1 //== RegionStore.cpp - Field-sensitive store model --------------*- C++ -*--==//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 //
9 // This file defines a basic region store model. In this model, we do have field
10 // sensitivity. But we assume nothing about the heap shape. So recursive data
11 // structures are largely ignored. Basically we do 1-limiting analysis.
12 // Parameter pointers are assumed with no aliasing. Pointee objects of
13 // parameters are created lazily.
14 //
15 //===----------------------------------------------------------------------===//
16 
17 #include "clang/AST/Attr.h"
18 #include "clang/AST/CharUnits.h"
19 #include "clang/ASTMatchers/ASTMatchFinder.h"
20 #include "clang/Analysis/Analyses/LiveVariables.h"
21 #include "clang/Analysis/AnalysisDeclContext.h"
22 #include "clang/Basic/JsonSupport.h"
23 #include "clang/Basic/TargetInfo.h"
24 #include "clang/StaticAnalyzer/Core/PathSensitive/AnalysisManager.h"
25 #include "clang/StaticAnalyzer/Core/PathSensitive/CallEvent.h"
26 #include "clang/StaticAnalyzer/Core/PathSensitive/ExprEngine.h"
27 #include "clang/StaticAnalyzer/Core/PathSensitive/MemRegion.h"
28 #include "clang/StaticAnalyzer/Core/PathSensitive/ProgramState.h"
29 #include "clang/StaticAnalyzer/Core/PathSensitive/ProgramStateTrait.h"
30 #include "llvm/ADT/ImmutableMap.h"
31 #include "llvm/ADT/Optional.h"
32 #include "llvm/Support/raw_ostream.h"
33 #include <utility>
34 
35 using namespace clang;
36 using namespace ento;
37 
38 //===----------------------------------------------------------------------===//
39 // Representation of binding keys.
40 //===----------------------------------------------------------------------===//
41 
42 namespace {
43 class BindingKey {
44 public:
45   enum Kind { Default = 0x0, Direct = 0x1 };
46 private:
47   enum { Symbolic = 0x2 };
48 
49   llvm::PointerIntPair<const MemRegion *, 2> P;
50   uint64_t Data;
51 
52   /// Create a key for a binding to region \p r, which has a symbolic offset
53   /// from region \p Base.
54   explicit BindingKey(const SubRegion *r, const SubRegion *Base, Kind k)
55     : P(r, k | Symbolic), Data(reinterpret_cast<uintptr_t>(Base)) {
56     assert(r && Base && "Must have known regions.");
57     assert(getConcreteOffsetRegion() == Base && "Failed to store base region");
58   }
59 
60   /// Create a key for a binding at \p offset from base region \p r.
61   explicit BindingKey(const MemRegion *r, uint64_t offset, Kind k)
62     : P(r, k), Data(offset) {
63     assert(r && "Must have known regions.");
64     assert(getOffset() == offset && "Failed to store offset");
65     assert((r == r->getBaseRegion() || isa<ObjCIvarRegion>(r) ||
66             isa <CXXDerivedObjectRegion>(r)) &&
67            "Not a base");
68   }
69 public:
70 
71   bool isDirect() const { return P.getInt() & Direct; }
72   bool hasSymbolicOffset() const { return P.getInt() & Symbolic; }
73 
74   const MemRegion *getRegion() const { return P.getPointer(); }
75   uint64_t getOffset() const {
76     assert(!hasSymbolicOffset());
77     return Data;
78   }
79 
80   const SubRegion *getConcreteOffsetRegion() const {
81     assert(hasSymbolicOffset());
82     return reinterpret_cast<const SubRegion *>(static_cast<uintptr_t>(Data));
83   }
84 
85   const MemRegion *getBaseRegion() const {
86     if (hasSymbolicOffset())
87       return getConcreteOffsetRegion()->getBaseRegion();
88     return getRegion()->getBaseRegion();
89   }
90 
91   void Profile(llvm::FoldingSetNodeID& ID) const {
92     ID.AddPointer(P.getOpaqueValue());
93     ID.AddInteger(Data);
94   }
95 
96   static BindingKey Make(const MemRegion *R, Kind k);
97 
98   bool operator<(const BindingKey &X) const {
99     if (P.getOpaqueValue() < X.P.getOpaqueValue())
100       return true;
101     if (P.getOpaqueValue() > X.P.getOpaqueValue())
102       return false;
103     return Data < X.Data;
104   }
105 
106   bool operator==(const BindingKey &X) const {
107     return P.getOpaqueValue() == X.P.getOpaqueValue() &&
108            Data == X.Data;
109   }
110 
111   LLVM_DUMP_METHOD void dump() const;
112 };
113 } // end anonymous namespace
114 
115 BindingKey BindingKey::Make(const MemRegion *R, Kind k) {
116   const RegionOffset &RO = R->getAsOffset();
117   if (RO.hasSymbolicOffset())
118     return BindingKey(cast<SubRegion>(R), cast<SubRegion>(RO.getRegion()), k);
119 
120   return BindingKey(RO.getRegion(), RO.getOffset(), k);
121 }
122 
123 namespace llvm {
124 static inline raw_ostream &operator<<(raw_ostream &Out, BindingKey K) {
125   Out << "\"kind\": \"" << (K.isDirect() ? "Direct" : "Default")
126       << "\", \"offset\": ";
127 
128   if (!K.hasSymbolicOffset())
129     Out << K.getOffset();
130   else
131     Out << "null";
132 
133   return Out;
134 }
135 
136 } // namespace llvm
137 
138 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
139 void BindingKey::dump() const { llvm::errs() << *this; }
140 #endif
141 
142 //===----------------------------------------------------------------------===//
143 // Actual Store type.
