Home | History | Annotate | Download | only in Analysis 
      1 //===------ MemoryBuiltins.cpp - Identify calls to memory builtins --------===//
      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 family of functions identifies calls to builtin functions that allocate
     11 // or free memory.
     12 //
     13 //===----------------------------------------------------------------------===//
     14 
     15 #define DEBUG_TYPE "memory-builtins"
     16 #include "llvm/Analysis/MemoryBuiltins.h"
     17 #include "llvm/ADT/STLExtras.h"
     18 #include "llvm/ADT/Statistic.h"
     19 #include "llvm/Analysis/ValueTracking.h"
     20 #include "llvm/IR/DataLayout.h"
     21 #include "llvm/IR/GlobalVariable.h"
     22 #include "llvm/IR/Instructions.h"
     23 #include "llvm/IR/Intrinsics.h"
     24 #include "llvm/IR/Metadata.h"
     25 #include "llvm/IR/Module.h"
     26 #include "llvm/Support/Debug.h"
     27 #include "llvm/Support/MathExtras.h"
     28 #include "llvm/Support/raw_ostream.h"
     29 #include "llvm/Target/TargetLibraryInfo.h"
     30 #include "llvm/Transforms/Utils/Local.h"
     31 using namespace llvm;
     32 
     33 enum AllocType {
     34   MallocLike         = 1<<0, // allocates
     35   CallocLike         = 1<<1, // allocates + bzero
     36   ReallocLike        = 1<<2, // reallocates
     37   StrDupLike         = 1<<3,
     38   AllocLike          = MallocLike | CallocLike | StrDupLike,
     39   AnyAlloc           = MallocLike | CallocLike | ReallocLike | StrDupLike
     40 };
     41 
     42 struct AllocFnsTy {
     43   LibFunc::Func Func;
     44   AllocType AllocTy;
     45   unsigned char NumParams;
     46   // First and Second size parameters (or -1 if unused)
     47   signed char FstParam, SndParam;
     48 };
     49 
     50 // FIXME: certain users need more information. E.g., SimplifyLibCalls needs to
     51 // know which functions are nounwind, noalias, nocapture parameters, etc.
     52 static const AllocFnsTy AllocationFnData[] = {
     53   {LibFunc::malloc,              MallocLike,  1, 0,  -1},
     54   {LibFunc::valloc,              MallocLike,  1, 0,  -1},
     55   {LibFunc::Znwj,                MallocLike,  1, 0,  -1}, // new(unsigned int)
     56   {LibFunc::ZnwjRKSt9nothrow_t,  MallocLike,  2, 0,  -1}, // new(unsigned int, nothrow)
     57   {LibFunc::Znwm,                MallocLike,  1, 0,  -1}, // new(unsigned long)
     58   {LibFunc::ZnwmRKSt9nothrow_t,  MallocLike,  2, 0,  -1}, // new(unsigned long, nothrow)
     59   {LibFunc::Znaj,                MallocLike,  1, 0,  -1}, // new[](unsigned int)
     60   {LibFunc::ZnajRKSt9nothrow_t,  MallocLike,  2, 0,  -1}, // new[](unsigned int, nothrow)
     61   {LibFunc::Znam,                MallocLike,  1, 0,  -1}, // new[](unsigned long)
     62   {LibFunc::ZnamRKSt9nothrow_t,  MallocLike,  2, 0,  -1}, // new[](unsigned long, nothrow)
     63   {LibFunc::posix_memalign,      MallocLike,  3, 2,  -1},
     64   {LibFunc::calloc,              CallocLike,  2, 0,   1},
     65   {LibFunc::realloc,             ReallocLike, 2, 1,  -1},
     66   {LibFunc::reallocf,            ReallocLike, 2, 1,  -1},
     67   {LibFunc::strdup,              StrDupLike,  1, -1, -1},
     68   {LibFunc::strndup,             StrDupLike,  2, 1,  -1}
     69 };
     70 
     71 
     72 static Function *getCalledFunction(const Value *V, bool LookThroughBitCast) {
     73   if (LookThroughBitCast)
     74     V = V->stripPointerCasts();
     75 
     76   CallSite CS(const_cast<Value*>(V));
     77   if (!CS.getInstruction())
     78     return 0;
     79 
     80   if (CS.isNoBuiltin())
     81     return 0;
     82 
     83   Function *Callee = CS.getCalledFunction();
     84   if (!Callee || !Callee->isDeclaration())
     85     return 0;
     86   return Callee;
     87 }
     88 
     89 /// \brief Returns the allocation data for the given value if it is a call to a
     90 /// known allocation function, and NULL otherwise.
