MemoryBuiltins.cpp revision a9cd5164c404905a8109da5bfedea9d68960990d 
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  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
133static 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).
142bool 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).
149bool 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).
159bool 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).
166bool 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).
173bool 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).
180bool 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.
188const CallInst *llvm::extractMallocCall(const Value *I,
189                                        const TargetLibraryInfo *TLI) {
190  return isMallocLikeFn(I, TLI) ? dyn_cast<CallInst>(I) : 0;
191}
192
193static Value *computeArraySize(const CallInst *CI, const DataLayout *TD,
194                               const TargetLibraryInfo *TLI,
195                               bool LookThroughSExt = false) {
196  if (!CI)
197    return 0;
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 0;
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 = 0;
212  if (ComputeMultiple(MallocArg, ElementSize, Multiple,
213                      LookThroughSExt))
214    return Multiple;
215
216  return 0;
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.
222const CallInst *llvm::isArrayMalloc(const Value *I,
223                                    const DataLayout *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 0;
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.
241PointerType *llvm::getMallocType(const CallInst *CI,
242                                 const TargetLibraryInfo *TLI) {
243  assert(isMallocLikeFn(CI, TLI) && "getMallocType and not malloc call");
244
245  PointerType *MallocType = 0;
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 0;
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.
273Type *llvm::getMallocAllocatedType(const CallInst *CI,
274                                   const TargetLibraryInfo *TLI) {
275  PointerType *PT = getMallocType(CI, TLI);
276  return PT ? PT->getElementType() : 0;
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.
284Value *llvm::getMallocArraySize(CallInst *CI, const DataLayout *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.
294const 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()
301const 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.
344bool llvm::getObjectSize(const Value *Ptr, uint64_t &Size, const DataLayout *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/// \brief Compute the size of the underlying object pointed by Ptr. Returns
364/// true and the object size in Size if successful, and false otherwise.
365/// If RoundToAlign is true, then Size is rounded up to the aligment of allocas,
366/// byval arguments, and global variables.
367bool llvm::getUnderlyingObjectSize(const Value *Ptr, uint64_t &Size,
368                                   const DataLayout *TD,
369                                   const TargetLibraryInfo *TLI,
370                                   bool RoundToAlign) {
371  if (!TD)
372    return false;
373
374  ObjectSizeOffsetVisitor Visitor(TD, TLI, Ptr->getContext(), RoundToAlign);
375  SizeOffsetType Data = Visitor.compute(const_cast<Value*>(Ptr));
376  if (!Visitor.knownSize(Data))
377    return false;
378
379  Size = Data.first.getZExtValue();
380  return true;
381}
382
383
384STATISTIC(ObjectVisitorArgument,
385          "Number of arguments with unsolved size and offset");
386STATISTIC(ObjectVisitorLoad,
387          "Number of load instructions with unsolved size and offset");
388
389
390APInt ObjectSizeOffsetVisitor::align(APInt Size, uint64_t Align) {
391  if (RoundToAlign && Align)
392    return APInt(IntTyBits, RoundUpToAlignment(Size.getZExtValue(), Align));
393  return Size;
394}
395
396ObjectSizeOffsetVisitor::ObjectSizeOffsetVisitor(const DataLayout *TD,
397                                                 const TargetLibraryInfo *TLI,
398                                                 LLVMContext &Context,
399                                                 bool RoundToAlign)
400: TD(TD), TLI(TLI), RoundToAlign(RoundToAlign) {
401  IntegerType *IntTy = TD->getIntPtrType(Context);
402  IntTyBits = IntTy->getBitWidth();
403  Zero = APInt::getNullValue(IntTyBits);
404}
405
406SizeOffsetType ObjectSizeOffsetVisitor::compute(Value *V) {
407  V = V->stripPointerCasts();
408
409  if (isa<Instruction>(V) || isa<GEPOperator>(V)) {
410    // Return cached value or insert unknown in cache if size of V was not
411    // computed yet in order to avoid recursions in PHis.
