MemoryBuiltins.cpp revision e43e2d8cf0a5a05a1d1ed5abbc2134834a9b06d2 
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"
31using namespace llvm;
32
33enum 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
42struct 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.
52static 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
72static 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.
88static const AllocFnsTy *getAllocationData(const Value *V, AllocType AllocTy,
89                                           const TargetLibraryInfo *TLI,
90                                           bool LookThroughBitCast = false) {
91  // Skip intrinsics
92  if (isa<IntrinsicInst>(V))
93    return 0;
94
95  Function *Callee = getCalledFunction(V, LookThroughBitCast);
96  if (!Callee)
97    return 0;
98
99  // Make sure that the function is available.
100  StringRef FnName = Callee->getName();
101  LibFunc::Func TLIFn;
102  if (!TLI || !TLI->getLibFunc(FnName, TLIFn) || !TLI->has(TLIFn))
103    return 0;
104
105  unsigned i = 0;
106  bool found = false;
107  for ( ; i < array_lengthof(AllocationFnData); ++i) {
108    if (AllocationFnData[i].Func == TLIFn) {
109      found = true;
110      break;
111    }
112  }
113  if (!found)
114    return 0;
115
116  const AllocFnsTy *FnData = &AllocationFnData[i];
117  if ((FnData->AllocTy & AllocTy) == 0)
118    return 0;
119
120  // Check function prototype.
121  int FstParam = FnData->FstParam;
122  int SndParam = FnData->SndParam;
123  FunctionType *FTy = Callee->getFunctionType();
124
125  if (FTy->getReturnType() == Type::getInt8PtrTy(FTy->getContext()) &&
126      FTy->getNumParams() == FnData->NumParams &&
127      (FstParam < 0 ||
128       (FTy->getParamType(FstParam)->isIntegerTy(32) ||
129        FTy->getParamType(FstParam)->isIntegerTy(64))) &&
130      (SndParam < 0 ||
131       FTy->getParamType(SndParam)->isIntegerTy(32) ||
132       FTy->getParamType(SndParam)->isIntegerTy(64)))
133    return FnData;
134  return 0;
135}
136
137static bool hasNoAliasAttr(const Value *V, bool LookThroughBitCast) {
138  ImmutableCallSite CS(LookThroughBitCast ? V->stripPointerCasts() : V);
139  return CS && CS.hasFnAttr(Attribute::NoAlias);
140}
141
142
143/// \brief Tests if a value is a call or invoke to a library function that
144/// allocates or reallocates memory (either malloc, calloc, realloc, or strdup
145/// like).
146bool llvm::isAllocationFn(const Value *V, const TargetLibraryInfo *TLI,
147                          bool LookThroughBitCast) {
148  return getAllocationData(V, AnyAlloc, TLI, LookThroughBitCast);
149}
150
151/// \brief Tests if a value is a call or invoke to a function that returns a
152/// NoAlias pointer (including malloc/calloc/realloc/strdup-like functions).
153bool llvm::isNoAliasFn(const Value *V, const TargetLibraryInfo *TLI,
154                       bool LookThroughBitCast) {
155  // it's safe to consider realloc as noalias since accessing the original
156  // pointer is undefined behavior
157  return isAllocationFn(V, TLI, LookThroughBitCast) ||
158         hasNoAliasAttr(V, LookThroughBitCast);
159}
160
161/// \brief Tests if a value is a call or invoke to a library function that
162/// allocates uninitialized memory (such as malloc).
163bool llvm::isMallocLikeFn(const Value *V, const TargetLibraryInfo *TLI,
164                          bool LookThroughBitCast) {
165  return getAllocationData(V, MallocLike, TLI, LookThroughBitCast);
166}
167
168/// \brief Tests if a value is a call or invoke to a library function that
169/// allocates zero-filled memory (such as calloc).
170bool llvm::isCallocLikeFn(const Value *V, const TargetLibraryInfo *TLI,
171                          bool LookThroughBitCast) {
172  return getAllocationData(V, CallocLike, TLI, LookThroughBitCast);
173}
174
175/// \brief Tests if a value is a call or invoke to a library function that
176/// allocates memory (either malloc, calloc, or strdup like).