144 //===----------------------------------------------------------------------===//
145 
146 typedef llvm::ImmutableMap<BindingKey, SVal>    ClusterBindings;
147 typedef llvm::ImmutableMapRef<BindingKey, SVal> ClusterBindingsRef;
148 typedef std::pair<BindingKey, SVal> BindingPair;
149 
150 typedef llvm::ImmutableMap<const MemRegion *, ClusterBindings>
151         RegionBindings;
152 
153 namespace {
154 class RegionBindingsRef : public llvm::ImmutableMapRef<const MemRegion *,
155                                  ClusterBindings> {
156   ClusterBindings::Factory *CBFactory;
157 
158   // This flag indicates whether the current bindings are within the analysis
159   // that has started from main(). It affects how we perform loads from
160   // global variables that have initializers: if we have observed the
161   // program execution from the start and we know that these variables
162   // have not been overwritten yet, we can be sure that their initializers
163   // are still relevant. This flag never gets changed when the bindings are
164   // updated, so it could potentially be moved into RegionStoreManager
165   // (as if it's the same bindings but a different loading procedure)
166   // however that would have made the manager needlessly stateful.
167   bool IsMainAnalysis;
168 
169 public:
170   typedef llvm::ImmutableMapRef<const MemRegion *, ClusterBindings>
171           ParentTy;
172 
173   RegionBindingsRef(ClusterBindings::Factory &CBFactory,
174                     const RegionBindings::TreeTy *T,
175                     RegionBindings::TreeTy::Factory *F,
176                     bool IsMainAnalysis)
177       : llvm::ImmutableMapRef<const MemRegion *, ClusterBindings>(T, F),
178         CBFactory(&CBFactory), IsMainAnalysis(IsMainAnalysis) {}
179 
180   RegionBindingsRef(const ParentTy &P,
181                     ClusterBindings::Factory &CBFactory,
182                     bool IsMainAnalysis)
183       : llvm::ImmutableMapRef<const MemRegion *, ClusterBindings>(P),
184         CBFactory(&CBFactory), IsMainAnalysis(IsMainAnalysis) {}
185 
186   RegionBindingsRef add(key_type_ref K, data_type_ref D) const {
187     return RegionBindingsRef(static_cast<const ParentTy *>(this)->add(K, D),
188                              *CBFactory, IsMainAnalysis);
189   }
190 
191   RegionBindingsRef remove(key_type_ref K) const {
192     return RegionBindingsRef(static_cast<const ParentTy *>(this)->remove(K),
193                              *CBFactory, IsMainAnalysis);
194   }
195 
196   RegionBindingsRef addBinding(BindingKey K, SVal V) const;
197 
198   RegionBindingsRef addBinding(const MemRegion *R,
199                                BindingKey::Kind k, SVal V) const;
200 
201   const SVal *lookup(BindingKey K) const;
202   const SVal *lookup(const MemRegion *R, BindingKey::Kind k) const;
203   using llvm::ImmutableMapRef<const MemRegion *, ClusterBindings>::lookup;
204 
205   RegionBindingsRef removeBinding(BindingKey K);
206 
207   RegionBindingsRef removeBinding(const MemRegion *R,
208                                   BindingKey::Kind k);
209 
210   RegionBindingsRef removeBinding(const MemRegion *R) {
211     return removeBinding(R, BindingKey::Direct).