     91 static const AllocFnsTy *getAllocationData(const Value *V, AllocType AllocTy,
     92                                            const TargetLibraryInfo *TLI,
     93                                            bool LookThroughBitCast = false) {
     94   // Skip intrinsics
     95   if (isa<IntrinsicInst>(V))
     96     return 0;
     97 
     98   Function *Callee = getCalledFunction(V, LookThroughBitCast);
     99   if (!Callee)
    100     return 0;
    101 
    102   // Make sure that the function is available.
    103   StringRef FnName = Callee->getName();
    104   LibFunc::Func TLIFn;
    105   if (!TLI || !TLI->getLibFunc(FnName, TLIFn) || !TLI->has(TLIFn))
    106     return 0;
    107 
    108   unsigned i = 0;
    109   bool found = false;
    110   for ( ; i < array_lengthof(AllocationFnData); ++i) {
    111     if (AllocationFnData[i].Func == TLIFn) {
    112       found = true;
    113       break;
    114     }
    115   }
    116   if (!found)
    117     return 0;
    118 
    119   const AllocFnsTy *FnData = &AllocationFnData[i];
    120   if ((FnData->AllocTy & AllocTy) == 0)
    121     return 0;
    122 
    123   // Check function prototype.
    124   int FstParam = FnData->FstParam;
    125   int SndParam = FnData->SndParam;
    126   FunctionType *FTy = Callee->getFunctionType();
    127 
    128   if (FTy->getReturnType() == Type::getInt8PtrTy(FTy->getContext()) &&
    129       FTy->getNumParams() == FnData->NumParams &&
    130       (FstParam < 0 ||
    131        (FTy->getParamType(FstParam)->isIntegerTy(32) ||
    132         FTy->getParamType(FstParam)->isIntegerTy(64))) &&
    133       (SndParam < 0 ||
    134        FTy->getParamType(SndParam)->isIntegerTy(32) ||
    135        FTy->getParamType(SndParam)->isIntegerTy(64)))
    136     return FnData;
    137   return 0;
    138 }
    139 
    140 static bool hasNoAliasAttr(const Value *V, bool LookThroughBitCast) {
    141   ImmutableCallSite CS(LookThroughBitCast ? V->stripPointerCasts() : V);
    142   return CS && CS.hasFnAttr(Attribute::NoAlias);
    143 }
    144 
    145 
    146 /// \brief Tests if a value is a call or invoke to a library function that
    147 /// allocates or reallocates memory (either malloc, calloc, realloc, or strdup
    148 /// like).
    149 bool llvm::isAllocationFn(const Value *V, const TargetLibraryInfo *TLI,
    150                           bool LookThroughBitCast) {
    151   return getAllocationData(V, AnyAlloc, TLI, LookThroughBitCast);
    152 }
    153 
    154 /// \brief Tests if a value is a call or invoke to a function that returns a
    155 /// NoAlias pointer (including malloc/calloc/realloc/strdup-like functions).
    156 bool llvm::isNoAliasFn(const Value *V, const TargetLibraryInfo *TLI,
    157                        bool LookThroughBitCast) {
    158   // it's safe to consider realloc as noalias since accessing the original
    159   // pointer is undefined behavior
    160   return isAllocationFn(V, TLI, LookThroughBitCast) ||
    161          hasNoAliasAttr(V, LookThroughBitCast);
    162 }
    163 
    164 /// \brief Tests if a value is a call or invoke to a library function that
    165 /// allocates uninitialized memory (such as malloc).
    166 bool llvm::isMallocLikeFn(const Value *V, const TargetLibraryInfo *TLI,
    167                           bool LookThroughBitCast) {
    168   return getAllocationData(V, MallocLike, TLI, LookThroughBitCast);
    169 }
    170 
    171 /// \brief Tests if a value is a call or invoke to a library function that
    172 /// allocates zero-filled memory (such as calloc).
    173 bool llvm::isCallocLikeFn(const Value *V, const TargetLibraryInfo *TLI,
    174                           bool LookThroughBitCast) {
    175   return getAllocationData(V, CallocLike, TLI, LookThroughBitCast);
    176 }
    177 
    178 /// \brief Tests if a value is a call or invoke to a library function that
    179 /// allocates memory (either malloc, calloc, or strdup like).
    180 bool llvm::isAllocLikeFn(const Value *V, const TargetLibraryInfo *TLI,
    181                          bool LookThroughBitCast) {
    182   return getAllocationData(V, AllocLike, TLI, LookThroughBitCast);
    183 }
    184 
    185 /// \brief Tests if a value is a call or invoke to a library function that
    186 /// reallocates memory (such as realloc).
