RangeConstraintManager.cpp revision 58fc86d68d53eb6c47cc34974b6f37627a5f386c 
1//== RangeConstraintManager.cpp - Manage range constraints.------*- C++ -*--==//
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 file defines RangeConstraintManager, a class that tracks simple
11//  equality and inequality constraints on symbolic values of ProgramState.
12//
13//===----------------------------------------------------------------------===//
14
15#include "SimpleConstraintManager.h"
16#include "clang/StaticAnalyzer/Core/PathSensitive/APSIntType.h"
17#include "clang/StaticAnalyzer/Core/PathSensitive/ProgramState.h"
18#include "clang/StaticAnalyzer/Core/PathSensitive/ProgramStateTrait.h"
19#include "llvm/Support/Debug.h"
20#include "llvm/ADT/FoldingSet.h"
21#include "llvm/ADT/ImmutableSet.h"
22#include "llvm/Support/raw_ostream.h"
23
24using namespace clang;
25using namespace ento;
26
27namespace { class ConstraintRange {}; }
28static int ConstraintRangeIndex = 0;
29
30/// A Range represents the closed range [from, to].  The caller must
31/// guarantee that from <= to.  Note that Range is immutable, so as not
32/// to subvert RangeSet's immutability.
33namespace {
34class Range : public std::pair<const llvm::APSInt*,
35                                                const llvm::APSInt*> {
36public:
37  Range(const llvm::APSInt &from, const llvm::APSInt &to)
38    : std::pair<const llvm::APSInt*, const llvm::APSInt*>(&from, &to) {
39    assert(from <= to);
40  }
41  bool Includes(const llvm::APSInt &v) const {
42    return *first <= v && v <= *second;
43  }
44  const llvm::APSInt &From() const {
45    return *first;
46  }
47  const llvm::APSInt &To() const {
48    return *second;
49  }
50  const llvm::APSInt *getConcreteValue() const {
51    return &From() == &To() ? &From() : NULL;
52  }
53
54  void Profile(llvm::FoldingSetNodeID &ID) const {
55    ID.AddPointer(&From());
56    ID.AddPointer(&To());
57  }
58};
59
60
61class RangeTrait : public llvm::ImutContainerInfo<Range> {
62public:
63  // When comparing if one Range is less than another, we should compare
64  // the actual APSInt values instead of their pointers.  This keeps the order
65  // consistent (instead of comparing by pointer values) and can potentially
66  // be used to speed up some of the operations in RangeSet.
67  static inline bool isLess(key_type_ref lhs, key_type_ref rhs) {
68    return *lhs.first < *rhs.first || (!(*rhs.first < *lhs.first) &&
69                                       *lhs.second < *rhs.second);
70  }
71};
72
73/// RangeSet contains a set of ranges. If the set is empty, then
74///  there the value of a symbol is overly constrained and there are no
75///  possible values for that symbol.
76class RangeSet {
77  typedef llvm::ImmutableSet<Range, RangeTrait> PrimRangeSet;
78  PrimRangeSet ranges; // no need to make const, since it is an
79                       // ImmutableSet - this allows default operator=
80                       // to work.
81public:
82  typedef PrimRangeSet::Factory Factory;
83  typedef PrimRangeSet::iterator iterator;
84
85  RangeSet(PrimRangeSet RS) : ranges(RS) {}
86
87  iterator begin() const { return ranges.begin(); }
88  iterator end() const { return ranges.end(); }
89
90  bool isEmpty() const { return ranges.isEmpty(); }
91
92  /// Construct a new RangeSet representing '{ [from, to] }'.
93  RangeSet(Factory &F, const llvm::APSInt &from, const llvm::APSInt &to)
94    : ranges(F.add(F.getEmptySet(), Range(from, to))) {}
95
96  /// Profile - Generates a hash profile of this RangeSet for use
97  ///  by FoldingSet.
