifier();
513    DescriptionOfGlobal += ")";
514    GlobalVariable *Name = createPrivateGlobalForString(M, DescriptionOfGlobal);
515
516    // Create a new global variable with enough space for a redzone.
517    GlobalVariable *NewGlobal = new GlobalVariable(
518        M, NewTy, G->isConstant(), G->getLinkage(),
519        NewInitializer, "", G, G->getThreadLocalMode());
520    NewGlobal->copyAttributesFrom(G);
521    NewGlobal->setAlignment(RedzoneSize);
522
523    Value *Indices2[2];
524    Indices2[0] = IRB.getInt32(0);
525    Indices2[1] = IRB.getInt32(0);
526
527    G->replaceAllUsesWith(
528        ConstantExpr::getGetElementPtr(NewGlobal, Indices2, true));
529    NewGlobal->takeName(G);
530    G->eraseFromParent();
531
532    Initializers[i] = ConstantStruct::get(
533        GlobalStructTy,
534        ConstantExpr::getPointerCast(NewGlobal, IntptrTy),
535        ConstantInt::get(IntptrTy, SizeInBytes),
536        ConstantInt::get(IntptrTy, SizeInBytes + RightRedzoneSize),
537        ConstantExpr::getPointerCast(Name, IntptrTy),
538        NULL);
539    DEBUG(dbgs() << "NEW GLOBAL:\n" << *NewGlobal);
540  }
541
542  ArrayType *ArrayOfGlobalStructTy = ArrayType::get(GlobalStructTy, n);
543  GlobalVariable *AllGlobals = new GlobalVariable(
544      M, ArrayOfGlobalStructTy, false, GlobalVariable::PrivateLinkage,
545      ConstantArray::get(ArrayOfGlobalStructTy, Initializers), "");
546
547  Function *AsanRegisterGlobals = checkInterfaceFunction(M.getOrInsertFunction(
548      kAsanRegisterGlobalsName, IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
549  AsanRegisterGlobals->setLinkage(Function::ExternalLinkage);
550
551  IRB.CreateCall2(AsanRegisterGlobals,
552                  IRB.CreatePointerCast(AllGlobals, IntptrTy),
553                  ConstantInt::get(IntptrTy, n));
554
555  // We also need to unregister globals at the end, e.g. when a shared library
556  // gets closed.
557  Function *AsanDtorFunction = Function::Create(
558      FunctionType::get(Type::getVoidTy(*C), false),
559      GlobalValue::InternalLinkage, kAsanModuleDtorName, &M);
560  BasicBlock *AsanDtorBB = BasicBlock::Create(*C, "", AsanDtorFunction);
561  IRBuilder<> IRB_Dtor(ReturnInst::Create(*C, AsanDtorBB));
562  Function *AsanUnregisterGlobals =
563      checkInterfaceFunction(M.getOrInsertFunction(
564          kAsanUnregisterGlobalsName,
565          IRB.getVoidTy(), IntptrTy, IntptrTy, NULL));
566  AsanUnregisterGlobals->setLinkage(Function::ExternalLinkage);
567
568  IRB_Dtor.CreateCall2(AsanUnregisterGlobals,
569                       IRB.CreatePointerCast(AllGlobals, IntptrTy),
570                       ConstantInt::get(IntptrTy, n));
571  appendToGlobalDtors(M, AsanDtorFunction, kAsanCtorAndCtorPriority);
572
573  DEBUG(dbgs() << M);
574  return true;
575}
576
577// virtual
578bool AddressSanitizer::runOnModule(Module &M) {
579  // Initialize the private fields. No one has accessed them before.