212            removeBinding(R, BindingKey::Default);
213   }
214 
215   Optional<SVal> getDirectBinding(const MemRegion *R) const;
216 
217   /// getDefaultBinding - Returns an SVal* representing an optional default
218   ///  binding associated with a region and its subregions.
219   Optional<SVal> getDefaultBinding(const MemRegion *R) const;
220 
221   /// Return the internal tree as a Store.
222   Store asStore() const {
223     llvm::PointerIntPair<Store, 1, bool> Ptr = {
224         asImmutableMap().getRootWithoutRetain(), IsMainAnalysis};
225     return reinterpret_cast<Store>(Ptr.getOpaqueValue());
226   }
227 
228   bool isMainAnalysis() const {
229     return IsMainAnalysis;
230   }
231 
232   void printJson(raw_ostream &Out, const char *NL = "\n",
233                  unsigned int Space = 0, bool IsDot = false) const {
234     for (iterator I = begin(); I != end(); ++I) {
235       // TODO: We might need a .printJson for I.getKey() as well.
236       Indent(Out, Space, IsDot)
237           << "{ \"cluster\": \"" << I.getKey() << "\", \"pointer\": \""
238           << (const void *)I.getKey() << "\", \"items\": [" << NL;
239 
240       ++Space;
241       const ClusterBindings &CB = I.getData();
242       for (ClusterBindings::iterator CI = CB.begin(); CI != CB.end(); ++CI) {
243         Indent(Out, Space, IsDot) << "{ " << CI.getKey() << ", \"value\": ";
244         CI.getData().printJson(Out, /*AddQuotes=*/true);
245         Out << " }";
246         if (std::next(CI) != CB.end())
247           Out << ',';
248         Out << NL;
249       }
250 
251       --Space;
252       Indent(Out, Space, IsDot) << "]}";
253       if (std::next(I) != end())
254         Out << ',';
255       Out << NL;
256     }
257   }
258 
259   LLVM_DUMP_METHOD void dump() const { printJson(llvm::errs()); }
260 };
261 } // end anonymous namespace
262 
263 typedef const RegionBindingsRef& RegionBindingsConstRef;
264 
265 Optional<SVal> RegionBindingsRef::getDirectBinding(const MemRegion *R) const {
266   return Optional<SVal>::create(lookup(R, BindingKey::Direct));
267 }
268 
269 Optional<SVal> RegionBindingsRef::getDefaultBinding(const MemRegion *R) const {
270   return Optional<SVal>::create(lookup(R, BindingKey::Default));
271 }
272 
273 RegionBindingsRef RegionBindingsRef::addBinding(BindingKey K, SVal V) const {
274   const MemRegion *Base = K.getBaseRegion();
275 
276   const ClusterBindings *ExistingCluster = lookup(Base);
277   ClusterBindings Cluster =
278       (ExistingCluster ? *ExistingCluster : CBFactory->getEmptyMap());
279 
280   ClusterBindings NewCluster = CBFactory->add(Cluster, K, V);
281   return add(Base, NewCluster);
282 }
283 
284 
285 RegionBindingsRef RegionBindingsRef::addBinding(const MemRegion *R,
286                                                 BindingKey::Kind k,
287                                                 SVal V) const {
288   return addBinding(BindingKey::Make(R, k), V);
289 }
290 
291 const SVal *RegionBindingsRef::lookup(BindingKey K) const {
292   const ClusterBindings *Cluster = lookup(K.getBaseRegion());
293   if (!Cluster)
294     return nullptr;
295   return Cluster->lookup(K);
296 }
297 
298 const SVal *RegionBindingsRef::lookup(const MemRegion *R,
299                                       BindingKey::Kind k) const {
300   return lookup(BindingKey::Make(R, k));
301 }
302 
303 RegionBindingsRef RegionBindingsRef::removeBinding(BindingKey K) {
304   const MemRegion *Base = K.getBaseRegion();
305   const ClusterBindings *Cluster = lookup(Base);
306   if (!Cluster)
307     return *this;
308 
309   ClusterBindings NewCluster = CBFactory->remove(*Cluster, K);
310   if (NewCluster.isEmpty())
311     return remove(Base);
312   return add(Base, NewCluster);
313 }
314 
315 RegionBindingsRef RegionBindingsRef::removeBinding(const MemRegion *R,
316                                                 BindingKey::Kind k){
317   return removeBinding(BindingKey::Make(R, k));
318 }
319 
320 //===----------------------------------------------------------------------===//
321 // Fine-grained control of RegionStoreManager.
