lib/Analysis/MemoryBuiltins.cpp

   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 #include "llvm/Analysis/MemoryBuiltins.h"
  16 #include "llvm/ADT/APInt.h"
  17 #include "llvm/ADT/None.h"
  18 #include "llvm/ADT/Optional.h"
  19 #include "llvm/ADT/STLExtras.h"
  20 #include "llvm/ADT/Statistic.h"
  21 #include "llvm/ADT/StringRef.h"
  22 #include "llvm/Analysis/TargetFolder.h"
  23 #include "llvm/Analysis/TargetLibraryInfo.h"
  24 #include "llvm/Analysis/Utils/Local.h"
  25 #include "llvm/Analysis/ValueTracking.h"
  26 #include "llvm/IR/Argument.h"
  27 #include "llvm/IR/Attributes.h"
  28 #include "llvm/IR/Constants.h"
  29 #include "llvm/IR/DataLayout.h"
  30 #include "llvm/IR/DerivedTypes.h"
  31 #include "llvm/IR/Function.h"
  32 #include "llvm/IR/GlobalAlias.h"
  33 #include "llvm/IR/GlobalVariable.h"
  34 #include "llvm/IR/Instruction.h"
  35 #include "llvm/IR/Instructions.h"
  36 #include "llvm/IR/IntrinsicInst.h"
  37 #include "llvm/IR/Operator.h"
  38 #include "llvm/IR/Type.h"
  39 #include "llvm/IR/Value.h"
  40 #include "llvm/Support/Casting.h"
  41 #include "llvm/Support/Debug.h"
  42 #include "llvm/Support/MathExtras.h"
  43 #include "llvm/Support/raw_ostream.h"
  44 #include <cassert>
  45 #include <cstdint>
  46 #include <iterator>
  47 #include <utility>
  48
  49 using namespace llvm;
  50
  51 #define DEBUG_TYPE "memory-builtins"
  52
  53 enum AllocType : uint8_t {
  54   OpNewLike          = 1<<0, // allocates; never returns null
  55   MallocLike         = 1<<1 | OpNewLike, // allocates; may return null
  56   CallocLike         = 1<<2, // allocates + bzero
  57   ReallocLike        = 1<<3, // reallocates
  58   StrDupLike         = 1<<4,
  59   MallocOrCallocLike = MallocLike | CallocLike,
  60   AllocLike          = MallocLike | CallocLike | StrDupLike,
  61   AnyAlloc           = AllocLike | ReallocLike
  62 };
  63
  64 struct AllocFnsTy {
  65   AllocType AllocTy;
  66   unsigned NumParams;
  67   // First and Second size parameters (or -1 if unused)
  68   int FstParam, SndParam;
  69 };
  70
  71 // FIXME: certain users need more information. E.g., SimplifyLibCalls needs to
  72 // know which functions are nounwind, noalias, nocapture parameters, etc.
  73 static const std::pair<LibFunc, AllocFnsTy> AllocationFnData[] = {
  74   {LibFunc_malloc,              {MallocLike,  1, 0,  -1}},
  75   {LibFunc_valloc,              {MallocLike,  1, 0,  -1}},
  76   {LibFunc_Znwj,                {OpNewLike,   1, 0,  -1}}, // new(unsigned int)
  77   {LibFunc_ZnwjRKSt9nothrow_t,  {MallocLike,  2, 0,  -1}}, // new(unsigned int, nothrow)
  78   {LibFunc_ZnwjSt11align_val_t, {OpNewLike,   2, 0,  -1}}, // new(unsigned int, align_val_t)
  79   {LibFunc_ZnwjSt11align_val_tRKSt9nothrow_t, // new(unsigned int, align_val_t, nothrow)
  80                                 {MallocLike,  3, 0,  -1}},
  81   {LibFunc_Znwm,                {OpNewLike,   1, 0,  -1}}, // new(unsigned long)
  82   {LibFunc_ZnwmRKSt9nothrow_t,  {MallocLike,  2, 0,  -1}}, // new(unsigned long, nothrow)
  83   {LibFunc_ZnwmSt11align_val_t, {OpNewLike,   2, 0,  -1}}, // new(unsigned long, align_val_t)
  84   {LibFunc_ZnwmSt11align_val_tRKSt9nothrow_t, // new(unsigned long, align_val_t, nothrow)
  85                                 {MallocLike,  3, 0,  -1}},
  86   {LibFunc_Znaj,                {OpNewLike,   1, 0,  -1}}, // new[](unsigned int)
  87   {LibFunc_ZnajRKSt9nothrow_t,  {MallocLike,  2, 0,  -1}}, // new[](unsigned int, nothrow)
  88   {LibFunc_ZnajSt11align_val_t, {OpNewLike,   2, 0,  -1}}, // new[](unsigned int, align_val_t)
  89   {LibFunc_ZnajSt11align_val_tRKSt9nothrow_t, // new[](unsigned int, align_val_t, nothrow)
  90                                 {MallocLike,  3, 0,  -1}},
  91   {LibFunc_Znam,                {OpNewLike,   1, 0,  -1}}, // new[](unsigned long)
  92   {LibFunc_ZnamRKSt9nothrow_t,  {MallocLike,  2, 0,  -1}}, // new[](unsigned long, nothrow)
  93   {LibFunc_ZnamSt11align_val_t, {OpNewLike,   2, 0,  -1}}, // new[](unsigned long, align_val_t)
  94   {LibFunc_ZnamSt11align_val_tRKSt9nothrow_t, // new[](unsigned long, align_val_t, nothrow)
  95                                  {MallocLike,  3, 0,  -1}},
  96   {LibFunc_msvc_new_int,         {OpNewLike,   1, 0,  -1}}, // new(unsigned int)
  97   {LibFunc_msvc_new_int_nothrow, {MallocLike,  2, 0,  -1}}, // new(unsigned int, nothrow)
  98   {LibFunc_msvc_new_longlong,         {OpNewLike,   1, 0,  -1}}, // new(unsigned long long)
  99   {LibFunc_msvc_new_longlong_nothrow, {MallocLike,  2, 0,  -1}}, // new(unsigned long long, nothrow)
 100   {LibFunc_msvc_new_array_int,         {OpNewLike,   1, 0,  -1}}, // new[](unsigned int)
 101   {LibFunc_msvc_new_array_int_nothrow, {MallocLike,  2, 0,  -1}}, // new[](unsigned int, nothrow)
 102   {LibFunc_msvc_new_array_longlong,         {OpNewLike,   1, 0,  -1}}, // new[](unsigned long long)
 103   {LibFunc_msvc_new_array_longlong_nothrow, {MallocLike,  2, 0,  -1}}, // new[](unsigned long long, nothrow)
 104   {LibFunc_calloc,              {CallocLike,  2, 0,   1}},
 105   {LibFunc_realloc,             {ReallocLike, 2, 1,  -1}},
 106   {LibFunc_reallocf,            {ReallocLike, 2, 1,  -1}},
 107   {LibFunc_strdup,              {StrDupLike,  1, -1, -1}},
 108   {LibFunc_strndup,             {StrDupLike,  2, 1,  -1}}
 109   // TODO: Handle "int posix_memalign(void **, size_t, size_t)"
 110 };
 111
 112 static const Function *getCalledFunction(const Value *V, bool LookThroughBitCast,
 113                                          bool &IsNoBuiltin) {
 114   // Don't care about intrinsics in this case.