177bool llvm::isAllocLikeFn(const Value *V, const TargetLibraryInfo *TLI,
178                         bool LookThroughBitCast) {
179  return getAllocationData(V, AllocLike, TLI, LookThroughBitCast);
180}
181
182/// \brief Tests if a value is a call or invoke to a library function that
183/// reallocates memory (such as realloc).
184bool llvm::isReallocLikeFn(const Value *V, const TargetLibraryInfo *TLI,
185                           bool LookThroughBitCast) {
186  return getAllocationData(V, ReallocLike, TLI, LookThroughBitCast);
187}
188
189/// extractMallocCall - Returns the corresponding CallInst if the instruction
190/// is a malloc call.  Since CallInst::CreateMalloc() only creates calls, we
191/// ignore InvokeInst here.
192const CallInst *llvm::extractMallocCall(const Value *I,
193                                        const TargetLibraryInfo *TLI) {
194  return isMallocLikeFn(I, TLI) ? dyn_cast<CallInst>(I) : 0;
195}
196
197static Value *computeArraySize(const CallInst *CI, const DataLayout *TD,
198                               const TargetLibraryInfo *TLI,
199                               bool LookThroughSExt = false) {
200  if (!CI)
201    return 0;
202
203  // The size of the malloc's result type must be known to determine array size.
204  Type *T = getMallocAllocatedType(CI, TLI);
205  if (!T || !T->isSized() || !TD)
206    return 0;
207
208  unsigned ElementSize = TD->getTypeAllocSize(T);
209  if (StructType *ST = dyn_cast<StructType>(T))
210    ElementSize = TD->getStructLayout(ST)->getSizeInBytes();
211
212  // If malloc call's arg can be determined to be a multiple of ElementSize,
213  // return the multiple.  Otherwise, return NULL.
214  Value *MallocArg = CI->getArgOperand(0);
215  Value *Multiple = 0;
216  if (ComputeMultiple(MallocArg, ElementSize, Multiple,
217                      LookThroughSExt))
218    return Multiple;
219
220  return 0;
221}
222
223/// isArrayMalloc - Returns the corresponding CallInst if the instruction
224/// is a call to malloc whose array size can be determined and the array size
225/// is not constant 1.  Otherwise, return NULL.
226const CallInst *llvm::isArrayMalloc(const Value *I,
227                                    const DataLayout *TD,
228                                    const TargetLibraryInfo *TLI) {
229  const CallInst *CI = extractMallocCall(I, TLI);
230  Value *ArraySize = computeArraySize(CI, TD, TLI);
231
232  if (ConstantInt *ConstSize = dyn_cast_or_null<ConstantInt>(ArraySize))
233    if (ConstSize->isOne())
234      return CI;
235
236  // CI is a non-array malloc or we can't figure out that it is an array malloc.
237  return 0;
238}
239
240/// getMallocType - Returns the PointerType resulting from the malloc call.
241/// The PointerType depends on the number of bitcast uses of the malloc call:
242///   0: PointerType is the calls' return type.
243///   1: PointerType is the bitcast's result type.
244///  >1: Unique PointerType cannot be determined, return NULL.
245PointerType *llvm::getMallocType(const CallInst *CI,
246                                 const TargetLibraryInfo *TLI) {
247  assert(isMallocLikeFn(CI, TLI) && "getMallocType and not malloc call");
248
249  PointerType *MallocType = 0;
250  unsigned NumOfBitCastUses = 0;
251
252  // Determine if CallInst has a bitcast use.
253  for (Value::const_use_iterator UI = CI->use_begin(), E = CI->use_end();
254       UI != E; )
255    if (const BitCastInst *BCI = dyn_cast<BitCastInst>(*UI++)) {
256      MallocType = cast<PointerType>(BCI->getDestTy());
257      NumOfBitCastUses++;
258    }
259
260  // Malloc call has 1 bitcast use, so type is the bitcast's destination type.
