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/ADT/Statistic.h"
     17 #include "llvm/ADT/STLExtras.h"
     18 #include "llvm/Analysis/MemoryBuiltins.h"
     19 #include "llvm/GlobalVariable.h"
     20 #include "llvm/Instructions.h"
     21 #include "llvm/Intrinsics.h"
     22 #include "llvm/Metadata.h"
     23 #include "llvm/Module.h"
     24 #include "llvm/Analysis/ValueTracking.h"
     25 #include "llvm/Support/Debug.h"
     26 #include "llvm/Support/MathExtras.h"
     27 #include "llvm/Support/raw_ostream.h"
     28 #include "llvm/Target/TargetData.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   Function *Callee = CS.getCalledFunction();
     81   if (!Callee || !Callee->isDeclaration())
     82     return 0;
     83   return Callee;
     84 }
     85 
     86 /// \brief Returns the allocation data for the given value if it is a call to a
     87 /// known allocation function, and NULL otherwise.
     88 static const AllocFnsTy *getAllocationData(const Value *V, AllocType AllocTy,
     89                                            const TargetLibraryInfo *TLI,
     90                                            bool LookThroughBitCast = false) {
     91   Function *Callee = getCalledFunction(V, LookThroughBitCast);
     92   if (!Callee)
     93     return 0;
     94 
     95   // Make sure that the function is available.
     96   StringRef FnName = Callee->getName();
     97   LibFunc::Func TLIFn;
     98   if (!TLI || !TLI->getLibFunc(FnName, TLIFn) || !TLI->has(TLIFn))
     99     return 0;
    100 
    101   unsigned i = 0;
    102   bool found = false;
    103   for ( ; i < array_lengthof(AllocationFnData); ++i) {
    104     if (AllocationFnData[i].Func == TLIFn) {
    105       found = true;
    106       break;
    107     }
    108   }
    109   if (!found)
    110     return 0;
    111 
    112   const AllocFnsTy *FnData = &AllocationFnData[i];
    113   if ((FnData->AllocTy & AllocTy) == 0)
    114     return 0;
    115 
    116   // Check function prototype.
    117   int FstParam = FnData->FstParam;
    118   int SndParam = FnData->SndParam;
    119   FunctionType *FTy = Callee->getFunctionType();
    120 
    121   if (FTy->getReturnType() == Type::getInt8PtrTy(FTy->getContext()) &&
    122       FTy->getNumParams() == FnData->NumParams &&
    123       (FstParam < 0 ||
    124        (FTy->getParamType(FstParam)->isIntegerTy(32) ||
    125         FTy->getParamType(FstParam)->isIntegerTy(64))) &&
    126       (SndParam < 0 ||
    127        FTy->getParamType(SndParam)->isIntegerTy(32) ||
    128        FTy->getParamType(SndParam)->isIntegerTy(64)))
    129     return FnData;
    130   return 0;
    131 }
    132 
    133 static bool hasNoAliasAttr(const Value *V, bool LookThroughBitCast) {
    134   ImmutableCallSite CS(LookThroughBitCast ? V->stripPointerCasts() : V);
    135   return CS && CS.hasFnAttr(Attribute::NoAlias);
    136 }
    137 
    138 
    139 /// \brief Tests if a value is a call or invoke to a library function that
    140 /// allocates or reallocates memory (either malloc, calloc, realloc, or strdup
    141 /// like).
    142 bool llvm::isAllocationFn(const Value *V, const TargetLibraryInfo *TLI,
    143                           bool LookThroughBitCast) {
    144   return getAllocationData(V, AnyAlloc, TLI, LookThroughBitCast);
    145 }
    146 
    147 /// \brief Tests if a value is a call or invoke to a function that returns a
    148 /// NoAlias pointer (including malloc/calloc/realloc/strdup-like functions).
    149 bool llvm::isNoAliasFn(const Value *V, const TargetLibraryInfo *TLI,
    150                        bool LookThroughBitCast) {
    151   // it's safe to consider realloc as noalias since accessing the original
    152   // pointer is undefined behavior
    153   return isAllocationFn(V, TLI, LookThroughBitCast) ||
    154          hasNoAliasAttr(V, LookThroughBitCast);
    155 }
    156 
    157 /// \brief Tests if a value is a call or invoke to a library function that
    158 /// allocates uninitialized memory (such as malloc).