 115   if (isa<IntrinsicInst>(V))
 116     return nullptr;
 117
 118   if (LookThroughBitCast)
 119     V = V->stripPointerCasts();
 120
 121   ImmutableCallSite CS(V);
 122   if (!CS.getInstruction())
 123     return nullptr;
 124
 125   IsNoBuiltin = CS.isNoBuiltin();
 126
 127   if (const Function *Callee = CS.getCalledFunction())
 128     return Callee;
 129   return nullptr;
 130 }
 131
 132 /// Returns the allocation data for the given value if it's either a call to a
 133 /// known allocation function, or a call to a function with the allocsize
 134 /// attribute.
 135 static Optional<AllocFnsTy>
 136 getAllocationDataForFunction(const Function *Callee, AllocType AllocTy,
 137                              const TargetLibraryInfo *TLI) {
 138   // Make sure that the function is available.
 139   StringRef FnName = Callee->getName();
 140   LibFunc TLIFn;
 141   if (!TLI || !TLI->getLibFunc(FnName, TLIFn) || !TLI->has(TLIFn))
 142     return None;
 143
 144   const auto *Iter = find_if(
 145       AllocationFnData, [TLIFn](const std::pair<LibFunc, AllocFnsTy> &P) {
 146         return P.first == TLIFn;
 147       });
 148
 149   if (Iter == std::end(AllocationFnData))
 150     return None;
 151
 152   const AllocFnsTy *FnData = &Iter->second;
 153   if ((FnData->AllocTy & AllocTy) != FnData->AllocTy)
 154     return None;
 155
 156   // Check function prototype.
 157   int FstParam = FnData->FstParam;
 158   int SndParam = FnData->SndParam;
 159   FunctionType *FTy = Callee->getFunctionType();
 160
 161   if (FTy->getReturnType() == Type::getInt8PtrTy(FTy->getContext()) &&
 162       FTy->getNumParams() == FnData->NumParams &&
 163       (FstParam < 0 ||
 164        (FTy->getParamType(FstParam)->isIntegerTy(32) ||
 165         FTy->getParamType(FstParam)->isIntegerTy(64))) &&
 166       (SndParam < 0 ||
 167        FTy->getParamType(SndParam)->isIntegerTy(32) ||
 168        FTy->getParamType(SndParam)->isIntegerTy(64)))
 169     return *FnData;
 170   return None;
 171 }
 172
 173 static Optional<AllocFnsTy> getAllocationData(const Value *V, AllocType AllocTy,
 174                                               const TargetLibraryInfo *TLI,
 175                                               bool LookThroughBitCast = false) {
 176   bool IsNoBuiltinCall;
 177   if (const Function *Callee =
 178           getCalledFunction(V, LookThroughBitCast, IsNoBuiltinCall))
 179     if (!IsNoBuiltinCall)
 180       return getAllocationDataForFunction(Callee, AllocTy, TLI);
 181   return None;
 182 }
 183
 184 static Optional<AllocFnsTy> getAllocationSize(const Value *V,
 185                                               const TargetLibraryInfo *TLI) {
 186   bool IsNoBuiltinCall;
 187   const Function *Callee =
 188       getCalledFunction(V, /*LookThroughBitCast=*/false, IsNoBuiltinCall);
 189   if (!Callee)
 190     return None;
 191
 192   // Prefer to use existing information over allocsize. This will give us an
 193   // accurate AllocTy.
 194   if (!IsNoBuiltinCall)
 195     if (Optional<AllocFnsTy> Data =
 196             getAllocationDataForFunction(Callee, AnyAlloc, TLI))
 197       return Data;
 198
 199   Attribute Attr = Callee->getFnAttribute(Attribute::AllocSize);
 200   if (Attr == Attribute())
 201     return None;
 202
 203   std::pair<unsigned, Optional<unsigned>> Args = Attr.getAllocSizeArgs();
 204
 205   AllocFnsTy Result;
 206   // Because allocsize only tells us how many bytes are allocated, we're not
 207   // really allowed to assume anything, so we use MallocLike.
 208   Result.AllocTy = MallocLike;
 209   Result.NumParams = Callee->getNumOperands();
 210   Result.FstParam = Args.first;
 211   Result.SndParam = Args.second.getValueOr(-1);
 212   return Result;
 213 }
 214
 215 static bool hasNoAliasAttr(const Value *V, bool LookThroughBitCast) {
 216   ImmutableCallSite CS(LookThroughBitCast ? V->stripPointerCasts() : V);
 217   return CS && CS.hasRetAttr(Attribute::NoAlias);
 218 }
 219
 220 /// Tests if a value is a call or invoke to a library function that
 221 /// allocates or reallocates memory (either malloc, calloc, realloc, or strdup
 222 /// like).