261  if (NumOfBitCastUses == 1)
262    return MallocType;
263
264  // Malloc call was not bitcast, so type is the malloc function's return type.
265  if (NumOfBitCastUses == 0)
266    return cast<PointerType>(CI->getType());
267
268  // Type could not be determined.
269  return 0;
270}
271
272/// getMallocAllocatedType - Returns the Type allocated by malloc call.
273/// The Type depends on the number of bitcast uses of the malloc call:
274///   0: PointerType is the malloc calls' return type.
275///   1: PointerType is the bitcast's result type.
276///  >1: Unique PointerType cannot be determined, return NULL.
277Type *llvm::getMallocAllocatedType(const CallInst *CI,
278                                   const TargetLibraryInfo *TLI) {
279  PointerType *PT = getMallocType(CI, TLI);
280  return PT ? PT->getElementType() : 0;
281}
282
283/// getMallocArraySize - Returns the array size of a malloc call.  If the
284/// argument passed to malloc is a multiple of the size of the malloced type,
285/// then return that multiple.  For non-array mallocs, the multiple is
286/// constant 1.  Otherwise, return NULL for mallocs whose array size cannot be
287/// determined.
288Value *llvm::getMallocArraySize(CallInst *CI, const DataLayout *TD,
289                                const TargetLibraryInfo *TLI,
290                                bool LookThroughSExt) {
291  assert(isMallocLikeFn(CI, TLI) && "getMallocArraySize and not malloc call");
292  return computeArraySize(CI, TD, TLI, LookThroughSExt);
293}
294
295
296/// extractCallocCall - Returns the corresponding CallInst if the instruction
297/// is a calloc call.
298const CallInst *llvm::extractCallocCall(const Value *I,
299                                        const TargetLibraryInfo *TLI) {
300  return isCallocLikeFn(I, TLI) ? cast<CallInst>(I) : 0;
301}
302
303
304/// isFreeCall - Returns non-null if the value is a call to the builtin free()
305const CallInst *llvm::isFreeCall(const Value *I, const TargetLibraryInfo *TLI) {
306  const CallInst *CI = dyn_cast<CallInst>(I);
307  if (!CI || isa<IntrinsicInst>(CI))
308    return 0;
309  Function *Callee = CI->getCalledFunction();
310  if (Callee == 0 || !Callee->isDeclaration())
311    return 0;
312
313  StringRef FnName = Callee->getName();
314  LibFunc::Func TLIFn;
315  if (!TLI || !TLI->getLibFunc(FnName, TLIFn) || !TLI->has(TLIFn))
316    return 0;
317
318  if (TLIFn != LibFunc::free &&
319      TLIFn != LibFunc::ZdlPv && // operator delete(void*)
320      TLIFn != LibFunc::ZdaPv)   // operator delete[](void*)
321    return 0;
322
323  // Check free prototype.
324  // FIXME: workaround for PR5130, this will be obsolete when a nobuiltin
325  // attribute will exist.
326  FunctionType *FTy = Callee->getFunctionType();
327  if (!FTy->getReturnType()->isVoidTy())
328    return 0;
329  if (FTy->getNumParams() != 1)
330    return 0;
331  if (FTy->getParamType(0) != Type::getInt8PtrTy(Callee->getContext()))
332    return 0;
333
334  return CI;
335}
336
337
338
339//===----------------------------------------------------------------------===//
340//  Utility functions to compute size of objects.
341//
342
343
344/// \brief Compute the size of the object pointed by Ptr. Returns true and the
345/// object size in Size if successful, and false otherwise.
346/// If RoundToAlign is true, then Size is rounded up to the aligment of allocas,
347/// byval arguments, and global variables.