    159 bool llvm::isMallocLikeFn(const Value *V, const TargetLibraryInfo *TLI,
    160                           bool LookThroughBitCast) {
    161   return getAllocationData(V, MallocLike, TLI, LookThroughBitCast);
    162 }
    163 
    164 /// \brief Tests if a value is a call or invoke to a library function that
    165 /// allocates zero-filled memory (such as calloc).
    166 bool llvm::isCallocLikeFn(const Value *V, const TargetLibraryInfo *TLI,
    167                           bool LookThroughBitCast) {
    168   return getAllocationData(V, CallocLike, TLI, LookThroughBitCast);
    169 }
    170 
    171 /// \brief Tests if a value is a call or invoke to a library function that
    172 /// allocates memory (either malloc, calloc, or strdup like).
    173 bool llvm::isAllocLikeFn(const Value *V, const TargetLibraryInfo *TLI,
    174                          bool LookThroughBitCast) {
    175   return getAllocationData(V, AllocLike, TLI, LookThroughBitCast);
    176 }
    177 
    178 /// \brief Tests if a value is a call or invoke to a library function that
    179 /// reallocates memory (such as realloc).
    180 bool llvm::isReallocLikeFn(const Value *V, const TargetLibraryInfo *TLI,
    181                            bool LookThroughBitCast) {
    182   return getAllocationData(V, ReallocLike, TLI, LookThroughBitCast);
    183 }
    184 
    185 /// extractMallocCall - Returns the corresponding CallInst if the instruction
    186 /// is a malloc call.  Since CallInst::CreateMalloc() only creates calls, we
    187 /// ignore InvokeInst here.
    188 const CallInst *llvm::extractMallocCall(const Value *I,
    189                                         const TargetLibraryInfo *TLI) {
    190   return isMallocLikeFn(I, TLI) ? dyn_cast<CallInst>(I) : 0;
    191 }
    192 
    193 static Value *computeArraySize(const CallInst *CI, const TargetData *TD,
    194                                const TargetLibraryInfo *TLI,
    195                                bool LookThroughSExt = false) {
    196   if (!CI)
    197     return NULL;
    198 
    199   // The size of the malloc's result type must be known to determine array size.
    200   Type *T = getMallocAllocatedType(CI, TLI);
    201   if (!T || !T->isSized() || !TD)
    202     return NULL;
    203 
    204   unsigned ElementSize = TD->getTypeAllocSize(T);
    205   if (StructType *ST = dyn_cast<StructType>(T))
    206     ElementSize = TD->getStructLayout(ST)->getSizeInBytes();
    207 
    208   // If malloc call's arg can be determined to be a multiple of ElementSize,
    209   // return the multiple.  Otherwise, return NULL.
    210   Value *MallocArg = CI->getArgOperand(0);
    211   Value *Multiple = NULL;
    212   if (ComputeMultiple(MallocArg, ElementSize, Multiple,
    213                       LookThroughSExt))
    214     return Multiple;
    215 
    216   return NULL;
    217 }
    218 
    219 /// isArrayMalloc - Returns the corresponding CallInst if the instruction
    220 /// is a call to malloc whose array size can be determined and the array size
    221 /// is not constant 1.  Otherwise, return NULL.
    222 const CallInst *llvm::isArrayMalloc(const Value *I,
    223                                     const TargetData *TD,
    224                                     const TargetLibraryInfo *TLI) {
    225   const CallInst *CI = extractMallocCall(I, TLI);
    226   Value *ArraySize = computeArraySize(CI, TD, TLI);
    227 
    228   if (ArraySize &&
    229       ArraySize != ConstantInt::get(CI->getArgOperand(0)->getType(), 1))
    230     return CI;
    231 
    232   // CI is a non-array malloc or we can't figure out that it is an array malloc.
    233   return NULL;
    234 }
    235 
    236 /// getMallocType - Returns the PointerType resulting from the malloc call.
    237 /// The PointerType depends on the number of bitcast uses of the malloc call:
    238 ///   0: PointerType is the calls' return type.