    187 bool llvm::isReallocLikeFn(const Value *V, const TargetLibraryInfo *TLI,
    188                            bool LookThroughBitCast) {
    189   return getAllocationData(V, ReallocLike, TLI, LookThroughBitCast);
    190 }
    191 
    192 /// extractMallocCall - Returns the corresponding CallInst if the instruction
    193 /// is a malloc call.  Since CallInst::CreateMalloc() only creates calls, we
    194 /// ignore InvokeInst here.
    195 const CallInst *llvm::extractMallocCall(const Value *I,
    196                                         const TargetLibraryInfo *TLI) {
    197   return isMallocLikeFn(I, TLI) ? dyn_cast<CallInst>(I) : 0;
    198 }
    199 
    200 static Value *computeArraySize(const CallInst *CI, const DataLayout *TD,
    201                                const TargetLibraryInfo *TLI,
    202                                bool LookThroughSExt = false) {
    203   if (!CI)
    204     return 0;
    205 
    206   // The size of the malloc's result type must be known to determine array size.
    207   Type *T = getMallocAllocatedType(CI, TLI);
    208   if (!T || !T->isSized() || !TD)
    209     return 0;
    210 
    211   unsigned ElementSize = TD->getTypeAllocSize(T);
    212   if (StructType *ST = dyn_cast<StructType>(T))
    213     ElementSize = TD->getStructLayout(ST)->getSizeInBytes();
    214 
    215   // If malloc call's arg can be determined to be a multiple of ElementSize,
    216   // return the multiple.  Otherwise, return NULL.
    217   Value *MallocArg = CI->getArgOperand(0);
    218   Value *Multiple = 0;
    219   if (ComputeMultiple(MallocArg, ElementSize, Multiple,
    220                       LookThroughSExt))
    221     return Multiple;
    222 
    223   return 0;
    224 }
    225 
    226 /// isArrayMalloc - Returns the corresponding CallInst if the instruction
    227 /// is a call to malloc whose array size can be determined and the array size
    228 /// is not constant 1.  Otherwise, return NULL.
    229 const CallInst *llvm::isArrayMalloc(const Value *I,
    230                                     const DataLayout *TD,
    231                                     const TargetLibraryInfo *TLI) {
    232   const CallInst *CI = extractMallocCall(I, TLI);
    233   Value *ArraySize = computeArraySize(CI, TD, TLI);
    234 
    235   if (ConstantInt *ConstSize = dyn_cast_or_null<ConstantInt>(ArraySize))
    236     if (ConstSize->isOne())
    237       return CI;
    238 
    239   // CI is a non-array malloc or we can't figure out that it is an array malloc.
    240   return 0;
    241 }
    242 
    243 /// getMallocType - Returns the PointerType resulting from the malloc call.
    244 /// The PointerType depends on the number of bitcast uses of the malloc call:
    245 ///   0: PointerType is the calls' return type.
    246 ///   1: PointerType is the bitcast's result type.
    247 ///  >1: Unique PointerType cannot be determined, return NULL.
    248 PointerType *llvm::getMallocType(const CallInst *CI,
    249                                  const TargetLibraryInfo *TLI) {
    250   assert(isMallocLikeFn(CI, TLI) && "getMallocType and not malloc call");
    251 
    252   PointerType *MallocType = 0;
    253   unsigned NumOfBitCastUses = 0;
    254 
    255   // Determine if CallInst has a bitcast use.
    256   for (Value::const_use_iterator UI = CI->use_begin(), E = CI->use_end();
    257        UI != E; )
    258     if (const BitCastInst *BCI = dyn_cast<BitCastInst>(*UI++)) {
    259       MallocType = cast<PointerType>(BCI->getDestTy());
    260       NumOfBitCastUses++;
    261     }
    262 
    263   // Malloc call has 1 bitcast use, so type is the bitcast's destination type.
    264   if (NumOfBitCastUses == 1)
    265     return MallocType;
    266 
    267   // Malloc call was not bitcast, so type is the malloc function's return type.
    268   if (NumOfBitCastUses == 0)
    269     return cast<PointerType>(CI->getType());
    270 
    271   // Type could not be determined.
    272   return 0;
    273 }
    274 
    275 /// getMallocAllocatedType - Returns the Type allocated by malloc call.
    276 /// The Type depends on the number of bitcast uses of the malloc call:
    277 ///   0: PointerType is the malloc calls' return type.
    278 ///   1: PointerType is the bitcast's result type.
    279 ///  >1: Unique PointerType cannot be determined, return NULL.
    280 Type *llvm::getMallocAllocatedType(const CallInst *CI,
    281                                    const TargetLibraryInfo *TLI) {
    282   PointerType *PT = getMallocType(CI, TLI);
    283   return PT ? PT->getElementType() : 0;
    284 }
    285 
    286 /// getMallocArraySize - Returns the array size of a malloc call.  If the
    287 /// argument passed to malloc is a multiple of the size of the malloced type,
    288 /// then return that multiple.  For non-array mallocs, the multiple is
    289 /// constant 1.  Otherwise, return NULL for mallocs whose array size cannot be
    290 /// determined.