98  void Profile(llvm::FoldingSetNodeID &ID) const { ranges.Profile(ID); }
99
100  /// getConcreteValue - If a symbol is contrained to equal a specific integer
101  ///  constant then this method returns that value.  Otherwise, it returns
102  ///  NULL.
103  const llvm::APSInt* getConcreteValue() const {
104    return ranges.isSingleton() ? ranges.begin()->getConcreteValue() : 0;
105  }
106
107private:
108  void IntersectInRange(BasicValueFactory &BV, Factory &F,
109                        const llvm::APSInt &Lower,
110                        const llvm::APSInt &Upper,
111                        PrimRangeSet &newRanges,
112                        PrimRangeSet::iterator &i,
113                        PrimRangeSet::iterator &e) const {
114    // There are six cases for each range R in the set:
115    //   1. R is entirely before the intersection range.
116    //   2. R is entirely after the intersection range.
117    //   3. R contains the entire intersection range.
118    //   4. R starts before the intersection range and ends in the middle.
119    //   5. R starts in the middle of the intersection range and ends after it.
120    //   6. R is entirely contained in the intersection range.
121    // These correspond to each of the conditions below.
122    for (/* i = begin(), e = end() */; i != e; ++i) {
123      if (i->To() < Lower) {
124        continue;
125      }
126      if (i->From() > Upper) {
127        break;
128      }
129
130      if (i->Includes(Lower)) {
131        if (i->Includes(Upper)) {
132          newRanges = F.add(newRanges, Range(BV.getValue(Lower),
133                                             BV.getValue(Upper)));
134          break;
135        } else
136          newRanges = F.add(newRanges, Range(BV.getValue(Lower), i->To()));
137      } else {
138        if (i->Includes(Upper)) {
139          newRanges = F.add(newRanges, Range(i->From(), BV.getValue(Upper)));
140          break;
141        } else
142          newRanges = F.add(newRanges, *i);
143      }
144    }
145  }
146
147  const llvm::APSInt &getMinValue() const {
148    assert(!isEmpty());
149    return ranges.begin()->From();
150  }
151
152  bool pin(llvm::APSInt &Lower, llvm::APSInt &Upper) const {
153    // This function has nine cases, the cartesian product of range-testing
154    // both the upper and lower bounds against the symbol's type.
155    // Each case requires a different pinning operation.
156    // The function returns false if the described range is entirely outside
157    // the range of values for the associated symbol.
158    APSIntType Type(getMinValue());
159    APSIntType::RangeTestResultKind LowerTest = Type.testInRange(Lower);
160    APSIntType::RangeTestResultKind UpperTest = Type.testInRange(Upper);
161
162    switch (LowerTest) {
163    case APSIntType::RTR_Below:
164      switch (UpperTest) {
165      case APSIntType::RTR_Below:
166        // The entire range is outside the symbol's set of possible values.
167        // If this is a conventionally-ordered range, the state is infeasible.
168        if (Lower < Upper)
169          return false;
170
171        // However, if the range wraps around, it spans all possible values.
172        Lower = Type.getMinValue();
173        Upper = Type.getMaxValue();
174        break;
175      case APSIntType::RTR_Within:
176        // The range starts below what's possible but ends within it. Pin.
177        Lower = Type.getMinValue();
178        Type.apply(Upper);
179        break;
180      case APSIntType::RTR_Above:
181        // The range spans all possible values for the symbol. Pin.
182        Lower = Type.getMinValue();
183        Upper = Type.getMaxValue();
184        break;
185      }
186      break;
187    case APSIntType::RTR_Within:
188      switch (UpperTest) {
189      case APSIntType::RTR_Below:
190        // The range wraps around, but all lower values are not possible.
191        Type.apply(Lower);
192        Upper = Type.getMaxValue();
193        break;
194      case APSIntType::RTR_Within:
195        // The range may or may not wrap around, but both limits are valid.
196        Type.apply(Lower);
197        Type.apply(Upper);
198        break;
199      case APSIntType::RTR_Above:
200        // The range starts within what's possible but ends above it. Pin.