    239 ///   1: PointerType is the bitcast's result type.
    240 ///  >1: Unique PointerType cannot be determined, return NULL.
    241 PointerType *llvm::getMallocType(const CallInst *CI,
    242                                  const TargetLibraryInfo *TLI) {
    243   assert(isMallocLikeFn(CI, TLI) && "getMallocType and not malloc call");
    244 
    245   PointerType *MallocType = NULL;
    246   unsigned NumOfBitCastUses = 0;
    247 
    248   // Determine if CallInst has a bitcast use.
    249   for (Value::const_use_iterator UI = CI->use_begin(), E = CI->use_end();
    250        UI != E; )
    251     if (const BitCastInst *BCI = dyn_cast<BitCastInst>(*UI++)) {
    252       MallocType = cast<PointerType>(BCI->getDestTy());
    253       NumOfBitCastUses++;
    254     }
    255 
    256   // Malloc call has 1 bitcast use, so type is the bitcast's destination type.
    257   if (NumOfBitCastUses == 1)
    258     return MallocType;
    259 
    260   // Malloc call was not bitcast, so type is the malloc function's return type.
    261   if (NumOfBitCastUses == 0)
    262     return cast<PointerType>(CI->getType());
    263 
    264   // Type could not be determined.
    265   return NULL;
    266 }
    267 
    268 /// getMallocAllocatedType - Returns the Type allocated by malloc call.
    269 /// The Type depends on the number of bitcast uses of the malloc call:
    270 ///   0: PointerType is the malloc calls' return type.
    271 ///   1: PointerType is the bitcast's result type.
    272 ///  >1: Unique PointerType cannot be determined, return NULL.
    273 Type *llvm::getMallocAllocatedType(const CallInst *CI,
    274                                    const TargetLibraryInfo *TLI) {
    275   PointerType *PT = getMallocType(CI, TLI);
    276   return PT ? PT->getElementType() : NULL;
    277 }
    278 
    279 /// getMallocArraySize - Returns the array size of a malloc call.  If the
    280 /// argument passed to malloc is a multiple of the size of the malloced type,
    281 /// then return that multiple.  For non-array mallocs, the multiple is
    282 /// constant 1.  Otherwise, return NULL for mallocs whose array size cannot be
    283 /// determined.
    284 Value *llvm::getMallocArraySize(CallInst *CI, const TargetData *TD,
    285                                 const TargetLibraryInfo *TLI,
    286                                 bool LookThroughSExt) {
    287   assert(isMallocLikeFn(CI, TLI) && "getMallocArraySize and not malloc call");
    288   return computeArraySize(CI, TD, TLI, LookThroughSExt);
    289 }
    290 
    291 
    292 /// extractCallocCall - Returns the corresponding CallInst if the instruction
    293 /// is a calloc call.
    294 const CallInst *llvm::extractCallocCall(const Value *I,
    295                                         const TargetLibraryInfo *TLI) {
    296   return isCallocLikeFn(I, TLI) ? cast<CallInst>(I) : 0;
    297 }
    298 
    299 
    300 /// isFreeCall - Returns non-null if the value is a call to the builtin free()
    301 const CallInst *llvm::isFreeCall(const Value *I, const TargetLibraryInfo *TLI) {
    302   const CallInst *CI = dyn_cast<CallInst>(I);
    303   if (!CI)
    304     return 0;
    305   Function *Callee = CI->getCalledFunction();
    306   if (Callee == 0 || !Callee->isDeclaration())
    307     return 0;
    308 
    309   StringRef FnName = Callee->getName();
    310   LibFunc::Func TLIFn;
    311   if (!TLI || !TLI->getLibFunc(FnName, TLIFn) || !TLI->has(TLIFn))
    312     return 0;
    313 
    314   if (TLIFn != LibFunc::free &&
    315       TLIFn != LibFunc::ZdlPv && // operator delete(void*)
    316       TLIFn != LibFunc::ZdaPv)   // operator delete[](void*)
    317     return 0;
    318 
    319   // Check free prototype.
    320   // FIXME: workaround for PR5130, this will be obsolete when a nobuiltin
    321   // attribute will exist.