580  TD = getAnalysisIfAvailable<TargetData>();
581  if (!TD)
582    return false;
583  BL.reset(new FunctionBlackList(ClBlackListFile));
584
585  C = &(M.getContext());
586  LongSize = TD->getPointerSizeInBits();
587  IntptrTy = Type::getIntNTy(*C, LongSize);
588  IntptrPtrTy = PointerType::get(IntptrTy, 0);
589
590  AsanCtorFunction = Function::Create(
591      FunctionType::get(Type::getVoidTy(*C), false),
592      GlobalValue::InternalLinkage, kAsanModuleCtorName, &M);
593  BasicBlock *AsanCtorBB = BasicBlock::Create(*C, "", AsanCtorFunction);
594  CtorInsertBefore = ReturnInst::Create(*C, AsanCtorBB);
595
596  // call __asan_init in the module ctor.
597  IRBuilder<> IRB(CtorInsertBefore);
598  AsanInitFunction = checkInterfaceFunction(
599      M.getOrInsertFunction(kAsanInitName, IRB.getVoidTy(), NULL));
600  AsanInitFunction->setLinkage(Function::ExternalLinkage);
601  IRB.CreateCall(AsanInitFunction);
602
603  // Create __asan_report* callbacks.
604  for (size_t AccessIsWrite = 0; AccessIsWrite <= 1; AccessIsWrite++) {
605    for (size_t AccessSizeIndex = 0; AccessSizeIndex < kNumberOfAccessSizes;
606         AccessSizeIndex++) {
607      // IsWrite and TypeSize are encoded in the function name.
608      std::string FunctionName = std::string(kAsanReportErrorTemplate) +
609          (AccessIsWrite ? "store" : "load") + itostr(1 << AccessSizeIndex);
610      AsanErrorCallback[AccessIsWrite][AccessSizeIndex] = cast<Function>(
611        M.getOrInsertFunction(FunctionName, IRB.getVoidTy(), IntptrTy, NULL));
612    }
613  }
614
615  llvm::Triple targetTriple(M.getTargetTriple());
616  bool isAndroid = targetTriple.getEnvironment() == llvm::Triple::ANDROIDEABI;
617
618  MappingOffset = isAndroid ? kDefaultShadowOffsetAndroid :
619    (LongSize == 32 ? kDefaultShadowOffset32 : kDefaultShadowOffset64);
620  if (ClMappingOffsetLog >= 0) {
621    if (ClMappingOffsetLog == 0) {
622      // special case
623      MappingOffset = 0;
624    } else {
625      MappingOffset = 1ULL << ClMappingOffsetLog;
626    }
627  }
628  MappingScale = kDefaultShadowScale;
629  if (ClMappingScale) {
630    MappingScale = ClMappingScale;
631  }
632  // Redzone used for stack and globals is at least 32 bytes.
633  // For scales 6 and 7, the redzone has to be 64 and 128 bytes respectively.
634  RedzoneSize = std::max(32, (int)(1 << MappingScale));
635
636  bool Res = false;
637
638  if (ClGlobals)
639    Res |= insertGlobalRedzones(M);
640
641  if (ClMappingOffsetLog >= 0) {
642    // Tell the run-time the current values of mapping offset and scale.
643    GlobalValue *asan_mapping_offset =
644        new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
645                       ConstantInt::get(IntptrTy, MappingOffset),
646                       kAsanMappingOffsetName);
647    // Read the global, otherwise it may be optimized away.
648    IRB.CreateLoad(asan_mapping_offset, true);
649  }
650  if (ClMappingScale) {
651    GlobalValue *asan_mapping_scale =
652        new GlobalVariable(M, IntptrTy, true, GlobalValue::LinkOnceODRLinkage,
653                           ConstantInt::get(IntptrTy, MappingScale),
654                           kAsanMappingScaleName);
655    // Read the global, otherwise it may be optimized away.
656    IRB.CreateLoad(asan_mapping_scale, true);
657  }
658
659
660  for (Module::iterator F = M.begin(), E = M.end(); F != E; ++F) {
661    if (F->isDeclaration()) continue;
662    Res |= handleFunction(M, *F);
663  }
664
665  appendToGlobalCtors(M, AsanCtorFunction, kAsanCtorAndCtorPriority);
666
667  return Res;
668}
669
670bool AddressSanitizer::maybeInsertAsanInitAtFunctionEntry(Function &F) {
671  // For each NSObject descendant having a +load method, this method is invoked
672  // by the ObjC runtime before any of the static constructors is called.