    291 Value *llvm::getMallocArraySize(CallInst *CI, const DataLayout *TD,
    292                                 const TargetLibraryInfo *TLI,
    293                                 bool LookThroughSExt) {
    294   assert(isMallocLikeFn(CI, TLI) && "getMallocArraySize and not malloc call");
    295   return computeArraySize(CI, TD, TLI, LookThroughSExt);
    296 }
    297 
    298 
    299 /// extractCallocCall - Returns the corresponding CallInst if the instruction
    300 /// is a calloc call.
    301 const CallInst *llvm::extractCallocCall(const Value *I,
    302                                         const TargetLibraryInfo *TLI) {
    303   return isCallocLikeFn(I, TLI) ? cast<CallInst>(I) : 0;
    304 }
    305 
    306 
    307 /// isFreeCall - Returns non-null if the value is a call to the builtin free()
    308 const CallInst *llvm::isFreeCall(const Value *I, const TargetLibraryInfo *TLI) {
    309   const CallInst *CI = dyn_cast<CallInst>(I);
    310   if (!CI || isa<IntrinsicInst>(CI))
    311     return 0;
    312   Function *Callee = CI->getCalledFunction();
    313   if (Callee == 0 || !Callee->isDeclaration())
    314     return 0;
    315 
    316   StringRef FnName = Callee->getName();
    317   LibFunc::Func TLIFn;
    318   if (!TLI || !TLI->getLibFunc(FnName, TLIFn) || !TLI->has(TLIFn))
    319     return 0;
    320 
    321   unsigned ExpectedNumParams;
    322   if (TLIFn == LibFunc::free ||
    323       TLIFn == LibFunc::ZdlPv || // operator delete(void*)
    324       TLIFn == LibFunc::ZdaPv)   // operator delete[](void*)
    325     ExpectedNumParams = 1;
    326   else if (TLIFn == LibFunc::ZdlPvRKSt9nothrow_t || // delete(void*, nothrow)
    327            TLIFn == LibFunc::ZdaPvRKSt9nothrow_t)   // delete[](void*, nothrow)
    328     ExpectedNumParams = 2;
    329   else
    330     return 0;
    331 
    332   // Check free prototype.
    333   // FIXME: workaround for PR5130, this will be obsolete when a nobuiltin
    334   // attribute will exist.
    335   FunctionType *FTy = Callee->getFunctionType();
    336   if (!FTy->getReturnType()->isVoidTy())
    337     return 0;
    338   if (FTy->getNumParams() != ExpectedNumParams)
    339     return 0;
    340   if (FTy->getParamType(0) != Type::getInt8PtrTy(Callee->getContext()))
    341     return 0;
    342 
    343   return CI;
    344 }
    345 
    346 
    347 
    348 //===----------------------------------------------------------------------===//
    349 //  Utility functions to compute size of objects.
    350 //
    351 
    352 
    353 /// \brief Compute the size of the object pointed by Ptr. Returns true and the
    354 /// object size in Size if successful, and false otherwise.
    355 /// If RoundToAlign is true, then Size is rounded up to the aligment of allocas,
    356 /// byval arguments, and global variables.
    357 bool llvm::getObjectSize(const Value *Ptr, uint64_t &Size, const DataLayout *TD,
    358                          const TargetLibraryInfo *TLI, bool RoundToAlign) {
    359   if (!TD)
    360     return false;
    361 
    362   ObjectSizeOffsetVisitor Visitor(TD, TLI, Ptr->getContext(), RoundToAlign);
    363   SizeOffsetType Data = Visitor.compute(const_cast<Value*>(Ptr));
    364   if (!Visitor.bothKnown(Data))
    365     return false;
    366 
    367   APInt ObjSize = Data.first, Offset = Data.second;
    368   // check for overflow
    369   if (Offset.slt(0) || ObjSize.ult(Offset))
    370     Size = 0;
    371   else
    372     Size = (ObjSize - Offset).getZExtValue();
    373   return true;
    374 }
    375 
    376 
    377 STATISTIC(ObjectVisitorArgument,
    378           "Number of arguments with unsolved size and offset");
    379 STATISTIC(ObjectVisitorLoad,
    380           "Number of load instructions with unsolved size and offset");
    381 
    382 
    383 APInt ObjectSizeOffsetVisitor::align(APInt Size, uint64_t Align) {
    384   if (RoundToAlign && Align)
    385     return APInt(IntTyBits, RoundUpToAlignment(Size.getZExtValue(), Align));
    386   return Size;
    387 }
    388 
    389 ObjectSizeOffsetVisitor::ObjectSizeOffsetVisitor(const DataLayout *TD,
    390                                                  const TargetLibraryInfo *TLI,
    391                                                  LLVMContext &Context,
    392                                                  bool RoundToAlign)
    393 : TD(TD), TLI(TLI), RoundToAlign(RoundToAlign) {
    394   IntegerType *IntTy = TD->getIntPtrType(Context);
    395   IntTyBits = IntTy->getBitWidth();
    396   Zero = APInt::getNullValue(IntTyBits);
    397 }
    398 
    399 SizeOffsetType ObjectSizeOffsetVisitor::compute(Value *V) {
    400   V = V->stripPointerCasts();
    401   if (Instruction *I = dyn_cast<Instruction>(V)) {
    402     // If we have already seen this instruction, bail out. Cycles can happen in
    403     // unreachable code after constant propagation.