    322   FunctionType *FTy = Callee->getFunctionType();
    323   if (!FTy->getReturnType()->isVoidTy())
    324     return 0;
    325   if (FTy->getNumParams() != 1)
    326     return 0;
    327   if (FTy->getParamType(0) != Type::getInt8PtrTy(Callee->getContext()))
    328     return 0;
    329 
    330   return CI;
    331 }
    332 
    333 
    334 
    335 //===----------------------------------------------------------------------===//
    336 //  Utility functions to compute size of objects.
    337 //
    338 
    339 
    340 /// \brief Compute the size of the object pointed by Ptr. Returns true and the
    341 /// object size in Size if successful, and false otherwise.
    342 /// If RoundToAlign is true, then Size is rounded up to the aligment of allocas,
    343 /// byval arguments, and global variables.
    344 bool llvm::getObjectSize(const Value *Ptr, uint64_t &Size, const TargetData *TD,
    345                          const TargetLibraryInfo *TLI, bool RoundToAlign) {
    346   if (!TD)
    347     return false;
    348 
    349   ObjectSizeOffsetVisitor Visitor(TD, TLI, Ptr->getContext(), RoundToAlign);
    350   SizeOffsetType Data = Visitor.compute(const_cast<Value*>(Ptr));
    351   if (!Visitor.bothKnown(Data))
    352     return false;
    353 
    354   APInt ObjSize = Data.first, Offset = Data.second;
    355   // check for overflow
    356   if (Offset.slt(0) || ObjSize.ult(Offset))
    357     Size = 0;
    358   else
    359     Size = (ObjSize - Offset).getZExtValue();
    360   return true;
    361 }
    362 
    363 
    364 STATISTIC(ObjectVisitorArgument,
    365           "Number of arguments with unsolved size and offset");
    366 STATISTIC(ObjectVisitorLoad,
    367           "Number of load instructions with unsolved size and offset");
    368 
    369 
    370 APInt ObjectSizeOffsetVisitor::align(APInt Size, uint64_t Align) {
    371   if (RoundToAlign && Align)
    372     return APInt(IntTyBits, RoundUpToAlignment(Size.getZExtValue(), Align));
    373   return Size;
    374 }
    375 
    376 ObjectSizeOffsetVisitor::ObjectSizeOffsetVisitor(const TargetData *TD,
    377                                                  const TargetLibraryInfo *TLI,
    378                                                  LLVMContext &Context,
    379                                                  bool RoundToAlign)
    380 : TD(TD), TLI(TLI), RoundToAlign(RoundToAlign) {
    381   IntegerType *IntTy = TD->getIntPtrType(Context);
    382   IntTyBits = IntTy->getBitWidth();
    383   Zero = APInt::getNullValue(IntTyBits);
    384 }
    385 
    386 SizeOffsetType ObjectSizeOffsetVisitor::compute(Value *V) {
    387   V = V->stripPointerCasts();
    388   if (Instruction *I = dyn_cast<Instruction>(V)) {
    389     // If we have already seen this instruction, bail out. Cycles can happen in
    390     // unreachable code after constant propagation.