673  // Therefore we need to instrument such methods with a call to __asan_init
674  // at the beginning in order to initialize our runtime before any access to
675  // the shadow memory.
676  // We cannot just ignore these methods, because they may call other
677  // instrumented functions.
678  if (F.getName().find(" load]") != std::string::npos) {
679    IRBuilder<> IRB(F.begin()->begin());
680    IRB.CreateCall(AsanInitFunction);
681    return true;
682  }
683  return false;
684}
685
686bool AddressSanitizer::handleFunction(Module &M, Function &F) {
687  if (BL->isIn(F)) return false;
688  if (&F == AsanCtorFunction) return false;
689
690  // If needed, insert __asan_init before checking for AddressSafety attr.
691  maybeInsertAsanInitAtFunctionEntry(F);
692
693  if (!F.hasFnAttr(Attribute::AddressSafety)) return false;
694
695  if (!ClDebugFunc.empty() && ClDebugFunc != F.getName())
696    return false;
697  // We want to instrument every address only once per basic block
698  // (unless there are calls between uses).
699  SmallSet<Value*, 16> TempsToInstrument;
700  SmallVector<Instruction*, 16> ToInstrument;
701  SmallVector<Instruction*, 8> NoReturnCalls;
702  bool IsWrite;
703
704  // Fill the set of memory operations to instrument.
705  for (Function::iterator FI = F.begin(), FE = F.end();
706       FI != FE; ++FI) {
707    TempsToInstrument.clear();
708    int NumInsnsPerBB = 0;
709    for (BasicBlock::iterator BI = FI->begin(), BE = FI->end();
710         BI != BE; ++BI) {
711      if (LooksLikeCodeInBug11395(BI)) return false;
712      if (Value *Addr = isInterestingMemoryAccess(BI, &IsWrite)) {
713        if (ClOpt && ClOptSameTemp) {
714          if (!TempsToInstrument.insert(Addr))
715            continue;  // We've seen this temp in the current BB.
716        }
717      } else if (isa<MemIntrinsic>(BI) && ClMemIntrin) {
718        // ok, take it.
719      } else {
720        if (CallInst *CI = dyn_cast<CallInst>(BI)) {
721          // A call inside BB.
722          TempsToInstrument.clear();
723          if (CI->doesNotReturn()) {
724            NoReturnCalls.push_back(CI);
725          }
726        }
727        continue;
728      }
729      ToInstrument.push_back(BI);
730      NumInsnsPerBB++;
731      if (NumInsnsPerBB >= ClMaxInsnsToInstrumentPerBB)
732        break;
733    }
734  }
735
736  // Instrument.
737  int NumInstrumented = 0;
738  for (size_t i = 0, n = ToInstrument.size(); i != n; i++) {
739    Instruction *Inst = ToInstrument[i];
740    if (ClDebugMin < 0 || ClDebugMax < 0 ||
741        (NumInstrumented >= ClDebugMin && NumInstrumented <= ClDebugMax)) {
742      if (isInterestingMemoryAccess(Inst, &IsWrite))
743        instrumentMop(Inst);
744      else
745        instrumentMemIntrinsic(cast<MemIntrinsic>(Inst));
746    }
747    NumInstrumented++;
748  }
749
750  DEBUG(dbgs() << F);
751
752  bool ChangedStack = poisonStackInFunction(M, F);
753
754  // We must unpoison the stack before every NoReturn call (throw, _exit, etc).