 223 bool llvm::isAllocationFn(const Value *V, const TargetLibraryInfo *TLI,
 224                           bool LookThroughBitCast) {
 225   return getAllocationData(V, AnyAlloc, TLI, LookThroughBitCast).hasValue();
 226 }
 227
 228 /// Tests if a value is a call or invoke to a function that returns a
 229 /// NoAlias pointer (including malloc/calloc/realloc/strdup-like functions).
 230 bool llvm::isNoAliasFn(const Value *V, const TargetLibraryInfo *TLI,
 231                        bool LookThroughBitCast) {
 232   // it's safe to consider realloc as noalias since accessing the original
 233   // pointer is undefined behavior
 234   return isAllocationFn(V, TLI, LookThroughBitCast) ||
 235          hasNoAliasAttr(V, LookThroughBitCast);
 236 }
 237
 238 /// Tests if a value is a call or invoke to a library function that
 239 /// allocates uninitialized memory (such as malloc).
 240 bool llvm::isMallocLikeFn(const Value *V, const TargetLibraryInfo *TLI,
 241                           bool LookThroughBitCast) {
 242   return getAllocationData(V, MallocLike, TLI, LookThroughBitCast).hasValue();
 243 }
 244
 245 /// Tests if a value is a call or invoke to a library function that
 246 /// allocates zero-filled memory (such as calloc).
 247 bool llvm::isCallocLikeFn(const Value *V, const TargetLibraryInfo *TLI,
 248                           bool LookThroughBitCast) {
 249   return getAllocationData(V, CallocLike, TLI, LookThroughBitCast).hasValue();
 250 }
 251
 252 /// Tests if a value is a call or invoke to a library function that
 253 /// allocates memory similar to malloc or calloc.
 254 bool llvm::isMallocOrCallocLikeFn(const Value *V, const TargetLibraryInfo *TLI,
 255                                   bool LookThroughBitCast) {
 256   return getAllocationData(V, MallocOrCallocLike, TLI,
 257                            LookThroughBitCast).hasValue();
 258 }
 259
 260 /// Tests if a value is a call or invoke to a library function that
 261 /// allocates memory (either malloc, calloc, or strdup like).
 262 bool llvm::isAllocLikeFn(const Value *V, const TargetLibraryInfo *TLI,
 263                          bool LookThroughBitCast) {
 264   return getAllocationData(V, AllocLike, TLI, LookThroughBitCast).hasValue();
 265 }
 266
 267 /// extractMallocCall - Returns the corresponding CallInst if the instruction
 268 /// is a malloc call.  Since CallInst::CreateMalloc() only creates calls, we
 269 /// ignore InvokeInst here.
 270 const CallInst *llvm::extractMallocCall(const Value *I,
 271                                         const TargetLibraryInfo *TLI) {
 272   return isMallocLikeFn(I, TLI) ? dyn_cast<CallInst>(I) : nullptr;
 273 }
 274
 275 static Value *computeArraySize(const CallInst *CI, const DataLayout &DL,
 276                                const TargetLibraryInfo *TLI,
 277                                bool LookThroughSExt = false) {
 278   if (!CI)
 279     return nullptr;
 280
 281   // The size of the malloc's result type must be known to determine array size.
 282   Type *T = getMallocAllocatedType(CI, TLI);
 283   if (!T || !T->isSized())
 284     return nullptr;
 285
 286   unsigned ElementSize = DL.getTypeAllocSize(T);
 287   if (StructType *ST = dyn_cast<StructType>(T))
 288     ElementSize = DL.getStructLayout(ST)->getSizeInBytes();
 289
 290   // If malloc call's arg can be determined to be a multiple of ElementSize,
 291   // return the multiple.  Otherwise, return NULL.
 292   Value *MallocArg = CI->getArgOperand(0);
 293   Value *Multiple = nullptr;
 294   if (ComputeMultiple(MallocArg, ElementSize, Multiple, LookThroughSExt))
 295     return Multiple;
 296
 297   return nullptr;
 298 }
 299
 300 /// getMallocType - Returns the PointerType resulting from the malloc call.
 301 /// The PointerType depends on the number of bitcast uses of the malloc call:
 302 ///   0: PointerType is the calls' return type.
 303 ///   1: PointerType is the bitcast's result type.
 304 ///  >1: Unique PointerType cannot be determined, return NULL.
 305 PointerType *llvm::getMallocType(const CallInst *CI,
 306                                  const TargetLibraryInfo *TLI) {
 307   assert(isMallocLikeFn(CI, TLI) && "getMallocType and not malloc call");
 308
 309   PointerType *MallocType = nullptr;
 310   unsigned NumOfBitCastUses = 0;
 311
 312   // Determine if CallInst has a bitcast use.
 313   for (Value::const_user_iterator UI = CI->user_begin(), E = CI->user_end();
 314        UI != E;)
 315     if (const BitCastInst *BCI = dyn_cast<BitCastInst>(*UI++)) {
 316       MallocType = cast<PointerType>(BCI->getDestTy());
 317       NumOfBitCastUses++;
 318     }
 319
 320   // Malloc call has 1 bitcast use, so type is the bitcast's destination type.
 321   if (NumOfBitCastUses == 1)
 322     return MallocType;
 323
 324   // Malloc call was not bitcast, so type is the malloc function's return type.
 325   if (NumOfBitCastUses == 0)
 326     return cast<PointerType>(CI->getType());
 327
 328   // Type could not be determined.
 329   return nullptr;
 330 }
 331
 332 /// getMallocAllocatedType - Returns the Type allocated by malloc call.
 333 /// The Type depends on the number of bitcast uses of the malloc call:
 334 ///   0: PointerType is the malloc calls' return type.
 335 ///   1: PointerType is the bitcast's result type.
 336 ///  >1: Unique PointerType cannot be determined, return NULL.