    391     if (!SeenInsts.insert(I))
    392       return unknown();
    393 
    394     if (GEPOperator *GEP = dyn_cast<GEPOperator>(V))
    395       return visitGEPOperator(*GEP);
    396     return visit(*I);
    397   }
    398   if (Argument *A = dyn_cast<Argument>(V))
    399     return visitArgument(*A);
    400   if (ConstantPointerNull *P = dyn_cast<ConstantPointerNull>(V))
    401     return visitConstantPointerNull(*P);
    402   if (GlobalVariable *GV = dyn_cast<GlobalVariable>(V))
    403     return visitGlobalVariable(*GV);
    404   if (UndefValue *UV = dyn_cast<UndefValue>(V))
    405     return visitUndefValue(*UV);
    406   if (ConstantExpr *CE = dyn_cast<ConstantExpr>(V)) {
    407     if (CE->getOpcode() == Instruction::IntToPtr)
    408       return unknown(); // clueless
    409     if (CE->getOpcode() == Instruction::GetElementPtr)
    410       return visitGEPOperator(cast<GEPOperator>(*CE));
    411   }
    412 
    413   DEBUG(dbgs() << "ObjectSizeOffsetVisitor::compute() unhandled value: " << *V
    414         << '\n');
    415   return unknown();
    416 }
    417 
    418 SizeOffsetType ObjectSizeOffsetVisitor::visitAllocaInst(AllocaInst &I) {
    419   if (!I.getAllocatedType()->isSized())
    420     return unknown();
    421 
    422   APInt Size(IntTyBits, TD->getTypeAllocSize(I.getAllocatedType()));
    423   if (!I.isArrayAllocation())
    424     return std::make_pair(align(Size, I.getAlignment()), Zero);
    425 
    426   Value *ArraySize = I.getArraySize();
    427   if (const ConstantInt *C = dyn_cast<ConstantInt>(ArraySize)) {
    428     Size *= C->getValue().zextOrSelf(IntTyBits);
    429     return std::make_pair(align(Size, I.getAlignment()), Zero);
    430   }
    431   return unknown();
    432 }
    433 
    434 SizeOffsetType ObjectSizeOffsetVisitor::visitArgument(Argument &A) {
    435   // no interprocedural analysis is done at the moment
    436   if (!A.hasByValAttr()) {
    437     ++ObjectVisitorArgument;
    438     return unknown();
    439   }
    440   PointerType *PT = cast<PointerType>(A.getType());
    441   APInt Size(IntTyBits, TD->getTypeAllocSize(PT->getElementType()));
    442   return std::make_pair(align(Size, A.getParamAlignment()), Zero);
    443 }
    444 
    445 SizeOffsetType ObjectSizeOffsetVisitor::visitCallSite(CallSite CS) {
    446   const AllocFnsTy *FnData = getAllocationData(CS.getInstruction(), AnyAlloc,
    447                                                TLI);
    448   if (!FnData)
    449     return unknown();
    450 
    451   // handle strdup-like functions separately
    452   if (FnData->AllocTy == StrDupLike) {
    453     APInt Size(IntTyBits, GetStringLength(CS.getArgument(0)));
    454     if (!Size)
    455       return unknown();
    456 
    457     // strndup limits strlen
    458     if (FnData->FstParam > 0) {
    459       ConstantInt *Arg= dyn_cast<ConstantInt>(CS.getArgument(FnData->FstParam));
    460       if (!Arg)
    461         return unknown();
    462 
    463       APInt MaxSize = Arg->getValue().zextOrSelf(IntTyBits);
    464       if (Size.ugt(MaxSize))
    465         Size = MaxSize + 1;
    466     }
    467     return std::make_pair(Size, Zero);
    468   }
    469 
    470   ConstantInt *Arg = dyn_cast<ConstantInt>(CS.getArgument(FnData->FstParam));
    471   if (!Arg)
    472     return unknown();
    473 
    474   APInt Size = Arg->getValue().zextOrSelf(IntTyBits);
    475   // size determined by just 1 parameter
    476   if (FnData->SndParam < 0)
    477     return std::make_pair(Size, Zero);
    478 
    479   Arg = dyn_cast<ConstantInt>(CS.getArgument(FnData->SndParam));
    480   if (!Arg)
    481     return unknown();
    482 
    483   Size *= Arg->getValue().zextOrSelf(IntTyBits);
    484   return std::make_pair(Size, Zero);
    485 
    486   // TODO: handle more standard functions (+ wchar cousins):
    487   // - strdup / strndup
    488   // - strcpy / strncpy
    489   // - strcat / strncat
    490   // - memcpy / memmove
    491   // - strcat / strncat
    492   // - memset
    493 }
    494 
    495 SizeOffsetType
    496 ObjectSizeOffsetVisitor::visitConstantPointerNull(ConstantPointerNull&) {
    497   return std::make_pair(Zero, Zero);
    498 }
    499 
    500 SizeOffsetType
    501 ObjectSizeOffsetVisitor::visitExtractElementInst(ExtractElementInst&) {
    502   return unknown();
    503 }
    504 
    505 SizeOffsetType
    506 ObjectSizeOffsetVisitor::visitExtractValueInst(ExtractValueInst&) {
    507   // Easy cases were already folded by previous passes.