    404     if (!SeenInsts.insert(I))
    405       return unknown();
    406 
    407     if (GEPOperator *GEP = dyn_cast<GEPOperator>(V))
    408       return visitGEPOperator(*GEP);
    409     return visit(*I);
    410   }
    411   if (Argument *A = dyn_cast<Argument>(V))
    412     return visitArgument(*A);
    413   if (ConstantPointerNull *P = dyn_cast<ConstantPointerNull>(V))
    414     return visitConstantPointerNull(*P);
    415   if (GlobalAlias *GA = dyn_cast<GlobalAlias>(V))
    416     return visitGlobalAlias(*GA);
    417   if (GlobalVariable *GV = dyn_cast<GlobalVariable>(V))
    418     return visitGlobalVariable(*GV);
    419   if (UndefValue *UV = dyn_cast<UndefValue>(V))
    420     return visitUndefValue(*UV);
    421   if (ConstantExpr *CE = dyn_cast<ConstantExpr>(V)) {
    422     if (CE->getOpcode() == Instruction::IntToPtr)
    423       return unknown(); // clueless
    424     if (CE->getOpcode() == Instruction::GetElementPtr)
    425       return visitGEPOperator(cast<GEPOperator>(*CE));
    426   }
    427 
    428   DEBUG(dbgs() << "ObjectSizeOffsetVisitor::compute() unhandled value: " << *V
    429         << '\n');
    430   return unknown();
    431 }
    432 
    433 SizeOffsetType ObjectSizeOffsetVisitor::visitAllocaInst(AllocaInst &I) {
    434   if (!I.getAllocatedType()->isSized())
    435     return unknown();
    436 
    437   APInt Size(IntTyBits, TD->getTypeAllocSize(I.getAllocatedType()));
    438   if (!I.isArrayAllocation())
    439     return std::make_pair(align(Size, I.getAlignment()), Zero);
    440 
    441   Value *ArraySize = I.getArraySize();
    442   if (const ConstantInt *C = dyn_cast<ConstantInt>(ArraySize)) {
    443     Size *= C->getValue().zextOrSelf(IntTyBits);
    444     return std::make_pair(align(Size, I.getAlignment()), Zero);
    445   }
    446   return unknown();
    447 }
    448 
    449 SizeOffsetType ObjectSizeOffsetVisitor::visitArgument(Argument &A) {
    450   // no interprocedural analysis is done at the moment
    451   if (!A.hasByValAttr()) {
    452     ++ObjectVisitorArgument;
    453     return unknown();
    454   }
    455   PointerType *PT = cast<PointerType>(A.getType());
    456   APInt Size(IntTyBits, TD->getTypeAllocSize(PT->getElementType()));
    457   return std::make_pair(align(Size, A.getParamAlignment()), Zero);
    458 }
    459 
    460 SizeOffsetType ObjectSizeOffsetVisitor::visitCallSite(CallSite CS) {
    461   const AllocFnsTy *FnData = getAllocationData(CS.getInstruction(), AnyAlloc,
    462                                                TLI);
    463   if (!FnData)
    464     return unknown();
    465 
    466   // handle strdup-like functions separately
    467   if (FnData->AllocTy == StrDupLike) {
    468     APInt Size(IntTyBits, GetStringLength(CS.getArgument(0)));
    469     if (!Size)
    470       return unknown();
    471 
    472     // strndup limits strlen
    473     if (FnData->FstParam > 0) {
    474       ConstantInt *Arg= dyn_cast<ConstantInt>(CS.getArgument(FnData->FstParam));
    475       if (!Arg)
    476         return unknown();
    477 
    478       APInt MaxSize = Arg->getValue().zextOrSelf(IntTyBits);
    479       if (Size.ugt(MaxSize))
    480         Size = MaxSize + 1;
    481     }
    482     return std::make_pair(Size, Zero);
    483   }
    484 
    485   ConstantInt *Arg = dyn_cast<ConstantInt>(CS.getArgument(FnData->FstParam));
    486   if (!Arg)
    487     return unknown();
    488 
    489   APInt Size = Arg->getValue().zextOrSelf(IntTyBits);
    490   // size determined by just 1 parameter
    491   if (FnData->SndParam < 0)
    492     return std::make_pair(Size, Zero);
    493 
    494   Arg = dyn_cast<ConstantInt>(CS.getArgument(FnData->SndParam));
    495   if (!Arg)
    496     return unknown();
    497 
    498   Size *= Arg->getValue().zextOrSelf(IntTyBits);