412    std::pair<CacheMapTy::iterator, bool> CacheVal =
413      CacheMap.insert(std::make_pair(V, unknown()));
414    if (!CacheVal.second)
415      return CacheVal.first->second;
416
417    SizeOffsetType Result;
418    if (GEPOperator *GEP = dyn_cast<GEPOperator>(V))
419      Result = visitGEPOperator(*GEP);
420    else
421      Result = visit(cast<Instruction>(*V));
422    return CacheMap[V] = Result;
423  }
424
425  if (Argument *A = dyn_cast<Argument>(V))
426    return visitArgument(*A);
427  if (ConstantPointerNull *P = dyn_cast<ConstantPointerNull>(V))
428    return visitConstantPointerNull(*P);
429  if (GlobalAlias *GA = dyn_cast<GlobalAlias>(V))
430    return visitGlobalAlias(*GA);
431  if (GlobalVariable *GV = dyn_cast<GlobalVariable>(V))
432    return visitGlobalVariable(*GV);
433  if (UndefValue *UV = dyn_cast<UndefValue>(V))
434    return visitUndefValue(*UV);
435  if (ConstantExpr *CE = dyn_cast<ConstantExpr>(V)) {
436    if (CE->getOpcode() == Instruction::IntToPtr)
437      return unknown(); // clueless
438  }
439
440  DEBUG(dbgs() << "ObjectSizeOffsetVisitor::compute() unhandled value: " << *V
441        << '\n');
442  return unknown();
443}
444
445SizeOffsetType ObjectSizeOffsetVisitor::visitAllocaInst(AllocaInst &I) {
446  if (!I.getAllocatedType()->isSized())
447    return unknown();
448
449  APInt Size(IntTyBits, TD->getTypeAllocSize(I.getAllocatedType()));
450  if (!I.isArrayAllocation())
451    return std::make_pair(align(Size, I.getAlignment()), Zero);
452
453  Value *ArraySize = I.getArraySize();
454  if (const ConstantInt *C = dyn_cast<ConstantInt>(ArraySize)) {
455    Size *= C->getValue().zextOrSelf(IntTyBits);
456    return std::make_pair(align(Size, I.getAlignment()), Zero);
457  }
458  return unknown();
459}
460
461SizeOffsetType ObjectSizeOffsetVisitor::visitArgument(Argument &A) {
462  // no interprocedural analysis is done at the moment
463  if (!A.hasByValAttr()) {
464    ++ObjectVisitorArgument;
465    return unknown();
466  }
467  PointerType *PT = cast<PointerType>(A.getType());
468  APInt Size(IntTyBits, TD->getTypeAllocSize(PT->getElementType()));
469  return std::make_pair(align(Size, A.getParamAlignment()), Zero);
470}
471
472SizeOffsetType ObjectSizeOffsetVisitor::visitCallSite(CallSite CS) {
473  const AllocFnsTy *FnData = getAllocationData(CS.getInstruction(), AnyAlloc,
474                                               TLI);
475  if (!FnData)
476    return unknown();
477
478  // handle strdup-like functions separately
479  if (FnData->AllocTy == StrDupLike) {
480    APInt Size(IntTyBits, GetStringLength(CS.getArgument(0)));
481    if (!Size)
482      return unknown();
483
484    // strndup limits strlen
485    if (FnData->FstParam > 0) {
486      ConstantInt *Arg= dyn_cast<ConstantInt>(CS.getArgument(FnData->FstParam));
487      if (!Arg)
488        return unknown();
489
490      APInt MaxSize = Arg->getValue().zextOrSelf(IntTyBits);
491      if (Size.ugt(MaxSize))
492        Size = MaxSize + 1;
493    }
494    return std::make_pair(Size, Zero);
495  }
496
497  ConstantInt *Arg = dyn_cast<ConstantInt>(CS.getArgument(FnData->FstParam));
498  if (!Arg)
499    return unknown();
500
501  APInt Size = Arg->getValue().zextOrSelf(IntTyBits);
502  // size determined by just 1 parameter
503  if (FnData->SndParam < 0)
504    return std::make_pair(Size, Zero);
505
506  Arg = dyn_cast<ConstantInt>(CS.getArgument(FnData->SndParam));
507  if (!Arg)
508    return unknown();
509
510  Size *= Arg->getValue().zextOrSelf(IntTyBits);
511  return std::make_pair(Size, Zero);
512
513  // TODO: handle more standard functions (+ wchar cousins):
514  // - strdup / strndup
515  // - strcpy / strncpy
516  // - strcat / strncat
517  // - memcpy / memmove
518  // - strcat / strncat
519  // - memset
520}
521
522SizeOffsetType
523ObjectSizeOffsetVisitor::visitConstantPointerNull(ConstantPointerNull&) {
524  return std::make_pair(Zero, Zero);
525}
526
527SizeOffsetType
528ObjectSizeOffsetVisitor::visitExtractElementInst(ExtractElementInst&) {
529  return unknown();
530}
531
532SizeOffsetType
533ObjectSizeOffsetVisitor::visitExtractValueInst(ExtractValueInst&) {
534  // Easy cases were already folded by previous passes.