348bool llvm::getObjectSize(const Value *Ptr, uint64_t &Size, const DataLayout *TD,
349                         const TargetLibraryInfo *TLI, bool RoundToAlign) {
350  if (!TD)
351    return false;
352
353  ObjectSizeOffsetVisitor Visitor(TD, TLI, Ptr->getContext(), RoundToAlign);
354  SizeOffsetType Data = Visitor.compute(const_cast<Value*>(Ptr));
355  if (!Visitor.bothKnown(Data))
356    return false;
357
358  APInt ObjSize = Data.first, Offset = Data.second;
359  // check for overflow
360  if (Offset.slt(0) || ObjSize.ult(Offset))
361    Size = 0;
362  else
363    Size = (ObjSize - Offset).getZExtValue();
364  return true;
365}
366
367/// \brief Compute the size of the underlying object pointed by Ptr. Returns
368/// true and the object size in Size if successful, and false otherwise.
369/// If RoundToAlign is true, then Size is rounded up to the aligment of allocas,
370/// byval arguments, and global variables.
371bool llvm::getUnderlyingObjectSize(const Value *Ptr, uint64_t &Size,
372                                   const DataLayout *TD,
373                                   const TargetLibraryInfo *TLI,
374                                   bool RoundToAlign) {
375  if (!TD)
376    return false;
377
378  ObjectSizeOffsetVisitor Visitor(TD, TLI, Ptr->getContext(), RoundToAlign);
379  SizeOffsetType Data = Visitor.compute(const_cast<Value*>(Ptr));
380  if (!Visitor.knownSize(Data))
381    return false;
382
383  Size = Data.first.getZExtValue();
384  return true;
385}
386
387
388STATISTIC(ObjectVisitorArgument,
389          "Number of arguments with unsolved size and offset");
390STATISTIC(ObjectVisitorLoad,
391          "Number of load instructions with unsolved size and offset");
392
393
394APInt ObjectSizeOffsetVisitor::align(APInt Size, uint64_t Align) {
395  if (RoundToAlign && Align)
396    return APInt(IntTyBits, RoundUpToAlignment(Size.getZExtValue(), Align));
397  return Size;
398}
399
400ObjectSizeOffsetVisitor::ObjectSizeOffsetVisitor(const DataLayout *TD,
401                                                 const TargetLibraryInfo *TLI,
402                                                 LLVMContext &Context,
403                                                 bool RoundToAlign)
404: TD(TD), TLI(TLI), RoundToAlign(RoundToAlign) {
405  IntegerType *IntTy = TD->getIntPtrType(Context);
406  IntTyBits = IntTy->getBitWidth();
407  Zero = APInt::getNullValue(IntTyBits);
408}
409
410SizeOffsetType ObjectSizeOffsetVisitor::compute(Value *V) {
411  V = V->stripPointerCasts();
412
413  if (isa<Instruction>(V) || isa<GEPOperator>(V)) {
414    // Return cached value or insert unknown in cache if size of V was not
415    // computed yet in order to avoid recursions in PHis.
416    std::pair<CacheMapTy::iterator, bool> CacheVal =
417      CacheMap.insert(std::make_pair(V, unknown()));
418    if (!CacheVal.second)
419      return CacheVal.first->second;
420
421    SizeOffsetType Result;
422    if (GEPOperator *GEP = dyn_cast<GEPOperator>(V))
423      Result = visitGEPOperator(*GEP);
424    else
425      Result = visit(cast<Instruction>(*V));
426    return CacheMap[V] = Result;
427  }
428
429  if (Argument *A = dyn_cast<Argument>(V))
430    return visitArgument(*A);
431  if (ConstantPointerNull *P = dyn_cast<ConstantPointerNull>(V))
432    return visitConstantPointerNull(*P);
433  if (GlobalAlias *GA = dyn_cast<GlobalAlias>(V))
434    return visitGlobalAlias(*GA);
435  if (GlobalVariable *GV = dyn_cast<GlobalVariable>(V))