755  // See e.g. http://code.google.com/p/address-sanitizer/issues/detail?id=37
756  for (size_t i = 0, n = NoReturnCalls.size(); i != n; i++) {
757    Instruction *CI = NoReturnCalls[i];
758    IRBuilder<> IRB(CI);
759    IRB.CreateCall(M.getOrInsertFunction(kAsanHandleNoReturnName,
760                                         IRB.getVoidTy(), NULL));
761  }
762
763  return NumInstrumented > 0 || ChangedStack || !NoReturnCalls.empty();
764}
765
766static uint64_t ValueForPoison(uint64_t PoisonByte, size_t ShadowRedzoneSize) {
767  if (ShadowRedzoneSize == 1) return PoisonByte;
768  if (ShadowRedzoneSize == 2) return (PoisonByte << 8) + PoisonByte;
769  if (ShadowRedzoneSize == 4)
770    return (PoisonByte << 24) + (PoisonByte << 16) +
771        (PoisonByte << 8) + (PoisonByte);
772  llvm_unreachable("ShadowRedzoneSize is either 1, 2 or 4");
773}
774
775static void PoisonShadowPartialRightRedzone(uint8_t *Shadow,
776                                            size_t Size,
777                                            size_t RedzoneSize,
778                                            size_t ShadowGranularity,
779                                            uint8_t Magic) {
780  for (size_t i = 0; i < RedzoneSize;
781       i+= ShadowGranularity, Shadow++) {
782    if (i + ShadowGranularity <= Size) {
783      *Shadow = 0;  // fully addressable
784    } else if (i >= Size) {
785      *Shadow = Magic;  // unaddressable
786    } else {
787      *Shadow = Size - i;  // first Size-i bytes are addressable
788    }
789  }
790}
791
792void AddressSanitizer::PoisonStack(const ArrayRef<AllocaInst*> &AllocaVec,
793                                   IRBuilder<> IRB,
794                                   Value *ShadowBase, bool DoPoison) {
795  size_t ShadowRZSize = RedzoneSize >> MappingScale;
796  assert(ShadowRZSize >= 1 && ShadowRZSize <= 4);
797  Type *RZTy = Type::getIntNTy(*C, ShadowRZSize * 8);
798  Type *RZPtrTy = PointerType::get(RZTy, 0);
799
800  Value *PoisonLeft  = ConstantInt::get(RZTy,
801    ValueForPoison(DoPoison ? kAsanStackLeftRedzoneMagic : 0LL, ShadowRZSize));
802  Value *PoisonMid   = ConstantInt::get(RZTy,
803    ValueForPoison(DoPoison ? kAsanStackMidRedzoneMagic : 0LL, ShadowRZSize));
804  Value *PoisonRight = ConstantInt::get(RZTy,
805    ValueForPoison(DoPoison ? kAsanStackRightRedzoneMagic : 0LL, ShadowRZSize));
806
807  // poison the first red zone.
808  IRB.CreateStore(PoisonLeft, IRB.CreateIntToPtr(ShadowBase, RZPtrTy));
809
810  // poison all other red zones.
811  uint64_t Pos = RedzoneSize;
812  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
813    AllocaInst *AI = AllocaVec[i];
814    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
815    uint64_t AlignedSize = getAlignedAllocaSize(AI);
816    assert(AlignedSize - SizeInBytes < RedzoneSize);
817    Value *Ptr = NULL;
818
819    Pos += AlignedSize;
820
821    assert(ShadowBase->getType() == IntptrTy);
822    if (SizeInBytes < AlignedSize) {
823      // Poison the partial redzone at right
824      Ptr = IRB.CreateAdd(
825          ShadowBase, ConstantInt::get(IntptrTy,
826                                       (Pos >> MappingScale) - ShadowRZSize));
827      size_t AddressableBytes = RedzoneSize - (AlignedSize - SizeInBytes);
828      uint32_t Poison = 0;
829      if (DoPoison) {
830        PoisonShadowPartialRightRedzone((uint8_t*)&Poison, AddressableBytes,
831                                        RedzoneSize,
832                                        1ULL << MappingScale,
833                                        kAsanStackPartialRedzoneMagic);
834      }
835      Value *PartialPoison = ConstantInt::get(RZTy, Poison);
836      IRB.CreateStore(PartialPoison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
837    }
838
839    // Poison the full redzone at right.