535  return unknown();
536}
537
538SizeOffsetType ObjectSizeOffsetVisitor::visitGEPOperator(GEPOperator &GEP) {
539  SizeOffsetType PtrData = compute(GEP.getPointerOperand());
540  APInt Offset(IntTyBits, 0);
541  if (!bothKnown(PtrData) || !GEP.accumulateConstantOffset(*TD, Offset))
542    return unknown();
543
544  return std::make_pair(PtrData.first, PtrData.second + Offset);
545}
546
547SizeOffsetType ObjectSizeOffsetVisitor::visitGlobalAlias(GlobalAlias &GA) {
548  if (GA.mayBeOverridden())
549    return unknown();
550  return compute(GA.getAliasee());
551}
552
553SizeOffsetType ObjectSizeOffsetVisitor::visitGlobalVariable(GlobalVariable &GV){
554  if (!GV.hasDefinitiveInitializer())
555    return unknown();
556
557  APInt Size(IntTyBits, TD->getTypeAllocSize(GV.getType()->getElementType()));
558  return std::make_pair(align(Size, GV.getAlignment()), Zero);
559}
560
561SizeOffsetType ObjectSizeOffsetVisitor::visitIntToPtrInst(IntToPtrInst&) {
562  // clueless
563  return unknown();
564}
565
566SizeOffsetType ObjectSizeOffsetVisitor::visitLoadInst(LoadInst&) {
567  ++ObjectVisitorLoad;
568  return unknown();
569}
570
571SizeOffsetType ObjectSizeOffsetVisitor::visitPHINode(PHINode &PHI) {
572  if (PHI.getNumIncomingValues() == 0)
573    return unknown();
574
575  SizeOffsetType Ret = compute(PHI.getIncomingValue(0));
576  if (!bothKnown(Ret))
577    return unknown();
578
579  // Verify that all PHI incoming pointers have the same size and offset.
580  for (unsigned i = 1, e = PHI.getNumIncomingValues(); i != e; ++i) {
581    SizeOffsetType EdgeData = compute(PHI.getIncomingValue(i));
582    if (!bothKnown(EdgeData) || EdgeData != Ret)
583      return unknown();
584  }
585  return Ret;
586}
587
588SizeOffsetType ObjectSizeOffsetVisitor::visitSelectInst(SelectInst &I) {
589  SizeOffsetType TrueSide  = compute(I.getTrueValue());
590  SizeOffsetType FalseSide = compute(I.getFalseValue());
591  if (bothKnown(TrueSide) && bothKnown(FalseSide) && TrueSide == FalseSide)
592    return TrueSide;
593  return unknown();
594}
595
596SizeOffsetType ObjectSizeOffsetVisitor::visitUndefValue(UndefValue&) {
597  return std::make_pair(Zero, Zero);
598}
599
600SizeOffsetType ObjectSizeOffsetVisitor::visitInstruction(Instruction &I) {
601  DEBUG(dbgs() << "ObjectSizeOffsetVisitor unknown instruction:" << I << '\n');
602  return unknown();
603}
604
605
606ObjectSizeOffsetEvaluator::ObjectSizeOffsetEvaluator(const DataLayout *TD,
607                                                   const TargetLibraryInfo *TLI,
608                                                     LLVMContext &Context)
609: TD(TD), TLI(TLI), Context(Context), Builder(Context, TargetFolder(TD)) {
610  IntTy = TD->getIntPtrType(Context);
611  Zero = ConstantInt::get(IntTy, 0);
612}
613
614SizeOffsetEvalType ObjectSizeOffsetEvaluator::compute(Value *V) {
615  SizeOffsetEvalType Result = compute_(V);
616
617  if (!bothKnown(Result)) {
618    // erase everything that was computed in this iteration from the cache, so
619    // that no dangling references are left behind. We could be a bit smarter if
620    // we kept a dependency graph. It's probably not worth the complexity.