 337 Type *llvm::getMallocAllocatedType(const CallInst *CI,
 338                                    const TargetLibraryInfo *TLI) {
 339   PointerType *PT = getMallocType(CI, TLI);
 340   return PT ? PT->getElementType() : nullptr;
 341 }
 342
 343 /// getMallocArraySize - Returns the array size of a malloc call.  If the
 344 /// argument passed to malloc is a multiple of the size of the malloced type,
 345 /// then return that multiple.  For non-array mallocs, the multiple is
 346 /// constant 1.  Otherwise, return NULL for mallocs whose array size cannot be
 347 /// determined.
 348 Value *llvm::getMallocArraySize(CallInst *CI, const DataLayout &DL,
 349                                 const TargetLibraryInfo *TLI,
 350                                 bool LookThroughSExt) {
 351   assert(isMallocLikeFn(CI, TLI) && "getMallocArraySize and not malloc call");
 352   return computeArraySize(CI, DL, TLI, LookThroughSExt);
 353 }
 354
 355 /// extractCallocCall - Returns the corresponding CallInst if the instruction
 356 /// is a calloc call.
 357 const CallInst *llvm::extractCallocCall(const Value *I,
 358                                         const TargetLibraryInfo *TLI) {
 359   return isCallocLikeFn(I, TLI) ? cast<CallInst>(I) : nullptr;
 360 }
 361
 362 /// isFreeCall - Returns non-null if the value is a call to the builtin free()
 363 const CallInst *llvm::isFreeCall(const Value *I, const TargetLibraryInfo *TLI) {
 364   bool IsNoBuiltinCall;
 365   const Function *Callee =
 366       getCalledFunction(I, /*LookThroughBitCast=*/false, IsNoBuiltinCall);
 367   if (Callee == nullptr || IsNoBuiltinCall)
 368     return nullptr;
 369
 370   StringRef FnName = Callee->getName();
 371   LibFunc TLIFn;
 372   if (!TLI || !TLI->getLibFunc(FnName, TLIFn) || !TLI->has(TLIFn))
 373     return nullptr;
 374
 375   unsigned ExpectedNumParams;
 376   if (TLIFn == LibFunc_free ||
 377       TLIFn == LibFunc_ZdlPv || // operator delete(void*)
 378       TLIFn == LibFunc_ZdaPv || // operator delete[](void*)
 379       TLIFn == LibFunc_msvc_delete_ptr32 || // operator delete(void*)
 380       TLIFn == LibFunc_msvc_delete_ptr64 || // operator delete(void*)
 381       TLIFn == LibFunc_msvc_delete_array_ptr32 || // operator delete[](void*)
 382       TLIFn == LibFunc_msvc_delete_array_ptr64)   // operator delete[](void*)
 383     ExpectedNumParams = 1;
 384   else if (TLIFn == LibFunc_ZdlPvj ||              // delete(void*, uint)
 385            TLIFn == LibFunc_ZdlPvm ||              // delete(void*, ulong)
 386            TLIFn == LibFunc_ZdlPvRKSt9nothrow_t || // delete(void*, nothrow)
 387            TLIFn == LibFunc_ZdlPvSt11align_val_t || // delete(void*, align_val_t)
 388            TLIFn == LibFunc_ZdaPvj ||              // delete[](void*, uint)
 389            TLIFn == LibFunc_ZdaPvm ||              // delete[](void*, ulong)
 390            TLIFn == LibFunc_ZdaPvRKSt9nothrow_t || // delete[](void*, nothrow)
 391            TLIFn == LibFunc_ZdaPvSt11align_val_t || // delete[](void*, align_val_t)
 392            TLIFn == LibFunc_msvc_delete_ptr32_int ||      // delete(void*, uint)
 393            TLIFn == LibFunc_msvc_delete_ptr64_longlong || // delete(void*, ulonglong)
 394            TLIFn == LibFunc_msvc_delete_ptr32_nothrow || // delete(void*, nothrow)
 395            TLIFn == LibFunc_msvc_delete_ptr64_nothrow || // delete(void*, nothrow)
 396            TLIFn == LibFunc_msvc_delete_array_ptr32_int ||      // delete[](void*, uint)
 397            TLIFn == LibFunc_msvc_delete_array_ptr64_longlong || // delete[](void*, ulonglong)
 398            TLIFn == LibFunc_msvc_delete_array_ptr32_nothrow || // delete[](void*, nothrow)
 399            TLIFn == LibFunc_msvc_delete_array_ptr64_nothrow)   // delete[](void*, nothrow)
 400     ExpectedNumParams = 2;
 401   else if (TLIFn == LibFunc_ZdaPvSt11align_val_tRKSt9nothrow_t || // delete(void*, align_val_t, nothrow)
 402            TLIFn == LibFunc_ZdlPvSt11align_val_tRKSt9nothrow_t) // delete[](void*, align_val_t, nothrow)
 403     ExpectedNumParams = 3;
 404   else
 405     return nullptr;
 406
 407   // Check free prototype.
 408   // FIXME: workaround for PR5130, this will be obsolete when a nobuiltin
 409   // attribute will exist.
 410   FunctionType *FTy = Callee->getFunctionType();
 411   if (!FTy->getReturnType()->isVoidTy())
 412     return nullptr;
 413   if (FTy->getNumParams() != ExpectedNumParams)
 414     return nullptr;
 415   if (FTy->getParamType(0) != Type::getInt8PtrTy(Callee->getContext()))
 416     return nullptr;
 417
 418   return dyn_cast<CallInst>(I);
 419 }
 420
 421 //===----------------------------------------------------------------------===//
 422 //  Utility functions to compute size of objects.
 423 //
 424 static APInt getSizeWithOverflow(const SizeOffsetType &Data) {
 425   if (Data.second.isNegative() || Data.first.ult(Data.second))
 426     return APInt(Data.first.getBitWidth(), 0);
 427   return Data.first - Data.second;
 428 }
 429
 430 /// Compute the size of the object pointed by Ptr. Returns true and the
 431 /// object size in Size if successful, and false otherwise.
 432 /// If RoundToAlign is true, then Size is rounded up to the alignment of
 433 /// allocas, byval arguments, and global variables.