840    Ptr = IRB.CreateAdd(ShadowBase,
841                        ConstantInt::get(IntptrTy, Pos >> MappingScale));
842    Value *Poison = i == AllocaVec.size() - 1 ? PoisonRight : PoisonMid;
843    IRB.CreateStore(Poison, IRB.CreateIntToPtr(Ptr, RZPtrTy));
844
845    Pos += RedzoneSize;
846  }
847}
848
849// Workaround for bug 11395: we don't want to instrument stack in functions
850// with large assembly blobs (32-bit only), otherwise reg alloc may crash.
851// FIXME: remove once the bug 11395 is fixed.
852bool AddressSanitizer::LooksLikeCodeInBug11395(Instruction *I) {
853  if (LongSize != 32) return false;
854  CallInst *CI = dyn_cast<CallInst>(I);
855  if (!CI || !CI->isInlineAsm()) return false;
856  if (CI->getNumArgOperands() <= 5) return false;
857  // We have inline assembly with quite a few arguments.
858  return true;
859}
860
861// Find all static Alloca instructions and put
862// poisoned red zones around all of them.
863// Then unpoison everything back before the function returns.
864//
865// Stack poisoning does not play well with exception handling.
866// When an exception is thrown, we essentially bypass the code
867// that unpoisones the stack. This is why the run-time library has
868// to intercept __cxa_throw (as well as longjmp, etc) and unpoison the entire
869// stack in the interceptor. This however does not work inside the
870// actual function which catches the exception. Most likely because the
871// compiler hoists the load of the shadow value somewhere too high.
872// This causes asan to report a non-existing bug on 453.povray.
873// It sounds like an LLVM bug.
874bool AddressSanitizer::poisonStackInFunction(Module &M, Function &F) {
875  if (!ClStack) return false;
876  SmallVector<AllocaInst*, 16> AllocaVec;
877  SmallVector<Instruction*, 8> RetVec;
878  uint64_t TotalSize = 0;
879
880  // Filter out Alloca instructions we want (and can) handle.
881  // Collect Ret instructions.
882  for (Function::iterator FI = F.begin(), FE = F.end();
883       FI != FE; ++FI) {
884    BasicBlock &BB = *FI;
885    for (BasicBlock::iterator BI = BB.begin(), BE = BB.end();
886         BI != BE; ++BI) {
887      if (isa<ReturnInst>(BI)) {
888          RetVec.push_back(BI);
889          continue;
890      }
891
892      AllocaInst *AI = dyn_cast<AllocaInst>(BI);
893      if (!AI) continue;
894      if (AI->isArrayAllocation()) continue;
895      if (!AI->isStaticAlloca()) continue;
896      if (!AI->getAllocatedType()->isSized()) continue;
897      if (AI->getAlignment() > RedzoneSize) continue;
898      AllocaVec.push_back(AI);
899      uint64_t AlignedSize =  getAlignedAllocaSize(AI);
900      TotalSize += AlignedSize;
901    }
902  }
903
904  if (AllocaVec.empty()) return false;
905
906  uint64_t LocalStackSize = TotalSize + (AllocaVec.size() + 1) * RedzoneSize;
907
908  bool DoStackMalloc = ClUseAfterReturn
909      && LocalStackSize <= kMaxStackMallocSize;
910
911  Instruction *InsBefore = AllocaVec[0];
912  IRBuilder<> IRB(InsBefore);
913
914
915  Type *ByteArrayTy = ArrayType::get(IRB.getInt8Ty(), LocalStackSize);
916  AllocaInst *MyAlloca =
917      new AllocaInst(ByteArrayTy, "MyAlloca", InsBefore);
918  MyAlloca->setAlignment(RedzoneSize);
919  assert(MyAlloca->isStaticAlloca());
920  Value *OrigStackBase = IRB.CreatePointerCast(MyAlloca, IntptrTy);
921  Value *LocalStackBase = OrigStackBase;
922
923  if (DoStackMalloc) {
924    Value *AsanStackMallocFunc = M.getOrInsertFunction(
925        kAsanStackMallocName, IntptrTy, IntptrTy, IntptrTy, NULL);
926    LocalStackBase = IRB.CreateCall2(AsanStackMallocFunc,
927        ConstantInt::get(IntptrTy, LocalStackSize), OrigStackBase);
928  }
929
930  // This string will be parsed by the run-time (DescribeStackAddress).