621    for (PtrSetTy::iterator I=SeenVals.begin(), E=SeenVals.end(); I != E; ++I) {
622      CacheMapTy::iterator CacheIt = CacheMap.find(*I);
623      // non-computable results can be safely cached
624      if (CacheIt != CacheMap.end() && anyKnown(CacheIt->second))
625        CacheMap.erase(CacheIt);
626    }
627  }
628
629  SeenVals.clear();
630  return Result;
631}
632
633SizeOffsetEvalType ObjectSizeOffsetEvaluator::compute_(Value *V) {
634  ObjectSizeOffsetVisitor Visitor(TD, TLI, Context);
635  SizeOffsetType Const = Visitor.compute(V);
636  if (Visitor.bothKnown(Const))
637    return std::make_pair(ConstantInt::get(Context, Const.first),
638                          ConstantInt::get(Context, Const.second));
639
640  V = V->stripPointerCasts();
641
642  // check cache
643  CacheMapTy::iterator CacheIt = CacheMap.find(V);
644  if (CacheIt != CacheMap.end())
645    return CacheIt->second;
646
647  // always generate code immediately before the instruction being
648  // processed, so that the generated code dominates the same BBs
649  Instruction *PrevInsertPoint = Builder.GetInsertPoint();
650  if (Instruction *I = dyn_cast<Instruction>(V))
651    Builder.SetInsertPoint(I);
652
653  // record the pointers that were handled in this run, so that they can be
654  // cleaned later if something fails
655  SeenVals.insert(V);
656
657  // now compute the size and offset
658  SizeOffsetEvalType Result;
659  if (GEPOperator *GEP = dyn_cast<GEPOperator>(V)) {
660    Result = visitGEPOperator(*GEP);
661  } else if (Instruction *I = dyn_cast<Instruction>(V)) {
662    Result = visit(*I);
663  } else if (isa<Argument>(V) ||
664             (isa<ConstantExpr>(V) &&
665              cast<ConstantExpr>(V)->getOpcode() == Instruction::IntToPtr) ||
666             isa<GlobalAlias>(V) ||
667             isa<GlobalVariable>(V)) {
668    // ignore values where we cannot do more than what ObjectSizeVisitor can
669    Result = unknown();
670  } else {
671    DEBUG(dbgs() << "ObjectSizeOffsetEvaluator::compute() unhandled value: "
672          << *V << '\n');
673    Result = unknown();
674  }
675
676  if (PrevInsertPoint)
677    Builder.SetInsertPoint(PrevInsertPoint);
678
679  // Don't reuse CacheIt since it may be invalid at this point.
680  CacheMap[V] = Result;
681  return Result;
682}
683
684SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitAllocaInst(AllocaInst &I) {
685  if (!I.getAllocatedType()->isSized())
686    return unknown();
687
688  // must be a VLA
689  assert(I.isArrayAllocation());
690  Value *ArraySize = I.getArraySize();
691  Value *Size = ConstantInt::get(ArraySize->getType(),
692                                 TD->getTypeAllocSize(I.getAllocatedType()));
693  Size = Builder.CreateMul(Size, ArraySize);
694  return std::make_pair(Size, Zero);
695}
696
697SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitCallSite(CallSite CS) {
698  const AllocFnsTy *FnData = getAllocationData(CS.getInstruction(), AnyAlloc,
699                                               TLI);
700  if (!FnData)
701    return unknown();
702
703  // handle strdup-like functions separately
704  if (FnData->AllocTy == StrDupLike) {
705    // TODO
706    return unknown();
707  }
708
709  Value *FirstArg = CS.getArgument(FnData->FstParam);
710  FirstArg = Builder.CreateZExt(FirstArg, IntTy);
711  if (FnData->SndParam < 0)
712    return std::make_pair(FirstArg, Zero);
713
714  Value *SecondArg = CS.getArgument(FnData->SndParam);
715  SecondArg = Builder.CreateZExt(SecondArg, IntTy);
716  Value *Size = Builder.CreateMul(FirstArg, SecondArg);
717  return std::make_pair(Size, Zero);