 434 bool llvm::getObjectSize(const Value *Ptr, uint64_t &Size, const DataLayout &DL,
 435                          const TargetLibraryInfo *TLI, ObjectSizeOpts Opts) {
 436   ObjectSizeOffsetVisitor Visitor(DL, TLI, Ptr->getContext(), Opts);
 437   SizeOffsetType Data = Visitor.compute(const_cast<Value*>(Ptr));
 438   if (!Visitor.bothKnown(Data))
 439     return false;
 440
 441   Size = getSizeWithOverflow(Data).getZExtValue();
 442   return true;
 443 }
 444
 445 ConstantInt *llvm::lowerObjectSizeCall(IntrinsicInst *ObjectSize,
 446                                        const DataLayout &DL,
 447                                        const TargetLibraryInfo *TLI,
 448                                        bool MustSucceed) {
 449   assert(ObjectSize->getIntrinsicID() == Intrinsic::objectsize &&
 450          "ObjectSize must be a call to llvm.objectsize!");
 451
 452   bool MaxVal = cast<ConstantInt>(ObjectSize->getArgOperand(1))->isZero();
 453   ObjectSizeOpts EvalOptions;
 454   // Unless we have to fold this to something, try to be as accurate as
 455   // possible.
 456   if (MustSucceed)
 457     EvalOptions.EvalMode =
 458         MaxVal ? ObjectSizeOpts::Mode::Max : ObjectSizeOpts::Mode::Min;
 459   else
 460     EvalOptions.EvalMode = ObjectSizeOpts::Mode::Exact;
 461
 462   EvalOptions.NullIsUnknownSize =
 463       cast<ConstantInt>(ObjectSize->getArgOperand(2))->isOne();
 464
 465   // FIXME: Does it make sense to just return a failure value if the size won't
 466   // fit in the output and `!MustSucceed`?
 467   uint64_t Size;
 468   auto *ResultType = cast<IntegerType>(ObjectSize->getType());
 469   if (getObjectSize(ObjectSize->getArgOperand(0), Size, DL, TLI, EvalOptions) &&
 470       isUIntN(ResultType->getBitWidth(), Size))
 471     return ConstantInt::get(ResultType, Size);
 472
 473   if (!MustSucceed)
 474     return nullptr;
 475
 476   return ConstantInt::get(ResultType, MaxVal ? -1ULL : 0);
 477 }
 478
 479 STATISTIC(ObjectVisitorArgument,
 480           "Number of arguments with unsolved size and offset");
 481 STATISTIC(ObjectVisitorLoad,
 482           "Number of load instructions with unsolved size and offset");
 483
 484 APInt ObjectSizeOffsetVisitor::align(APInt Size, uint64_t Align) {
 485   if (Options.RoundToAlign && Align)
 486     return APInt(IntTyBits, alignTo(Size.getZExtValue(), Align));
 487   return Size;
 488 }
 489
 490 ObjectSizeOffsetVisitor::ObjectSizeOffsetVisitor(const DataLayout &DL,
 491                                                  const TargetLibraryInfo *TLI,
 492                                                  LLVMContext &Context,
 493                                                  ObjectSizeOpts Options)
 494     : DL(DL), TLI(TLI), Options(Options) {
 495   // Pointer size must be rechecked for each object visited since it could have
 496   // a different address space.
 497 }
 498
 499 SizeOffsetType ObjectSizeOffsetVisitor::compute(Value *V) {
 500   IntTyBits = DL.getPointerTypeSizeInBits(V->getType());
 501   Zero = APInt::getNullValue(IntTyBits);
 502
 503   V = V->stripPointerCasts();
 504   if (Instruction *I = dyn_cast<Instruction>(V)) {
 505     // If we have already seen this instruction, bail out. Cycles can happen in
 506     // unreachable code after constant propagation.
 507     if (!SeenInsts.insert(I).second)
 508       return unknown();
 509
 510     if (GEPOperator *GEP = dyn_cast<GEPOperator>(V))
 511       return visitGEPOperator(*GEP);
 512     return visit(*I);
 513   }
 514   if (Argument *A = dyn_cast<Argument>(V))
 515     return visitArgument(*A);
 516   if (ConstantPointerNull *P = dyn_cast<ConstantPointerNull>(V))
 517     return visitConstantPointerNull(*P);
 518   if (GlobalAlias *GA = dyn_cast<GlobalAlias>(V))
 519     return visitGlobalAlias(*GA);
 520   if (GlobalVariable *GV = dyn_cast<GlobalVariable>(V))
 521     return visitGlobalVariable(*GV);
 522   if (UndefValue *UV = dyn_cast<UndefValue>(V))
 523     return visitUndefValue(*UV);
 524   if (ConstantExpr *CE = dyn_cast<ConstantExpr>(V)) {
 525     if (CE->getOpcode() == Instruction::IntToPtr)
 526       return unknown(); // clueless
 527     if (CE->getOpcode() == Instruction::GetElementPtr)
 528       return visitGEPOperator(cast<GEPOperator>(*CE));
 529   }
 530
 531   LLVM_DEBUG(dbgs() << "ObjectSizeOffsetVisitor::compute() unhandled value: "
 532                     << *V << '\n');
 533   return unknown();
 534 }
 535
 536 /// When we're compiling N-bit code, and the user uses parameters that are
 537 /// greater than N bits (e.g. uint64_t on a 32-bit build), we can run into
 538 /// trouble with APInt size issues. This function handles resizing + overflow
 539 /// checks for us. Check and zext or trunc \p I depending on IntTyBits and
 540 /// I's value.
 541 bool ObjectSizeOffsetVisitor::CheckedZextOrTrunc(APInt &I) {
 542   // More bits than we can handle. Checking the bit width isn't necessary, but
 543   // it's faster than checking active bits, and should give `false` in the
 544   // vast majority of cases.