    508   return unknown();
    509 }
    510 
    511 SizeOffsetType ObjectSizeOffsetVisitor::visitGEPOperator(GEPOperator &GEP) {
    512   SizeOffsetType PtrData = compute(GEP.getPointerOperand());
    513   if (!bothKnown(PtrData) || !GEP.hasAllConstantIndices())
    514     return unknown();
    515 
    516   SmallVector<Value*, 8> Ops(GEP.idx_begin(), GEP.idx_end());
    517   APInt Offset(IntTyBits,TD->getIndexedOffset(GEP.getPointerOperandType(),Ops));
    518   return std::make_pair(PtrData.first, PtrData.second + Offset);
    519 }
    520 
    521 SizeOffsetType ObjectSizeOffsetVisitor::visitGlobalVariable(GlobalVariable &GV){
    522   if (!GV.hasDefinitiveInitializer())
    523     return unknown();
    524 
    525   APInt Size(IntTyBits, TD->getTypeAllocSize(GV.getType()->getElementType()));
    526   return std::make_pair(align(Size, GV.getAlignment()), Zero);
    527 }
    528 
    529 SizeOffsetType ObjectSizeOffsetVisitor::visitIntToPtrInst(IntToPtrInst&) {
    530   // clueless
    531   return unknown();
    532 }
    533 
    534 SizeOffsetType ObjectSizeOffsetVisitor::visitLoadInst(LoadInst&) {
    535   ++ObjectVisitorLoad;
    536   return unknown();
    537 }
    538 
    539 SizeOffsetType ObjectSizeOffsetVisitor::visitPHINode(PHINode&) {
    540   // too complex to analyze statically.
    541   return unknown();
    542 }
    543 
    544 SizeOffsetType ObjectSizeOffsetVisitor::visitSelectInst(SelectInst &I) {
    545   SizeOffsetType TrueSide  = compute(I.getTrueValue());
    546   SizeOffsetType FalseSide = compute(I.getFalseValue());
    547   if (bothKnown(TrueSide) && bothKnown(FalseSide) && TrueSide == FalseSide)
    548     return TrueSide;
    549   return unknown();
    550 }
    551 
    552 SizeOffsetType ObjectSizeOffsetVisitor::visitUndefValue(UndefValue&) {
    553   return std::make_pair(Zero, Zero);
    554 }
    555 
    556 SizeOffsetType ObjectSizeOffsetVisitor::visitInstruction(Instruction &I) {
    557   DEBUG(dbgs() << "ObjectSizeOffsetVisitor unknown instruction:" << I << '\n');
    558   return unknown();
    559 }
    560 
    561 
    562 ObjectSizeOffsetEvaluator::ObjectSizeOffsetEvaluator(const TargetData *TD,
    563                                                    const TargetLibraryInfo *TLI,
    564                                                      LLVMContext &Context)
    565 : TD(TD), TLI(TLI), Context(Context), Builder(Context, TargetFolder(TD)) {
    566   IntTy = TD->getIntPtrType(Context);
    567   Zero = ConstantInt::get(IntTy, 0);
    568 }
    569 
    570 SizeOffsetEvalType ObjectSizeOffsetEvaluator::compute(Value *V) {
    571   SizeOffsetEvalType Result = compute_(V);
    572 
    573   if (!bothKnown(Result)) {
    574     // erase everything that was computed in this iteration from the cache, so
    575     // that no dangling references are left behind. We could be a bit smarter if
    576     // we kept a dependency graph. It's probably not worth the complexity.