436    return visitGlobalVariable(*GV);
437  if (UndefValue *UV = dyn_cast<UndefValue>(V))
438    return visitUndefValue(*UV);
439  if (ConstantExpr *CE = dyn_cast<ConstantExpr>(V)) {
440    if (CE->getOpcode() == Instruction::IntToPtr)
441      return unknown(); // clueless
442  }
443
444  DEBUG(dbgs() << "ObjectSizeOffsetVisitor::compute() unhandled value: " << *V
445        << '\n');
446  return unknown();
447}
448
449SizeOffsetType ObjectSizeOffsetVisitor::visitAllocaInst(AllocaInst &I) {
450  if (!I.getAllocatedType()->isSized())
451    return unknown();
452
453  APInt Size(IntTyBits, TD->getTypeAllocSize(I.getAllocatedType()));
454  if (!I.isArrayAllocation())
455    return std::make_pair(align(Size, I.getAlignment()), Zero);
456
457  Value *ArraySize = I.getArraySize();
458  if (const ConstantInt *C = dyn_cast<ConstantInt>(ArraySize)) {
459    Size *= C->getValue().zextOrSelf(IntTyBits);
460    return std::make_pair(align(Size, I.getAlignment()), Zero);
461  }
462  return unknown();
463}
464
465SizeOffsetType ObjectSizeOffsetVisitor::visitArgument(Argument &A) {
466  // no interprocedural analysis is done at the moment
467  if (!A.hasByValAttr()) {
468    ++ObjectVisitorArgument;
469    return unknown();
470  }
471  PointerType *PT = cast<PointerType>(A.getType());
472  APInt Size(IntTyBits, TD->getTypeAllocSize(PT->getElementType()));
473  return std::make_pair(align(Size, A.getParamAlignment()), Zero);
474}
475
476SizeOffsetType ObjectSizeOffsetVisitor::visitCallSite(CallSite CS) {
477  const AllocFnsTy *FnData = getAllocationData(CS.getInstruction(), AnyAlloc,
478                                               TLI);
479  if (!FnData)
480    return unknown();
481
482  // handle strdup-like functions separately
483  if (FnData->AllocTy == StrDupLike) {
484    APInt Size(IntTyBits, GetStringLength(CS.getArgument(0)));
485    if (!Size)
486      return unknown();
487
488    // strndup limits strlen
489    if (FnData->FstParam > 0) {
490      ConstantInt *Arg= dyn_cast<ConstantInt>(CS.getArgument(FnData->FstParam));
491      if (!Arg)
492        return unknown();
493
494      APInt MaxSize = Arg->getValue().zextOrSelf(IntTyBits);
495      if (Size.ugt(MaxSize))
496        Size = MaxSize + 1;
497    }
498    return std::make_pair(Size, Zero);
499  }
500
501  ConstantInt *Arg = dyn_cast<ConstantInt>(CS.getArgument(FnData->FstParam));
502  if (!Arg)
503    return unknown();
504
505  APInt Size = Arg->getValue().zextOrSelf(IntTyBits);
506  // size determined by just 1 parameter
507  if (FnData->SndParam < 0)
508    return std::make_pair(Size, Zero);
509
510  Arg = dyn_cast<ConstantInt>(CS.getArgument(FnData->SndParam));
511  if (!Arg)
512    return unknown();
513
514  Size *= Arg->getValue().zextOrSelf(IntTyBits);
515  return std::make_pair(Size, Zero);
516
517  // TODO: handle more standard functions (+ wchar cousins):
518  // - strdup / strndup
519  // - strcpy / strncpy
520  // - strcat / strncat
521  // - memcpy / memmove
522  // - strcat / strncat
523  // - memset
524}
525
526SizeOffsetType
527ObjectSizeOffsetVisitor::visitConstantPointerNull(ConstantPointerNull&) {
528  return std::make_pair(Zero, Zero);
529}
530
531SizeOffsetType
532ObjectSizeOffsetVisitor::visitExtractElementInst(ExtractElementInst&) {
533  return unknown();
534}
535
536SizeOffsetType
537ObjectSizeOffsetVisitor::visitExtractValueInst(ExtractValueInst&) {
538  // Easy cases were already folded by previous passes.