 545   if (I.getBitWidth() > IntTyBits && I.getActiveBits() > IntTyBits)
 546     return false;
 547   if (I.getBitWidth() != IntTyBits)
 548     I = I.zextOrTrunc(IntTyBits);
 549   return true;
 550 }
 551
 552 SizeOffsetType ObjectSizeOffsetVisitor::visitAllocaInst(AllocaInst &I) {
 553   if (!I.getAllocatedType()->isSized())
 554     return unknown();
 555
 556   APInt Size(IntTyBits, DL.getTypeAllocSize(I.getAllocatedType()));
 557   if (!I.isArrayAllocation())
 558     return std::make_pair(align(Size, I.getAlignment()), Zero);
 559
 560   Value *ArraySize = I.getArraySize();
 561   if (const ConstantInt *C = dyn_cast<ConstantInt>(ArraySize)) {
 562     APInt NumElems = C->getValue();
 563     if (!CheckedZextOrTrunc(NumElems))
 564       return unknown();
 565
 566     bool Overflow;
 567     Size = Size.umul_ov(NumElems, Overflow);
 568     return Overflow ? unknown() : std::make_pair(align(Size, I.getAlignment()),
 569                                                  Zero);
 570   }
 571   return unknown();
 572 }
 573
 574 SizeOffsetType ObjectSizeOffsetVisitor::visitArgument(Argument &A) {
 575   // No interprocedural analysis is done at the moment.
 576   if (!A.hasByValOrInAllocaAttr()) {
 577     ++ObjectVisitorArgument;
 578     return unknown();
 579   }
 580   PointerType *PT = cast<PointerType>(A.getType());
 581   APInt Size(IntTyBits, DL.getTypeAllocSize(PT->getElementType()));
 582   return std::make_pair(align(Size, A.getParamAlignment()), Zero);
 583 }
 584
 585 SizeOffsetType ObjectSizeOffsetVisitor::visitCallSite(CallSite CS) {
 586   Optional<AllocFnsTy> FnData = getAllocationSize(CS.getInstruction(), TLI);
 587   if (!FnData)
 588     return unknown();
 589
 590   // Handle strdup-like functions separately.
 591   if (FnData->AllocTy == StrDupLike) {
 592     APInt Size(IntTyBits, GetStringLength(CS.getArgument(0)));
 593     if (!Size)
 594       return unknown();
 595
 596     // Strndup limits strlen.
 597     if (FnData->FstParam > 0) {
 598       ConstantInt *Arg =
 599           dyn_cast<ConstantInt>(CS.getArgument(FnData->FstParam));
 600       if (!Arg)
 601         return unknown();
 602
 603       APInt MaxSize = Arg->getValue().zextOrSelf(IntTyBits);
 604       if (Size.ugt(MaxSize))
 605         Size = MaxSize + 1;
 606     }
 607     return std::make_pair(Size, Zero);
 608   }
 609
 610   ConstantInt *Arg = dyn_cast<ConstantInt>(CS.getArgument(FnData->FstParam));
 611   if (!Arg)
 612     return unknown();
 613
 614   APInt Size = Arg->getValue();
 615   if (!CheckedZextOrTrunc(Size))
 616     return unknown();
 617
 618   // Size is determined by just 1 parameter.
 619   if (FnData->SndParam < 0)
 620     return std::make_pair(Size, Zero);
 621
 622   Arg = dyn_cast<ConstantInt>(CS.getArgument(FnData->SndParam));
 623   if (!Arg)
 624     return unknown();
 625
 626   APInt NumElems = Arg->getValue();
 627   if (!CheckedZextOrTrunc(NumElems))
 628     return unknown();
 629
 630   bool Overflow;
 631   Size = Size.umul_ov(NumElems, Overflow);
 632   return Overflow ? unknown() : std::make_pair(Size, Zero);
 633
 634   // TODO: handle more standard functions (+ wchar cousins):
 635   // - strdup / strndup
 636   // - strcpy / strncpy
 637   // - strcat / strncat
 638   // - memcpy / memmove
 639   // - strcat / strncat
 640   // - memset
 641 }
 642
 643 SizeOffsetType
 644 ObjectSizeOffsetVisitor::visitConstantPointerNull(ConstantPointerNull& CPN) {
 645   // If null is unknown, there's nothing we can do. Additionally, non-zero
 646   // address spaces can make use of null, so we don't presume to know anything
 647   // about that.
 648   //
 649   // TODO: How should this work with address space casts? We currently just drop
 650   // them on the floor, but it's unclear what we should do when a NULL from
 651   // addrspace(1) gets casted to addrspace(0) (or vice-versa).
 652   if (Options.NullIsUnknownSize || CPN.getType()->getAddressSpace())
 653     return unknown();
 654   return std::make_pair(Zero, Zero);
 655 }
 656
 657 SizeOffsetType
 658 ObjectSizeOffsetVisitor::visitExtractElementInst(ExtractElementInst&) {
 659   return unknown();
 660 }
 661
 662 SizeOffsetType
 663 ObjectSizeOffsetVisitor::visitExtractValueInst(ExtractValueInst&) {
 664   // Easy cases were already folded by previous passes.
 665   return unknown();
 666 }
 667
 668 SizeOffsetType ObjectSizeOffsetVisitor::visitGEPOperator(GEPOperator &GEP) {
 669   SizeOffsetType PtrData = compute(GEP.getPointerOperand());
 670   APInt Offset(IntTyBits, 0);
 671   if (!bothKnown(PtrData) || !GEP.accumulateConstantOffset(DL, Offset))
 672     return unknown();
 673
 674   return std::make_pair(PtrData.first, PtrData.second + Offset);
 675 }
 676
 677 SizeOffsetType ObjectSizeOffsetVisitor::visitGlobalAlias(GlobalAlias &GA) {
 678   if (GA.isInterposable())
 679     return unknown();
 680   return compute(GA.getAliasee());
 681 }
 682
 683 SizeOffsetType ObjectSizeOffsetVisitor::visitGlobalVariable(GlobalVariable &GV){
 684   if (!GV.hasDefinitiveInitializer())
 685     return unknown();
 686
 687   APInt Size(IntTyBits, DL.getTypeAllocSize(GV.getType()->getElementType()));
 688   return std::make_pair(align(Size, GV.getAlignment()), Zero);
 689 }
 690
 691 SizeOffsetType ObjectSizeOffsetVisitor::visitIntToPtrInst(IntToPtrInst&) {
 692   // clueless
 693   return unknown();
 694 }
 695
 696 SizeOffsetType ObjectSizeOffsetVisitor::visitLoadInst(LoadInst&) {
 697   ++ObjectVisitorLoad;
 698   return unknown();
 699 }
 700
 701 SizeOffsetType ObjectSizeOffsetVisitor::visitPHINode(PHINode&) {
 702   // too complex to analyze statically.