201        Type.apply(Lower);
202        Upper = Type.getMaxValue();
203        break;
204      }
205      break;
206    case APSIntType::RTR_Above:
207      switch (UpperTest) {
208      case APSIntType::RTR_Below:
209        // The range wraps but is outside the symbol's set of possible values.
210        return false;
211      case APSIntType::RTR_Within:
212        // The range starts above what's possible but ends within it (wrap).
213        Lower = Type.getMinValue();
214        Type.apply(Upper);
215        break;
216      case APSIntType::RTR_Above:
217        // The entire range is outside the symbol's set of possible values.
218        // If this is a conventionally-ordered range, the state is infeasible.
219        if (Lower < Upper)
220          return false;
221
222        // However, if the range wraps around, it spans all possible values.
223        Lower = Type.getMinValue();
224        Upper = Type.getMaxValue();
225        break;
226      }
227      break;
228    }
229
230    return true;
231  }
232
233public:
234  // Returns a set containing the values in the receiving set, intersected with
235  // the closed range [Lower, Upper]. Unlike the Range type, this range uses
236  // modular arithmetic, corresponding to the common treatment of C integer
237  // overflow. Thus, if the Lower bound is greater than the Upper bound, the
238  // range is taken to wrap around. This is equivalent to taking the
239  // intersection with the two ranges [Min, Upper] and [Lower, Max],
240  // or, alternatively, /removing/ all integers between Upper and Lower.
241  RangeSet Intersect(BasicValueFactory &BV, Factory &F,
242                     llvm::APSInt Lower, llvm::APSInt Upper) const {
243    if (!pin(Lower, Upper))
244      return F.getEmptySet();
245
246    PrimRangeSet newRanges = F.getEmptySet();
247
248    PrimRangeSet::iterator i = begin(), e = end();
249    if (Lower <= Upper)
250      IntersectInRange(BV, F, Lower, Upper, newRanges, i, e);
251    else {
252      // The order of the next two statements is important!
253      // IntersectInRange() does not reset the iteration state for i and e.
254      // Therefore, the lower range most be handled first.
255      IntersectInRange(BV, F, BV.getMinValue(Upper), Upper, newRanges, i, e);
256      IntersectInRange(BV, F, Lower, BV.getMaxValue(Lower), newRanges, i, e);
257    }
258
259    return newRanges;
260  }
261
262  void print(raw_ostream &os) const {
263    bool isFirst = true;
264    os << "{ ";
265    for (iterator i = begin(), e = end(); i != e; ++i) {
266      if (isFirst)
267        isFirst = false;
268      else
269        os << ", ";
270
271      os << '[' << i->From().toString(10) << ", " << i->To().toString(10)
272         << ']';
273    }
274    os << " }";
275  }
276
277  bool operator==(const RangeSet &other) const {
278    return ranges == other.ranges;
279  }
280};
281} // end anonymous namespace
282
283typedef llvm::ImmutableMap<SymbolRef,RangeSet> ConstraintRangeTy;
284
285namespace clang {
286namespace ento {
287template<>
288struct ProgramStateTrait<ConstraintRange>
289  : public ProgramStatePartialTrait<ConstraintRangeTy> {
290  static inline void *GDMIndex() { return &ConstraintRangeIndex; }
291};
292}
293}
294
295namespace {
296class RangeConstraintManager : public SimpleConstraintManager{
297  RangeSet GetRange(ProgramStateRef state, SymbolRef sym);
298public:
299  RangeConstraintManager(SubEngine &subengine, BasicValueFactory &BVF)
300    : SimpleConstraintManager(subengine, BVF) {}
301
302  ProgramStateRef assumeSymNE(ProgramStateRef state, SymbolRef sym,
303                             const llvm::APSInt& Int,
304                             const llvm::APSInt& Adjustment);
305
306  ProgramStateRef assumeSymEQ(ProgramStateRef state, SymbolRef sym,
307                             const llvm::APSInt& Int,
308                             const llvm::APSInt& Adjustment);
309
310  ProgramStateRef assumeSymLT(ProgramStateRef state, SymbolRef sym,
311                             const llvm::APSInt& Int,
312                             const llvm::APSInt& Adjustment);
313
314  ProgramStateRef assumeSymGT(ProgramStateRef state, SymbolRef sym,
315                             const llvm::APSInt& Int,
316                             const llvm::APSInt& Adjustment);
317
318  ProgramStateRef assumeSymGE(ProgramStateRef state, SymbolRef sym,
319                             const llvm::APSInt& Int,
320                             const llvm::APSInt& Adjustment);
321
322  ProgramStateRef assumeSymLE(ProgramStateRef state, SymbolRef sym,
323                             const llvm::APSInt& Int,
324                             const llvm::APSInt& Adjustment);
325
326  const llvm::APSInt* getSymVal(ProgramStateRef St, SymbolRef sym) const;
327
328  // FIXME: Refactor into SimpleConstraintManager?