931  SmallString<2048> StackDescriptionStorage;
932  raw_svector_ostream StackDescription(StackDescriptionStorage);
933  StackDescription << F.getName() << " " << AllocaVec.size() << " ";
934
935  uint64_t Pos = RedzoneSize;
936  // Replace Alloca instructions with base+offset.
937  for (size_t i = 0, n = AllocaVec.size(); i < n; i++) {
938    AllocaInst *AI = AllocaVec[i];
939    uint64_t SizeInBytes = getAllocaSizeInBytes(AI);
940    StringRef Name = AI->getName();
941    StackDescription << Pos << " " << SizeInBytes << " "
942                     << Name.size() << " " << Name << " ";
943    uint64_t AlignedSize = getAlignedAllocaSize(AI);
944    assert((AlignedSize % RedzoneSize) == 0);
945    AI->replaceAllUsesWith(
946        IRB.CreateIntToPtr(
947            IRB.CreateAdd(LocalStackBase, ConstantInt::get(IntptrTy, Pos)),
948            AI->getType()));
949    Pos += AlignedSize + RedzoneSize;
950  }
951  assert(Pos == LocalStackSize);
952
953  // Write the Magic value and the frame description constant to the redzone.
954  Value *BasePlus0 = IRB.CreateIntToPtr(LocalStackBase, IntptrPtrTy);
955  IRB.CreateStore(ConstantInt::get(IntptrTy, kCurrentStackFrameMagic),
956                  BasePlus0);
957  Value *BasePlus1 = IRB.CreateAdd(LocalStackBase,
958                                   ConstantInt::get(IntptrTy, LongSize/8));
959  BasePlus1 = IRB.CreateIntToPtr(BasePlus1, IntptrPtrTy);
960  Value *Description = IRB.CreatePointerCast(
961      createPrivateGlobalForString(M, StackDescription.str()),
962      IntptrTy);
963  IRB.CreateStore(Description, BasePlus1);
964
965  // Poison the stack redzones at the entry.
966  Value *ShadowBase = memToShadow(LocalStackBase, IRB);
967  PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRB, ShadowBase, true);
968
969  Value *AsanStackFreeFunc = NULL;
970  if (DoStackMalloc) {
971    AsanStackFreeFunc = M.getOrInsertFunction(
972        kAsanStackFreeName, IRB.getVoidTy(),
973        IntptrTy, IntptrTy, IntptrTy, NULL);
974  }
975
976  // Unpoison the stack before all ret instructions.
977  for (size_t i = 0, n = RetVec.size(); i < n; i++) {
978    Instruction *Ret = RetVec[i];
979    IRBuilder<> IRBRet(Ret);
980
981    // Mark the current frame as retired.
982    IRBRet.CreateStore(ConstantInt::get(IntptrTy, kRetiredStackFrameMagic),
983                       BasePlus0);
984    // Unpoison the stack.
985    PoisonStack(ArrayRef<AllocaInst*>(AllocaVec), IRBRet, ShadowBase, false);
986
987    if (DoStackMalloc) {
988      IRBRet.CreateCall3(AsanStackFreeFunc, LocalStackBase,
989                         ConstantInt::get(IntptrTy, LocalStackSize),
990                         OrigStackBase);
991    }
992  }
993
994  if (ClDebugStack) {
995    DEBUG(dbgs() << F);
996  }
997
998  return true;
999}
1000