    577     for (PtrSetTy::iterator I=SeenVals.begin(), E=SeenVals.end(); I != E; ++I) {
    578       CacheMapTy::iterator CacheIt = CacheMap.find(*I);
    579       // non-computable results can be safely cached
    580       if (CacheIt != CacheMap.end() && anyKnown(CacheIt->second))
    581         CacheMap.erase(CacheIt);
    582     }
    583   }
    584 
    585   SeenVals.clear();
    586   return Result;
    587 }
    588 
    589 SizeOffsetEvalType ObjectSizeOffsetEvaluator::compute_(Value *V) {
    590   ObjectSizeOffsetVisitor Visitor(TD, TLI, Context);
    591   SizeOffsetType Const = Visitor.compute(V);
    592   if (Visitor.bothKnown(Const))
    593     return std::make_pair(ConstantInt::get(Context, Const.first),
    594                           ConstantInt::get(Context, Const.second));
    595 
    596   V = V->stripPointerCasts();
    597 
    598   // check cache
    599   CacheMapTy::iterator CacheIt = CacheMap.find(V);
    600   if (CacheIt != CacheMap.end())
    601     return CacheIt->second;
    602 
    603   // always generate code immediately before the instruction being
    604   // processed, so that the generated code dominates the same BBs
    605   Instruction *PrevInsertPoint = Builder.GetInsertPoint();
    606   if (Instruction *I = dyn_cast<Instruction>(V))
    607     Builder.SetInsertPoint(I);
    608 
    609   // record the pointers that were handled in this run, so that they can be
    610   // cleaned later if something fails
    611   SeenVals.insert(V);
    612 
    613   // now compute the size and offset
    614   SizeOffsetEvalType Result;
    615   if (GEPOperator *GEP = dyn_cast<GEPOperator>(V)) {
    616     Result = visitGEPOperator(*GEP);
    617   } else if (Instruction *I = dyn_cast<Instruction>(V)) {
    618     Result = visit(*I);
    619   } else if (isa<Argument>(V) ||
    620              (isa<ConstantExpr>(V) &&
    621               cast<ConstantExpr>(V)->getOpcode() == Instruction::IntToPtr) ||
    622              isa<GlobalVariable>(V)) {
    623     // ignore values where we cannot do more than what ObjectSizeVisitor can
    624     Result = unknown();
    625   } else {
    626     DEBUG(dbgs() << "ObjectSizeOffsetEvaluator::compute() unhandled value: "
    627           << *V << '\n');
    628     Result = unknown();
    629   }
    630 
    631   if (PrevInsertPoint)
    632     Builder.SetInsertPoint(PrevInsertPoint);
    633 
    634   // Don't reuse CacheIt since it may be invalid at this point.
    635   CacheMap[V] = Result;
    636   return Result;
    637 }
    638 
    639 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitAllocaInst(AllocaInst &I) {
    640   if (!I.getAllocatedType()->isSized())
    641     return unknown();
    642 
    643   // must be a VLA
    644   assert(I.isArrayAllocation());
    645   Value *ArraySize = I.getArraySize();
    646   Value *Size = ConstantInt::get(ArraySize->getType(),
    647                                  TD->getTypeAllocSize(I.getAllocatedType()));
    648   Size = Builder.CreateMul(Size, ArraySize);
    649   return std::make_pair(Size, Zero);
    650 }
    651 
    652 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitCallSite(CallSite CS) {
    653   const AllocFnsTy *FnData = getAllocationData(CS.getInstruction(), AnyAlloc,
    654                                                TLI);
    655   if (!FnData)
    656     return unknown();
    657 
    658   // handle strdup-like functions separately
    659   if (FnData->AllocTy == StrDupLike) {
    660     // TODO
    661     return unknown();
    662   }
    663 
    664   Value *FirstArg = CS.getArgument(FnData->FstParam);
    665   FirstArg = Builder.CreateZExt(FirstArg, IntTy);
    666   if (FnData->SndParam < 0)
    667     return std::make_pair(FirstArg, Zero);
    668 
    669   Value *SecondArg = CS.getArgument(FnData->SndParam);
    670   SecondArg = Builder.CreateZExt(SecondArg, IntTy);
    671   Value *Size = Builder.CreateMul(FirstArg, SecondArg);
    672   return std::make_pair(Size, Zero);
    673 