329  bool isEqual(ProgramStateRef St, SymbolRef sym, const llvm::APSInt& V) const {
330    const llvm::APSInt *i = getSymVal(St, sym);
331    return i ? *i == V : false;
332  }
333
334  ProgramStateRef removeDeadBindings(ProgramStateRef St, SymbolReaper& SymReaper);
335
336  void print(ProgramStateRef St, raw_ostream &Out,
337             const char* nl, const char *sep);
338
339private:
340  RangeSet::Factory F;
341};
342
343} // end anonymous namespace
344
345ConstraintManager *
346ento::CreateRangeConstraintManager(ProgramStateManager &StMgr, SubEngine &Eng) {
347  return new RangeConstraintManager(Eng, StMgr.getBasicVals());
348}
349
350const llvm::APSInt* RangeConstraintManager::getSymVal(ProgramStateRef St,
351                                                      SymbolRef sym) const {
352  const ConstraintRangeTy::data_type *T = St->get<ConstraintRange>(sym);
353  return T ? T->getConcreteValue() : NULL;
354}
355
356/// Scan all symbols referenced by the constraints. If the symbol is not alive
357/// as marked in LSymbols, mark it as dead in DSymbols.
358ProgramStateRef
359RangeConstraintManager::removeDeadBindings(ProgramStateRef state,
360                                           SymbolReaper& SymReaper) {
361
362  ConstraintRangeTy CR = state->get<ConstraintRange>();
363  ConstraintRangeTy::Factory& CRFactory = state->get_context<ConstraintRange>();
364
365  for (ConstraintRangeTy::iterator I = CR.begin(), E = CR.end(); I != E; ++I) {
366    SymbolRef sym = I.getKey();
367    if (SymReaper.maybeDead(sym))
368      CR = CRFactory.remove(CR, sym);
369  }
370
371  return state->set<ConstraintRange>(CR);
372}
373
374RangeSet
375RangeConstraintManager::GetRange(ProgramStateRef state, SymbolRef sym) {
376  if (ConstraintRangeTy::data_type* V = state->get<ConstraintRange>(sym))
377    return *V;
378
379  // Lazily generate a new RangeSet representing all possible values for the
380  // given symbol type.
381  BasicValueFactory &BV = getBasicVals();
382  QualType T = sym->getType(BV.getContext());
383  return RangeSet(F, BV.getMinValue(T), BV.getMaxValue(T));
384}
385
386//===------------------------------------------------------------------------===
387// assumeSymX methods: public interface for RangeConstraintManager.
388//===------------------------------------------------------------------------===/
389
390// The syntax for ranges below is mathematical, using [x, y] for closed ranges
391// and (x, y) for open ranges. These ranges are modular, corresponding with
392// a common treatment of C integer overflow. This means that these methods
393// do not have to worry about overflow; RangeSet::Intersect can handle such a
394// "wraparound" range.