    499   return std::make_pair(Size, Zero);
    500 
    501   // TODO: handle more standard functions (+ wchar cousins):
    502   // - strdup / strndup
    503   // - strcpy / strncpy
    504   // - strcat / strncat
    505   // - memcpy / memmove
    506   // - strcat / strncat
    507   // - memset
    508 }
    509 
    510 SizeOffsetType
    511 ObjectSizeOffsetVisitor::visitConstantPointerNull(ConstantPointerNull&) {
    512   return std::make_pair(Zero, Zero);
    513 }
    514 
    515 SizeOffsetType
    516 ObjectSizeOffsetVisitor::visitExtractElementInst(ExtractElementInst&) {
    517   return unknown();
    518 }
    519 
    520 SizeOffsetType
    521 ObjectSizeOffsetVisitor::visitExtractValueInst(ExtractValueInst&) {
    522   // Easy cases were already folded by previous passes.
    523   return unknown();
    524 }
    525 
    526 SizeOffsetType ObjectSizeOffsetVisitor::visitGEPOperator(GEPOperator &GEP) {
    527   SizeOffsetType PtrData = compute(GEP.getPointerOperand());
    528   APInt Offset(IntTyBits, 0);
    529   if (!bothKnown(PtrData) || !GEP.accumulateConstantOffset(*TD, Offset))
    530     return unknown();
    531 
    532   return std::make_pair(PtrData.first, PtrData.second + Offset);
    533 }
    534 
    535 SizeOffsetType ObjectSizeOffsetVisitor::visitGlobalAlias(GlobalAlias &GA) {
    536   if (GA.mayBeOverridden())
    537     return unknown();
    538   return compute(GA.getAliasee());
    539 }
    540 
    541 SizeOffsetType ObjectSizeOffsetVisitor::visitGlobalVariable(GlobalVariable &GV){
    542   if (!GV.hasDefinitiveInitializer())
    543     return unknown();
    544 
    545   APInt Size(IntTyBits, TD->getTypeAllocSize(GV.getType()->getElementType()));
    546   return std::make_pair(align(Size, GV.getAlignment()), Zero);
    547 }
    548 
    549 SizeOffsetType ObjectSizeOffsetVisitor::visitIntToPtrInst(IntToPtrInst&) {
    550   // clueless
    551   return unknown();
    552 }
    553 
    554 SizeOffsetType ObjectSizeOffsetVisitor::visitLoadInst(LoadInst&) {
    555   ++ObjectVisitorLoad;
    556   return unknown();
    557 }
    558 
    559 SizeOffsetType ObjectSizeOffsetVisitor::visitPHINode(PHINode&) {
    560   // too complex to analyze statically.
    561   return unknown();
    562 }
    563 
    564 SizeOffsetType ObjectSizeOffsetVisitor::visitSelectInst(SelectInst &I) {
    565   SizeOffsetType TrueSide  = compute(I.getTrueValue());
    566   SizeOffsetType FalseSide = compute(I.getFalseValue());
    567   if (bothKnown(TrueSide) && bothKnown(FalseSide) && TrueSide == FalseSide)
    568     return TrueSide;
    569   return unknown();
    570 }
    571 
    572 SizeOffsetType ObjectSizeOffsetVisitor::visitUndefValue(UndefValue&) {
    573   return std::make_pair(Zero, Zero);
    574 }
    575 
    576 SizeOffsetType ObjectSizeOffsetVisitor::visitInstruction(Instruction &I) {
    577   DEBUG(dbgs() << "ObjectSizeOffsetVisitor unknown instruction:" << I << '\n');
    578   return unknown();
    579 }
    580 
    581 
    582 ObjectSizeOffsetEvaluator::ObjectSizeOffsetEvaluator(const DataLayout *TD,
    583                                                    const TargetLibraryInfo *TLI,
    584                                                      LLVMContext &Context)
    585 : TD(TD), TLI(TLI), Context(Context), Builder(Context, TargetFolder(TD)) {
    586   IntTy = TD->getIntPtrType(Context);
    587   Zero = ConstantInt::get(IntTy, 0);
    588 }
    589 
    590 SizeOffsetEvalType ObjectSizeOffsetEvaluator::compute(Value *V) {
    591   SizeOffsetEvalType Result = compute_(V);
    592 
    593   if (!bothKnown(Result)) {
    594     // erase everything that was computed in this iteration from the cache, so
    595     // that no dangling references are left behind. We could be a bit smarter if
    596     // we kept a dependency graph. It's probably not worth the complexity.