539  return unknown();
540}
541
542SizeOffsetType ObjectSizeOffsetVisitor::visitGEPOperator(GEPOperator &GEP) {
543  SizeOffsetType PtrData = compute(GEP.getPointerOperand());
544  APInt Offset(IntTyBits, 0);
545  if (!bothKnown(PtrData) || !GEP.accumulateConstantOffset(*TD, Offset))
546    return unknown();
547
548  return std::make_pair(PtrData.first, PtrData.second + Offset);
549}
550
551SizeOffsetType ObjectSizeOffsetVisitor::visitGlobalAlias(GlobalAlias &GA) {
552  if (GA.mayBeOverridden())
553    return unknown();
554  return compute(GA.getAliasee());
555}
556
557SizeOffsetType ObjectSizeOffsetVisitor::visitGlobalVariable(GlobalVariable &GV){
558  if (!GV.hasDefinitiveInitializer())
559    return unknown();
560
561  APInt Size(IntTyBits, TD->getTypeAllocSize(GV.getType()->getElementType()));
562  return std::make_pair(align(Size, GV.getAlignment()), Zero);
563}
564
565SizeOffsetType ObjectSizeOffsetVisitor::visitIntToPtrInst(IntToPtrInst&) {
566  // clueless
567  return unknown();
568}
569
570SizeOffsetType ObjectSizeOffsetVisitor::visitLoadInst(LoadInst&) {
571  ++ObjectVisitorLoad;
572  return unknown();
573}
574
575SizeOffsetType ObjectSizeOffsetVisitor::visitPHINode(PHINode &PHI) {
576  if (PHI.getNumIncomingValues() == 0)
577    return unknown();
578
579  SizeOffsetType Ret = compute(PHI.getIncomingValue(0));
580  if (!bothKnown(Ret))
581    return unknown();
582
583  // Verify that all PHI incoming pointers have the same size and offset.
584  for (unsigned i = 1, e = PHI.getNumIncomingValues(); i != e; ++i) {
585    SizeOffsetType EdgeData = compute(PHI.getIncomingValue(i));
586    if (!bothKnown(EdgeData) || EdgeData != Ret)
587      return unknown();
588  }
589  return Ret;
590}
591
592SizeOffsetType ObjectSizeOffsetVisitor::visitSelectInst(SelectInst &I) {
593  SizeOffsetType TrueSide  = compute(I.getTrueValue());
594  SizeOffsetType FalseSide = compute(I.getFalseValue());
595  if (bothKnown(TrueSide) && bothKnown(FalseSide) && TrueSide == FalseSide)
596    return TrueSide;
597  return unknown();
598}
599
600SizeOffsetType ObjectSizeOffsetVisitor::visitUndefValue(UndefValue&) {
601  return std::make_pair(Zero, Zero);
602}
603
604SizeOffsetType ObjectSizeOffsetVisitor::visitInstruction(Instruction &I) {
605  DEBUG(dbgs() << "ObjectSizeOffsetVisitor unknown instruction:" << I << '\n');
606  return unknown();
607}
608
609
610ObjectSizeOffsetEvaluator::ObjectSizeOffsetEvaluator(const DataLayout *TD,
611                                                   const TargetLibraryInfo *TLI,
612                                                     LLVMContext &Context)
613: TD(TD), TLI(TLI), Context(Context), Builder(Context, TargetFolder(TD)) {
614  IntTy = TD->getIntPtrType(Context);
615  Zero = ConstantInt::get(IntTy, 0);
616}
617
618SizeOffsetEvalType ObjectSizeOffsetEvaluator::compute(Value *V) {
619  SizeOffsetEvalType Result = compute_(V);
620
621  if (!bothKnown(Result)) {
622    // erase everything that was computed in this iteration from the cache, so
623    // that no dangling references are left behind. We could be a bit smarter if
624    // we kept a dependency graph. It's probably not worth the complexity.