395// As an example, the range [UINT_MAX-1, 3) contains five values: UINT_MAX-1,
396// UINT_MAX, 0, 1, and 2.
397
398ProgramStateRef
399RangeConstraintManager::assumeSymNE(ProgramStateRef St, SymbolRef Sym,
400                                    const llvm::APSInt &Int,
401                                    const llvm::APSInt &Adjustment) {
402  // Before we do any real work, see if the value can even show up.
403  APSIntType AdjustmentType(Adjustment);
404  if (AdjustmentType.testInRange(Int) != APSIntType::RTR_Within)
405    return St;
406
407  llvm::APSInt Lower = AdjustmentType.convert(Int) - Adjustment;
408  llvm::APSInt Upper = Lower;
409  --Lower;
410  ++Upper;
411
412  // [Int-Adjustment+1, Int-Adjustment-1]
413  // Notice that the lower bound is greater than the upper bound.
414  RangeSet New = GetRange(St, Sym).Intersect(getBasicVals(), F, Upper, Lower);
415  return New.isEmpty() ? NULL : St->set<ConstraintRange>(Sym, New);
416}
417
418ProgramStateRef
419RangeConstraintManager::assumeSymEQ(ProgramStateRef St, SymbolRef Sym,
420                                    const llvm::APSInt &Int,
421                                    const llvm::APSInt &Adjustment) {
422  // Before we do any real work, see if the value can even show up.
423  APSIntType AdjustmentType(Adjustment);
424  if (AdjustmentType.testInRange(Int) != APSIntType::RTR_Within)
425    return NULL;
426
427  // [Int-Adjustment, Int-Adjustment]
428  llvm::APSInt AdjInt = AdjustmentType.convert(Int) - Adjustment;
429  RangeSet New = GetRange(St, Sym).Intersect(getBasicVals(), F, AdjInt, AdjInt);
430  return New.isEmpty() ? NULL : St->set<ConstraintRange>(Sym, New);
431}
432
433ProgramStateRef
434RangeConstraintManager::assumeSymLT(ProgramStateRef St, SymbolRef Sym,
435                                    const llvm::APSInt &Int,
436                                    const llvm::APSInt &Adjustment) {
437  // Before we do any real work, see if the value can even show up.
438  APSIntType AdjustmentType(Adjustment);
439  switch (AdjustmentType.testInRange(Int)) {
440  case APSIntType::RTR_Below:
441    return NULL;
442  case APSIntType::RTR_Within:
443    break;
444  case APSIntType::RTR_Above:
445    return St;
446  }
447
448  // Special case for Int == Min. This is always false.
449  llvm::APSInt ComparisonVal = AdjustmentType.convert(Int);
450  llvm::APSInt Min = AdjustmentType.getMinValue();
451  if (ComparisonVal == Min)
452    return NULL;
453
454  llvm::APSInt Lower = Min-Adjustment;
455  llvm::APSInt Upper = ComparisonVal-Adjustment;
456  --Upper;
457
458  RangeSet New = GetRange(St, Sym).Intersect(getBasicVals(), F, Lower, Upper);
459  return New.isEmpty() ? NULL : St->set<ConstraintRange>(Sym, New);
460}
461
462ProgramStateRef
463RangeConstraintManager::assumeSymGT(ProgramStateRef St, SymbolRef Sym,
464                                    const llvm::APSInt &Int,
465                                    const llvm::APSInt &Adjustment) {
466  // Before we do any real work, see if the value can even show up.
467  APSIntType AdjustmentType(Adjustment);
468  switch (AdjustmentType.testInRange(Int)) {
469  case APSIntType::RTR_Below:
470    return St;
471  case APSIntType::RTR_Within:
472    break;
473  case APSIntType::RTR_Above:
474    return NULL;
475  }
476
477  // Special case for Int == Max. This is always false.