718
719  // TODO: handle more standard functions (+ wchar cousins):
720  // - strdup / strndup
721  // - strcpy / strncpy
722  // - strcat / strncat
723  // - memcpy / memmove
724  // - strcat / strncat
725  // - memset
726}
727
728SizeOffsetEvalType
729ObjectSizeOffsetEvaluator::visitExtractElementInst(ExtractElementInst&) {
730  return unknown();
731}
732
733SizeOffsetEvalType
734ObjectSizeOffsetEvaluator::visitExtractValueInst(ExtractValueInst&) {
735  return unknown();
736}
737
738SizeOffsetEvalType
739ObjectSizeOffsetEvaluator::visitGEPOperator(GEPOperator &GEP) {
740  SizeOffsetEvalType PtrData = compute_(GEP.getPointerOperand());
741  if (!bothKnown(PtrData))
742    return unknown();
743
744  Value *Offset = EmitGEPOffset(&Builder, *TD, &GEP, /*NoAssumptions=*/true);
745  Offset = Builder.CreateAdd(PtrData.second, Offset);
746  return std::make_pair(PtrData.first, Offset);
747}
748
749SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitIntToPtrInst(IntToPtrInst&) {
750  // clueless
751  return unknown();
752}
753
754SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitLoadInst(LoadInst&) {
755  return unknown();
756}
757
758SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitPHINode(PHINode &PHI) {
759  // create 2 PHIs: one for size and another for offset
760  PHINode *SizePHI   = Builder.CreatePHI(IntTy, PHI.getNumIncomingValues());
761  PHINode *OffsetPHI = Builder.CreatePHI(IntTy, PHI.getNumIncomingValues());
762
763  // insert right away in the cache to handle recursive PHIs
764  CacheMap[&PHI] = std::make_pair(SizePHI, OffsetPHI);
765
766  // compute offset/size for each PHI incoming pointer
767  for (unsigned i = 0, e = PHI.getNumIncomingValues(); i != e; ++i) {
768    Builder.SetInsertPoint(PHI.getIncomingBlock(i)->getFirstInsertionPt());
769    SizeOffsetEvalType EdgeData = compute_(PHI.getIncomingValue(i));
770
771    if (!bothKnown(EdgeData)) {
772      OffsetPHI->replaceAllUsesWith(UndefValue::get(IntTy));
773      OffsetPHI->eraseFromParent();
774      SizePHI->replaceAllUsesWith(UndefValue::get(IntTy));
775      SizePHI->eraseFromParent();
776      return unknown();
777    }
778    SizePHI->addIncoming(EdgeData.first, PHI.getIncomingBlock(i));
779    OffsetPHI->addIncoming(EdgeData.second, PHI.getIncomingBlock(i));
780  }
781
782  Value *Size = SizePHI, *Offset = OffsetPHI, *Tmp;
783  if ((Tmp = SizePHI->hasConstantValue())) {
784    Size = Tmp;
785    SizePHI->replaceAllUsesWith(Size);
786    SizePHI->eraseFromParent();
787  }
788  if ((Tmp = OffsetPHI->hasConstantValue())) {
789    Offset = Tmp;
790    OffsetPHI->replaceAllUsesWith(Offset);
791    OffsetPHI->eraseFromParent();
792  }
793  return std::make_pair(Size, Offset);
794}
795
796SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitSelectInst(SelectInst &I) {
797  SizeOffsetEvalType TrueSide  = compute_(I.getTrueValue());
798  SizeOffsetEvalType FalseSide = compute_(I.getFalseValue());
799
800  if (!bothKnown(TrueSide) || !bothKnown(FalseSide))
801    return unknown();
802  if (TrueSide == FalseSide)
803    return TrueSide;
804
805  Value *Size = Builder.CreateSelect(I.getCondition(), TrueSide.first,
806                                     FalseSide.first);
807  Value *Offset = Builder.CreateSelect(I.getCondition(), TrueSide.second,
808                                       FalseSide.second);
809  return std::make_pair(Size, Offset);
810}
811
812SizeOffsetEvalType ObjectSizeOffsetEvaluator::visitInstruction(Instruction &I) {
813  DEBUG(dbgs() << "ObjectSizeOffsetEvaluator unknown instruction:" << I <<'\n');
814  return unknown();
815}
816