625    for (PtrSetTy::iterator I=SeenVals.begin(), E=SeenVals.end(); I != E; ++I) {
626      CacheMapTy::iterator CacheIt = CacheMap.find(*I);
627      // non-computable results can be safely cached
628      if (CacheIt != CacheMap.end() && anyKnown(CacheIt->second))
629        CacheMap.erase(CacheIt);
630    }
631  }
632
633  SeenVals.clear();
634  return Result;
635}
636
637SizeOffsetEvalType ObjectSizeOffsetEvaluator::compute_(Value *V) {
638  ObjectSizeOffsetVisitor Visitor(TD, TLI, Context);
639  SizeOffsetType Const = Visitor.compute(V);
640  if (Visitor.bothKnown(Const))
641    return std::make_pair(ConstantInt::get(Context, Const.first),
642                          ConstantInt::get(Context, Const.second));
643
644  V = V->stripPointerCasts();
645
646  // check cache
647  CacheMapTy::iterator CacheIt = CacheMap.find(V);
648  if (CacheIt != CacheMap.end())
649    return CacheIt->second;
650
651  // always generate code immediately before the instruction being
652  // processed, so that the generated code dominates the same BBs
653  Instruction *PrevInsertPoint = Builder.GetInsertPoint();
654  if (Instruction *I = dyn_cast<Instruction>(V))
655    Builder.SetInsertPoint(I);
656
657  // record the pointers that were handled in this run, so that they can be
658  // cleaned later if something fails
659  SeenVals.insert(V);
660
661  // now compute the size and offset
662  SizeOffsetEvalType Result;
663  if (GEPOperator *GEP = dyn_cast<GEPOperator>(V)) {
664    Result = visitGEPOperator(*GEP);
665  } else if (Instruction *I = dyn_cast<Instruction>(V)) {
666    Result = visit(*I);
667  } else if (isa<Argument>(V) ||
668             (isa<ConstantExpr>(V) &&
669              cast<ConstantExpr>(V)->getOpcode() == Instruction::IntToPtr) ||
670             isa<GlobalAlias>(V) ||
671             isa<GlobalVariable>(V)) {
672    // ignore values where we cannot do more than what ObjectSizeVisitor can
673    Result = unknown();
674  } else {
675    DEBUG(dbgs() << "ObjectSizeOffsetEvaluator::compute() unhandled value: "
676          << *V << '\n');
677    Result = unknown();
678  }
679
680  if (PrevInsertPoint)
681    Builder.SetInsertPoint(PrevInsertPoint);
682
683  // Don't reuse CacheIt since it may be invalid at this point.
684  CacheMap[V] = Result;
685  return Result;
686}
687
688SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitAllocaInst(AllocaInst &I) {
689  if (!I.getAllocatedType()->isSized())
690    return unknown();
691
692  // must be a VLA
693  assert(I.isArrayAllocation());
694  Value *ArraySize = I.getArraySize();
695  Value *Size = ConstantInt::get(ArraySize->getType(),
696                                 TD->getTypeAllocSize(I.getAllocatedType()));
697  Size = Builder.CreateMul(Size, ArraySize);
698  return std::make_pair(Size, Zero);
699}
700
701SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitCallSite(CallSite CS) {
702  const AllocFnsTy *FnData = getAllocationData(CS.getInstruction(), AnyAlloc,
703                                               TLI);
704  if (!FnData)
705    return unknown();
706
707  // handle strdup-like functions separately
708  if (FnData->AllocTy == StrDupLike) {
709    // TODO
710    return unknown();
711  }
712
713  Value *FirstArg = CS.getArgument(FnData->FstParam);
714  FirstArg = Builder.CreateZExt(FirstArg, IntTy);
715  if (FnData->SndParam < 0)
716    return std::make_pair(FirstArg, Zero);
717
718  Value *SecondArg = CS.getArgument(FnData->SndParam);
719  SecondArg = Builder.CreateZExt(SecondArg, IntTy);
720  Value *Size = Builder.CreateMul(FirstArg, SecondArg);
721  return std::make_pair(Size, Zero);