478  llvm::APSInt ComparisonVal = AdjustmentType.convert(Int);
479  llvm::APSInt Max = AdjustmentType.getMaxValue();
480  if (ComparisonVal == Max)
481    return NULL;
482
483  llvm::APSInt Lower = ComparisonVal-Adjustment;
484  llvm::APSInt Upper = Max-Adjustment;
485  ++Lower;
486
487  RangeSet New = GetRange(St, Sym).Intersect(getBasicVals(), F, Lower, Upper);
488  return New.isEmpty() ? NULL : St->set<ConstraintRange>(Sym, New);
489}
490
491ProgramStateRef
492RangeConstraintManager::assumeSymGE(ProgramStateRef St, SymbolRef Sym,
493                                    const llvm::APSInt &Int,
494                                    const llvm::APSInt &Adjustment) {
495  // Before we do any real work, see if the value can even show up.
496  APSIntType AdjustmentType(Adjustment);
497  switch (AdjustmentType.testInRange(Int)) {
498  case APSIntType::RTR_Below:
499    return St;
500  case APSIntType::RTR_Within:
501    break;
502  case APSIntType::RTR_Above:
503    return NULL;
504  }
505
506  // Special case for Int == Min. This is always feasible.
507  llvm::APSInt ComparisonVal = AdjustmentType.convert(Int);
508  llvm::APSInt Min = AdjustmentType.getMinValue();
509  if (ComparisonVal == Min)
510    return St;
511
512  llvm::APSInt Max = AdjustmentType.getMaxValue();
513  llvm::APSInt Lower = ComparisonVal-Adjustment;
514  llvm::APSInt Upper = Max-Adjustment;
515
516  RangeSet New = GetRange(St, Sym).Intersect(getBasicVals(), F, Lower, Upper);
517  return New.isEmpty() ? NULL : St->set<ConstraintRange>(Sym, New);
518}
519
520ProgramStateRef
521RangeConstraintManager::assumeSymLE(ProgramStateRef St, SymbolRef Sym,
522                                    const llvm::APSInt &Int,
523                                    const llvm::APSInt &Adjustment) {
524  // Before we do any real work, see if the value can even show up.
525  APSIntType AdjustmentType(Adjustment);
526  switch (AdjustmentType.testInRange(Int)) {
527  case APSIntType::RTR_Below:
528    return NULL;
529  case APSIntType::RTR_Within:
530    break;
531  case APSIntType::RTR_Above:
532    return St;
533  }
534
535  // Special case for Int == Max. This is always feasible.
536  llvm::APSInt ComparisonVal = AdjustmentType.convert(Int);
537  llvm::APSInt Max = AdjustmentType.getMaxValue();
538  if (ComparisonVal == Max)
539    return St;
540
541  llvm::APSInt Min = AdjustmentType.getMinValue();
542  llvm::APSInt Lower = Min-Adjustment;
543  llvm::APSInt Upper = ComparisonVal-Adjustment;
544
545  RangeSet New = GetRange(St, Sym).Intersect(getBasicVals(), F, Lower, Upper);
546  return New.isEmpty() ? NULL : St->set<ConstraintRange>(Sym, New);
547}
548
549//===------------------------------------------------------------------------===
550// Pretty-printing.
551//===------------------------------------------------------------------------===/
552
553void RangeConstraintManager::print(ProgramStateRef St, raw_ostream &Out,
554                                   const char* nl, const char *sep) {
555
556  ConstraintRangeTy Ranges = St->get<ConstraintRange>();
557
558  if (Ranges.isEmpty()) {
559    Out << nl << sep << "Ranges are empty." << nl;
560    return;
561  }
562
563  Out << nl << sep << "Ranges of symbol values:";
564  for (ConstraintRangeTy::iterator I=Ranges.begin(), E=Ranges.end(); I!=E; ++I){
565    Out << nl << ' ' << I.getKey() << " : ";
566    I.getData().print(Out);
567  }
568  Out << nl;
569}
570