    597     for (PtrSetTy::iterator I=SeenVals.begin(), E=SeenVals.end(); I != E; ++I) {
    598       CacheMapTy::iterator CacheIt = CacheMap.find(*I);
    599       // non-computable results can be safely cached
    600       if (CacheIt != CacheMap.end() && anyKnown(CacheIt->second))
    601         CacheMap.erase(CacheIt);
    602     }
    603   }
    604 
    605   SeenVals.clear();
    606   return Result;
    607 }
    608 
    609 SizeOffsetEvalType ObjectSizeOffsetEvaluator::compute_(Value *V) {
    610   ObjectSizeOffsetVisitor Visitor(TD, TLI, Context);
    611   SizeOffsetType Const = Visitor.compute(V);
    612   if (Visitor.bothKnown(Const))
    613     return std::make_pair(ConstantInt::get(Context, Const.first),
    614                           ConstantInt::get(Context, Const.second));
    615 
    616   V = V->stripPointerCasts();
    617 
    618   // check cache
    619   CacheMapTy::iterator CacheIt = CacheMap.find(V);
    620   if (CacheIt != CacheMap.end())
    621     return CacheIt->second;
    622 
    623   // always generate code immediately before the instruction being
    624   // processed, so that the generated code dominates the same BBs
    625   Instruction *PrevInsertPoint = Builder.GetInsertPoint();
    626   if (Instruction *I = dyn_cast<Instruction>(V))
    627     Builder.SetInsertPoint(I);
    628 
    629   // record the pointers that were handled in this run, so that they can be
    630   // cleaned later if something fails
    631   SeenVals.insert(V);
    632 
    633   // now compute the size and offset
    634   SizeOffsetEvalType Result;
    635   if (GEPOperator *GEP = dyn_cast<GEPOperator>(V)) {
    636     Result = visitGEPOperator(*GEP);
    637   } else if (Instruction *I = dyn_cast<Instruction>(V)) {
    638     Result = visit(*I);
    639   } else if (isa<Argument>(V) ||
    640              (isa<ConstantExpr>(V) &&
    641               cast<ConstantExpr>(V)->getOpcode() == Instruction::IntToPtr) ||
    642              isa<GlobalAlias>(V) ||
    643              isa<GlobalVariable>(V)) {
    644     // ignore values where we cannot do more than what ObjectSizeVisitor can
    645     Result = unknown();
    646   } else {
    647     DEBUG(dbgs() << "ObjectSizeOffsetEvaluator::compute() unhandled value: "
    648           << *V << '\n');
    649     Result = unknown();
    650   }
    651 
    652   if (PrevInsertPoint)
    653     Builder.SetInsertPoint(PrevInsertPoint);
    654 
    655   // Don't reuse CacheIt since it may be invalid at this point.
    656   CacheMap[V] = Result;
    657   return Result;
    658 }
    659 
    660 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitAllocaInst(AllocaInst &I) {
    661   if (!I.getAllocatedType()->isSized())
    662     return unknown();
    663 
    664   // must be a VLA
    665   assert(I.isArrayAllocation());
    666   Value *ArraySize = I.getArraySize();
    667   Value *Size = ConstantInt::get(ArraySize->getType(),
    668                                  TD->getTypeAllocSize(I.getAllocatedType()));
    669   Size = Builder.CreateMul(Size, ArraySize);
    670   return std::make_pair(Size, Zero);
    671 }
    672 
    673 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitCallSite(CallSite CS) {
    674   const AllocFnsTy *FnData = getAllocationData(CS.getInstruction(), AnyAlloc,
    675                                                TLI);
    676   if (!FnData)
    677     return unknown();
    678 
    679   // handle strdup-like functions separately
    680   if (FnData->AllocTy == StrDupLike) {
    681     // TODO
    682     return unknown();
    683   }
    684 
    685   Value *FirstArg = CS.getArgument(FnData->FstParam);
    686   FirstArg = Builder.CreateZExt(FirstArg, IntTy);
    687   if (FnData->SndParam < 0)
    688     return std::make_pair(FirstArg, Zero);
    689 
    690   Value *SecondArg = CS.getArgument(FnData->SndParam);
    691   SecondArg = Builder.CreateZExt(SecondArg, IntTy);
    692   Value *Size = Builder.CreateMul(FirstArg, SecondArg);
    693   return std::make_pair(Size, Zero);