722
723  // TODO: handle more standard functions (+ wchar cousins):
724  // - strdup / strndup
725  // - strcpy / strncpy
726  // - strcat / strncat
727  // - memcpy / memmove
728  // - strcat / strncat
729  // - memset
730}
731
732SizeOffsetEvalType
733ObjectSizeOffsetEvaluator::visitExtractElementInst(ExtractElementInst&) {
734  return unknown();
735}
736
737SizeOffsetEvalType
738ObjectSizeOffsetEvaluator::visitExtractValueInst(ExtractValueInst&) {
739  return unknown();
740}
741
742SizeOffsetEvalType
743ObjectSizeOffsetEvaluator::visitGEPOperator(GEPOperator &GEP) {
744  SizeOffsetEvalType PtrData = compute_(GEP.getPointerOperand());
745  if (!bothKnown(PtrData))
746    return unknown();
747
748  Value *Offset = EmitGEPOffset(&Builder, *TD, &GEP, /*NoAssumptions=*/true);
749  Offset = Builder.CreateAdd(PtrData.second, Offset);
750  return std::make_pair(PtrData.first, Offset);
751}
752
753SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitIntToPtrInst(IntToPtrInst&) {
754  // clueless
755  return unknown();
756}
757
758SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitLoadInst(LoadInst&) {
759  return unknown();
760}
761
762SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitPHINode(PHINode &PHI) {
763  // create 2 PHIs: one for size and another for offset
764  PHINode *SizePHI   = Builder.CreatePHI(IntTy, PHI.getNumIncomingValues());
765  PHINode *OffsetPHI = Builder.CreatePHI(IntTy, PHI.getNumIncomingValues());
766
767  // insert right away in the cache to handle recursive PHIs
768  CacheMap[&PHI] = std::make_pair(SizePHI, OffsetPHI);
769
770  // compute offset/size for each PHI incoming pointer
771  for (unsigned i = 0, e = PHI.getNumIncomingValues(); i != e; ++i) {
772    Builder.SetInsertPoint(PHI.getIncomingBlock(i)->getFirstInsertionPt());
773    SizeOffsetEvalType EdgeData = compute_(PHI.getIncomingValue(i));
774
775    if (!bothKnown(EdgeData)) {
776      OffsetPHI->replaceAllUsesWith(UndefValue::get(IntTy));
777      OffsetPHI->eraseFromParent();
778      SizePHI->replaceAllUsesWith(UndefValue::get(IntTy));
779      SizePHI->eraseFromParent();
780      return unknown();
781    }
782    SizePHI->addIncoming(EdgeData.first, PHI.getIncomingBlock(i));
783    OffsetPHI->addIncoming(EdgeData.second, PHI.getIncomingBlock(i));
784  }
785
786  Value *Size = SizePHI, *Offset = OffsetPHI, *Tmp;
787  if ((Tmp = SizePHI->hasConstantValue())) {
788    Size = Tmp;
789    SizePHI->replaceAllUsesWith(Size);
790    SizePHI->eraseFromParent();
791  }
792  if ((Tmp = OffsetPHI->hasConstantValue())) {
793    Offset = Tmp;
794    OffsetPHI->replaceAllUsesWith(Offset);
795    OffsetPHI->eraseFromParent();
796  }
797  return std::make_pair(Size, Offset);
798}
799
800SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitSelectInst(SelectInst &I) {
801  SizeOffsetEvalType TrueSide  = compute_(I.getTrueValue());
802  SizeOffsetEvalType FalseSide = compute_(I.getFalseValue());
803
804  if (!bothKnown(TrueSide) || !bothKnown(FalseSide))
805    return unknown();
806  if (TrueSide == FalseSide)
807    return TrueSide;
808
809  Value *Size = Builder.CreateSelect(I.getCondition(), TrueSide.first,
810                                     FalseSide.first);
811  Value *Offset = Builder.CreateSelect(I.getCondition(), TrueSide.second,
812                                       FalseSide.second);
813  return std::make_pair(Size, Offset);
814}
815
816SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitInstruction(Instruction &I) {
817  DEBUG(dbgs() << "ObjectSizeOffsetEvaluator unknown instruction:" << I <<'\n');
818  return unknown();
819}
820