 703   return unknown();
 704 }
 705
 706 SizeOffsetType ObjectSizeOffsetVisitor::visitSelectInst(SelectInst &I) {
 707   SizeOffsetType TrueSide  = compute(I.getTrueValue());
 708   SizeOffsetType FalseSide = compute(I.getFalseValue());
 709   if (bothKnown(TrueSide) && bothKnown(FalseSide)) {
 710     if (TrueSide == FalseSide) {
 711         return TrueSide;
 712     }
 713
 714     APInt TrueResult = getSizeWithOverflow(TrueSide);
 715     APInt FalseResult = getSizeWithOverflow(FalseSide);
 716
 717     if (TrueResult == FalseResult) {
 718       return TrueSide;
 719     }
 720     if (Options.EvalMode == ObjectSizeOpts::Mode::Min) {
 721       if (TrueResult.slt(FalseResult))
 722         return TrueSide;
 723       return FalseSide;
 724     }
 725     if (Options.EvalMode == ObjectSizeOpts::Mode::Max) {
 726       if (TrueResult.sgt(FalseResult))
 727         return TrueSide;
 728       return FalseSide;
 729     }
 730   }
 731   return unknown();
 732 }
 733
 734 SizeOffsetType ObjectSizeOffsetVisitor::visitUndefValue(UndefValue&) {
 735   return std::make_pair(Zero, Zero);
 736 }
 737
 738 SizeOffsetType ObjectSizeOffsetVisitor::visitInstruction(Instruction &I) {
 739   LLVM_DEBUG(dbgs() << "ObjectSizeOffsetVisitor unknown instruction:" << I
 740                     << '\n');
 741   return unknown();
 742 }
 743
 744 ObjectSizeOffsetEvaluator::ObjectSizeOffsetEvaluator(
 745     const DataLayout &DL, const TargetLibraryInfo *TLI, LLVMContext &Context,
 746     bool RoundToAlign)
 747     : DL(DL), TLI(TLI), Context(Context), Builder(Context, TargetFolder(DL)),
 748       RoundToAlign(RoundToAlign) {
 749   // IntTy and Zero must be set for each compute() since the address space may
 750   // be different for later objects.
 751 }
 752
 753 SizeOffsetEvalType ObjectSizeOffsetEvaluator::compute(Value *V) {
 754   // XXX - Are vectors of pointers possible here?
 755   IntTy = cast<IntegerType>(DL.getIntPtrType(V->getType()));
 756   Zero = ConstantInt::get(IntTy, 0);
 757
 758   SizeOffsetEvalType Result = compute_(V);
 759
 760   if (!bothKnown(Result)) {
 761     // Erase everything that was computed in this iteration from the cache, so
 762     // that no dangling references are left behind. We could be a bit smarter if
 763     // we kept a dependency graph. It's probably not worth the complexity.
 764     for (const Value *SeenVal : SeenVals) {
 765       CacheMapTy::iterator CacheIt = CacheMap.find(SeenVal);
 766       // non-computable results can be safely cached
 767       if (CacheIt != CacheMap.end() && anyKnown(CacheIt->second))
 768         CacheMap.erase(CacheIt);
 769     }
 770   }
 771
 772   SeenVals.clear();
 773   return Result;
 774 }
 775
 776 SizeOffsetEvalType ObjectSizeOffsetEvaluator::compute_(Value *V) {
 777   ObjectSizeOpts ObjSizeOptions;
 778   ObjSizeOptions.RoundToAlign = RoundToAlign;
 779
 780   ObjectSizeOffsetVisitor Visitor(DL, TLI, Context, ObjSizeOptions);
 781   SizeOffsetType Const = Visitor.compute(V);
 782   if (Visitor.bothKnown(Const))
 783     return std::make_pair(ConstantInt::get(Context, Const.first),
 784                           ConstantInt::get(Context, Const.second));
 785
 786   V = V->stripPointerCasts();
 787
 788   // Check cache.
 789   CacheMapTy::iterator CacheIt = CacheMap.find(V);
 790   if (CacheIt != CacheMap.end())
 791     return CacheIt->second;
 792
 793   // Always generate code immediately before the instruction being
 794   // processed, so that the generated code dominates the same BBs.
 795   BuilderTy::InsertPointGuard Guard(Builder);
 796   if (Instruction *I = dyn_cast<Instruction>(V))
 797     Builder.SetInsertPoint(I);
 798
 799   // Now compute the size and offset.
 800   SizeOffsetEvalType Result;
 801
 802   // Record the pointers that were handled in this run, so that they can be
 803   // cleaned later if something fails. We also use this set to break cycles that
 804   // can occur in dead code.
 805   if (!SeenVals.insert(V).second) {
 806     Result = unknown();
 807   } else if (GEPOperator *GEP = dyn_cast<GEPOperator>(V)) {
 808     Result = visitGEPOperator(*GEP);
 809   } else if (Instruction *I = dyn_cast<Instruction>(V)) {
 810     Result = visit(*I);
 811   } else if (isa<Argument>(V) ||
 812              (isa<ConstantExpr>(V) &&
 813               cast<ConstantExpr>(V)->getOpcode() == Instruction::IntToPtr) ||
 814              isa<GlobalAlias>(V) ||
 815              isa<GlobalVariable>(V)) {
 816     // Ignore values where we cannot do more than ObjectSizeVisitor.
 817     Result = unknown();
 818   } else {
 819     LLVM_DEBUG(
 820         dbgs() << "ObjectSizeOffsetEvaluator::compute() unhandled value: " << *V
 821                << '\n');
 822     Result = unknown();
 823   }
 824
 825   // Don't reuse CacheIt since it may be invalid at this point.