    674   // TODO: handle more standard functions (+ wchar cousins):
    675   // - strdup / strndup
    676   // - strcpy / strncpy
    677   // - strcat / strncat
    678   // - memcpy / memmove
    679   // - strcat / strncat
    680   // - memset
    681 }
    682 
    683 SizeOffsetEvalType
    684 ObjectSizeOffsetEvaluator::visitExtractElementInst(ExtractElementInst&) {
    685   return unknown();
    686 }
    687 
    688 SizeOffsetEvalType
    689 ObjectSizeOffsetEvaluator::visitExtractValueInst(ExtractValueInst&) {
    690   return unknown();
    691 }
    692 
    693 SizeOffsetEvalType
    694 ObjectSizeOffsetEvaluator::visitGEPOperator(GEPOperator &GEP) {
    695   SizeOffsetEvalType PtrData = compute_(GEP.getPointerOperand());
    696   if (!bothKnown(PtrData))
    697     return unknown();
    698 
    699   Value *Offset = EmitGEPOffset(&Builder, *TD, &GEP, /*NoAssumptions=*/true);
    700   Offset = Builder.CreateAdd(PtrData.second, Offset);
    701   return std::make_pair(PtrData.first, Offset);
    702 }
    703 
    704 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitIntToPtrInst(IntToPtrInst&) {
    705   // clueless
    706   return unknown();
    707 }
    708 
    709 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitLoadInst(LoadInst&) {
    710   return unknown();
    711 }
    712 
    713 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitPHINode(PHINode &PHI) {
    714   // create 2 PHIs: one for size and another for offset
    715   PHINode *SizePHI   = Builder.CreatePHI(IntTy, PHI.getNumIncomingValues());
    716   PHINode *OffsetPHI = Builder.CreatePHI(IntTy, PHI.getNumIncomingValues());
    717 
    718   // insert right away in the cache to handle recursive PHIs
    719   CacheMap[&PHI] = std::make_pair(SizePHI, OffsetPHI);
    720 
    721   // compute offset/size for each PHI incoming pointer
    722   for (unsigned i = 0, e = PHI.getNumIncomingValues(); i != e; ++i) {
    723     Builder.SetInsertPoint(PHI.getIncomingBlock(i)->getFirstInsertionPt());
    724     SizeOffsetEvalType EdgeData = compute_(PHI.getIncomingValue(i));
    725 
    726     if (!bothKnown(EdgeData)) {
    727       OffsetPHI->replaceAllUsesWith(UndefValue::get(IntTy));
    728       OffsetPHI->eraseFromParent();
    729       SizePHI->replaceAllUsesWith(UndefValue::get(IntTy));
    730       SizePHI->eraseFromParent();
    731       return unknown();
    732     }
    733     SizePHI->addIncoming(EdgeData.first, PHI.getIncomingBlock(i));
    734     OffsetPHI->addIncoming(EdgeData.second, PHI.getIncomingBlock(i));
    735   }
    736 
    737   Value *Size = SizePHI, *Offset = OffsetPHI, *Tmp;
    738   if ((Tmp = SizePHI->hasConstantValue())) {
    739     Size = Tmp;
    740     SizePHI->replaceAllUsesWith(Size);
    741     SizePHI->eraseFromParent();
    742   }
    743   if ((Tmp = OffsetPHI->hasConstantValue())) {
    744     Offset = Tmp;
    745     OffsetPHI->replaceAllUsesWith(Offset);
    746     OffsetPHI->eraseFromParent();
    747   }
    748   return std::make_pair(Size, Offset);
    749 }
    750 
    751 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitSelectInst(SelectInst &I) {
    752   SizeOffsetEvalType TrueSide  = compute_(I.getTrueValue());
    753   SizeOffsetEvalType FalseSide = compute_(I.getFalseValue());
    754 
    755   if (!bothKnown(TrueSide) || !bothKnown(FalseSide))
    756     return unknown();
    757   if (TrueSide == FalseSide)
    758     return TrueSide;
    759 
    760   Value *Size = Builder.CreateSelect(I.getCondition(), TrueSide.first,
    761                                      FalseSide.first);
    762   Value *Offset = Builder.CreateSelect(I.getCondition(), TrueSide.second,
    763                                        FalseSide.second);
    764   return std::make_pair(Size, Offset);
    765 }
    766 
    767 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitInstruction(Instruction &I) {
    768   DEBUG(dbgs() << "ObjectSizeOffsetEvaluator unknown instruction:" << I <<'\n');
    769   return unknown();
    770 }
    771