    694 
    695   // TODO: handle more standard functions (+ wchar cousins):
    696   // - strdup / strndup
    697   // - strcpy / strncpy
    698   // - strcat / strncat
    699   // - memcpy / memmove
    700   // - strcat / strncat
    701   // - memset
    702 }
    703 
    704 SizeOffsetEvalType
    705 ObjectSizeOffsetEvaluator::visitExtractElementInst(ExtractElementInst&) {
    706   return unknown();
    707 }
    708 
    709 SizeOffsetEvalType
    710 ObjectSizeOffsetEvaluator::visitExtractValueInst(ExtractValueInst&) {
    711   return unknown();
    712 }
    713 
    714 SizeOffsetEvalType
    715 ObjectSizeOffsetEvaluator::visitGEPOperator(GEPOperator &GEP) {
    716   SizeOffsetEvalType PtrData = compute_(GEP.getPointerOperand());
    717   if (!bothKnown(PtrData))
    718     return unknown();
    719 
    720   Value *Offset = EmitGEPOffset(&Builder, *TD, &GEP, /*NoAssumptions=*/true);
    721   Offset = Builder.CreateAdd(PtrData.second, Offset);
    722   return std::make_pair(PtrData.first, Offset);
    723 }
    724 
    725 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitIntToPtrInst(IntToPtrInst&) {
    726   // clueless
    727   return unknown();
    728 }
    729 
    730 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitLoadInst(LoadInst&) {
    731   return unknown();
    732 }
    733 
    734 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitPHINode(PHINode &PHI) {
    735   // create 2 PHIs: one for size and another for offset
    736   PHINode *SizePHI   = Builder.CreatePHI(IntTy, PHI.getNumIncomingValues());
    737   PHINode *OffsetPHI = Builder.CreatePHI(IntTy, PHI.getNumIncomingValues());
    738 
    739   // insert right away in the cache to handle recursive PHIs
    740   CacheMap[&PHI] = std::make_pair(SizePHI, OffsetPHI);
    741 
    742   // compute offset/size for each PHI incoming pointer
    743   for (unsigned i = 0, e = PHI.getNumIncomingValues(); i != e; ++i) {
    744     Builder.SetInsertPoint(PHI.getIncomingBlock(i)->getFirstInsertionPt());
    745     SizeOffsetEvalType EdgeData = compute_(PHI.getIncomingValue(i));
    746 
    747     if (!bothKnown(EdgeData)) {
    748       OffsetPHI->replaceAllUsesWith(UndefValue::get(IntTy));
    749       OffsetPHI->eraseFromParent();
    750       SizePHI->replaceAllUsesWith(UndefValue::get(IntTy));
    751       SizePHI->eraseFromParent();
    752       return unknown();
    753     }
    754     SizePHI->addIncoming(EdgeData.first, PHI.getIncomingBlock(i));
    755     OffsetPHI->addIncoming(EdgeData.second, PHI.getIncomingBlock(i));
    756   }
    757 
    758   Value *Size = SizePHI, *Offset = OffsetPHI, *Tmp;
    759   if ((Tmp = SizePHI->hasConstantValue())) {
    760     Size = Tmp;
    761     SizePHI->replaceAllUsesWith(Size);
    762     SizePHI->eraseFromParent();
    763   }
    764   if ((Tmp = OffsetPHI->hasConstantValue())) {
    765     Offset = Tmp;
    766     OffsetPHI->replaceAllUsesWith(Offset);
    767     OffsetPHI->eraseFromParent();
    768   }
    769   return std::make_pair(Size, Offset);
    770 }
    771 
    772 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitSelectInst(SelectInst &I) {
    773   SizeOffsetEvalType TrueSide  = compute_(I.getTrueValue());
    774   SizeOffsetEvalType FalseSide = compute_(I.getFalseValue());
    775 
    776   if (!bothKnown(TrueSide) || !bothKnown(FalseSide))
    777     return unknown();
    778   if (TrueSide == FalseSide)
    779     return TrueSide;
    780 
    781   Value *Size = Builder.CreateSelect(I.getCondition(), TrueSide.first,
    782                                      FalseSide.first);
    783   Value *Offset = Builder.CreateSelect(I.getCondition(), TrueSide.second,
    784                                        FalseSide.second);
    785   return std::make_pair(Size, Offset);
    786 }
    787 
    788 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitInstruction(Instruction &I) {
    789   DEBUG(dbgs() << "ObjectSizeOffsetEvaluator unknown instruction:" << I <<'\n');
    790   return unknown();
    791 }
    792