 826   CacheMap[V] = Result;
 827   return Result;
 828 }
 829
 830 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitAllocaInst(AllocaInst &I) {
 831   if (!I.getAllocatedType()->isSized())
 832     return unknown();
 833
 834   // must be a VLA
 835   assert(I.isArrayAllocation());
 836   Value *ArraySize = I.getArraySize();
 837   Value *Size = ConstantInt::get(ArraySize->getType(),
 838                                  DL.getTypeAllocSize(I.getAllocatedType()));
 839   Size = Builder.CreateMul(Size, ArraySize);
 840   return std::make_pair(Size, Zero);
 841 }
 842
 843 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitCallSite(CallSite CS) {
 844   Optional<AllocFnsTy> FnData = getAllocationSize(CS.getInstruction(), TLI);
 845   if (!FnData)
 846     return unknown();
 847
 848   // Handle strdup-like functions separately.
 849   if (FnData->AllocTy == StrDupLike) {
 850     // TODO
 851     return unknown();
 852   }
 853
 854   Value *FirstArg = CS.getArgument(FnData->FstParam);
 855   FirstArg = Builder.CreateZExt(FirstArg, IntTy);
 856   if (FnData->SndParam < 0)
 857     return std::make_pair(FirstArg, Zero);
 858
 859   Value *SecondArg = CS.getArgument(FnData->SndParam);
 860   SecondArg = Builder.CreateZExt(SecondArg, IntTy);
 861   Value *Size = Builder.CreateMul(FirstArg, SecondArg);
 862   return std::make_pair(Size, Zero);
 863
 864   // TODO: handle more standard functions (+ wchar cousins):
 865   // - strdup / strndup
 866   // - strcpy / strncpy
 867   // - strcat / strncat
 868   // - memcpy / memmove
 869   // - strcat / strncat
 870   // - memset
 871 }
 872
 873 SizeOffsetEvalType
 874 ObjectSizeOffsetEvaluator::visitExtractElementInst(ExtractElementInst&) {
 875   return unknown();
 876 }
 877
 878 SizeOffsetEvalType
 879 ObjectSizeOffsetEvaluator::visitExtractValueInst(ExtractValueInst&) {
 880   return unknown();
 881 }
 882
 883 SizeOffsetEvalType
 884 ObjectSizeOffsetEvaluator::visitGEPOperator(GEPOperator &GEP) {
 885   SizeOffsetEvalType PtrData = compute_(GEP.getPointerOperand());
 886   if (!bothKnown(PtrData))
 887     return unknown();
 888
 889   Value *Offset = EmitGEPOffset(&Builder, DL, &GEP, /*NoAssumptions=*/true);
 890   Offset = Builder.CreateAdd(PtrData.second, Offset);
 891   return std::make_pair(PtrData.first, Offset);
 892 }
 893
 894 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitIntToPtrInst(IntToPtrInst&) {
 895   // clueless
 896   return unknown();
 897 }
 898
 899 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitLoadInst(LoadInst&) {
 900   return unknown();
 901 }
 902
 903 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitPHINode(PHINode &PHI) {
 904   // Create 2 PHIs: one for size and another for offset.
 905   PHINode *SizePHI   = Builder.CreatePHI(IntTy, PHI.getNumIncomingValues());
 906   PHINode *OffsetPHI = Builder.CreatePHI(IntTy, PHI.getNumIncomingValues());
 907
 908   // Insert right away in the cache to handle recursive PHIs.
 909   CacheMap[&PHI] = std::make_pair(SizePHI, OffsetPHI);
 910
 911   // Compute offset/size for each PHI incoming pointer.
 912   for (unsigned i = 0, e = PHI.getNumIncomingValues(); i != e; ++i) {
 913     Builder.SetInsertPoint(&*PHI.getIncomingBlock(i)->getFirstInsertionPt());
 914     SizeOffsetEvalType EdgeData = compute_(PHI.getIncomingValue(i));
 915
 916     if (!bothKnown(EdgeData)) {
 917       OffsetPHI->replaceAllUsesWith(UndefValue::get(IntTy));
 918       OffsetPHI->eraseFromParent();
 919       SizePHI->replaceAllUsesWith(UndefValue::get(IntTy));
 920       SizePHI->eraseFromParent();
 921       return unknown();
 922     }
 923     SizePHI->addIncoming(EdgeData.first, PHI.getIncomingBlock(i));
 924     OffsetPHI->addIncoming(EdgeData.second, PHI.getIncomingBlock(i));
 925   }
 926
 927   Value *Size = SizePHI, *Offset = OffsetPHI, *Tmp;
 928   if ((Tmp = SizePHI->hasConstantValue())) {
 929     Size = Tmp;
 930     SizePHI->replaceAllUsesWith(Size);
 931     SizePHI->eraseFromParent();
 932   }
 933   if ((Tmp = OffsetPHI->hasConstantValue())) {
 934     Offset = Tmp;
 935     OffsetPHI->replaceAllUsesWith(Offset);
 936     OffsetPHI->eraseFromParent();
 937   }
 938   return std::make_pair(Size, Offset);
 939 }
 940
 941 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitSelectInst(SelectInst &I) {
 942   SizeOffsetEvalType TrueSide  = compute_(I.getTrueValue());
 943   SizeOffsetEvalType FalseSide = compute_(I.getFalseValue());
 944
 945   if (!bothKnown(TrueSide) || !bothKnown(FalseSide))
 946     return unknown();
 947   if (TrueSide == FalseSide)
 948     return TrueSide;
 949
 950   Value *Size = Builder.CreateSelect(I.getCondition(), TrueSide.first,
 951                                      FalseSide.first);
 952   Value *Offset = Builder.CreateSelect(I.getCondition(), TrueSide.second,
 953                                        FalseSide.second);
 954   return std::make_pair(Size, Offset);
 955 }
 956
 957 SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitInstruction(Instruction &I) {
 958   LLVM_DEBUG(dbgs() << "ObjectSizeOffsetEvaluator unknown instruction:" << I
 959                     << '\n');
 960   return unknown();
 961 }