322 //===----------------------------------------------------------------------===//
323 
324 namespace {
325 struct minimal_features_tag {};
326 struct maximal_features_tag {};
327 
328 class RegionStoreFeatures {
329   bool SupportsFields;
330 public:
331   RegionStoreFeatures(minimal_features_tag) :
332     SupportsFields(false) {}
333 
334   RegionStoreFeatures(maximal_features_tag) :
335     SupportsFields(true) {}
336 
337   void enableFields(bool t) { SupportsFields = t; }
338 
339   bool supportsFields() const { return SupportsFields; }
340 };
341 }
342 
343 //===----------------------------------------------------------------------===//
344 // Main RegionStore logic.
345 //===----------------------------------------------------------------------===//
346 
347 namespace {
348 class InvalidateRegionsWorker;
349 
350 class RegionStoreManager : public StoreManager {
351 public:
352   const RegionStoreFeatures Features;
353 
354   RegionBindings::Factory RBFactory;
355   mutable ClusterBindings::Factory CBFactory;
356 
357   typedef std::vector<SVal> SValListTy;
358 private:
359   typedef llvm::DenseMap<const LazyCompoundValData *,
360                          SValListTy> LazyBindingsMapTy;
361   LazyBindingsMapTy LazyBindingsMap;
362 
363   /// The largest number of fields a struct can have and still be
364   /// considered "small".
365   ///
366   /// This is currently used to decide whether or not it is worth "forcing" a
367   /// LazyCompoundVal on bind.
368   ///
369   /// This is controlled by 'region-store-small-struct-limit' option.
370   /// To disable all small-struct-dependent behavior, set the option to "0".
371   unsigned SmallStructLimit;
372 
373   /// A helper used to populate the work list with the given set of
374   /// regions.
375   void populateWorkList(InvalidateRegionsWorker &W,
376                         ArrayRef<SVal> Values,
377                         InvalidatedRegions *TopLevelRegions);
378 
379 public:
380   RegionStoreManager(ProgramStateManager& mgr, const RegionStoreFeatures &f)
381     : StoreManager(mgr), Features(f),
382       RBFactory(mgr.getAllocator()), CBFactory(mgr.getAllocator()),
383       SmallStructLimit(0) {
384     ExprEngine &Eng = StateMgr.getOwningEngine();
385     AnalyzerOptions &Options = Eng.getAnalysisManager().options;
386     SmallStructLimit = Options.RegionStoreSmallStructLimit;
387   }
388 
389 
390   /// setImplicitDefaultValue - Set the default binding for the provided
391   ///  MemRegion to the value implicitly defined for compound literals when
392   ///  the value is not specified.
393   RegionBindingsRef setImplicitDefaultValue(RegionBindingsConstRef B,
394                                             const MemRegion *R, QualType T);
395 
396   /// ArrayToPointer - Emulates the "decay" of an array to a pointer
397   ///  type.  'Array' represents the lvalue of the array being decayed
398   ///  to a pointer, and the returned SVal represents the decayed
399   ///  version of that lvalue (i.e., a pointer to the first element of
400   ///  the array).  This is called by ExprEngine when evaluating
401   ///  casts from arrays to pointers.
402   SVal ArrayToPointer(Loc Array, QualType ElementTy) override;
403 
404   /// Creates the Store that correctly represents memory contents before
405   /// the beginning of the analysis of the given top-level stack frame.
406   StoreRef getInitialStore(const LocationContext *InitLoc) override {
407     bool IsMainAnalysis = false;
408     if (const auto *FD = dyn_cast<FunctionDecl>(InitLoc->getDecl()))
409       IsMainAnalysis = FD->isMain() && !Ctx.getLangOpts().CPlusPlus;
410     return StoreRef(RegionBindingsRef(
411         RegionBindingsRef::ParentTy(RBFactory.getEmptyMap(), RBFactory),
412         CBFactory, IsMainAnalysis).asStore(), *this);
413   }
414 
415   //===-------------------------------------------------------------------===//
416   // Binding values to regions.
417   //===-------------------------------------------------------------------===//
418   RegionBindingsRef invalidateGlobalRegion(MemRegion::Kind K,
419                                            const Expr *Ex,
420                                            unsigned Count,
421                                            const LocationContext *LCtx,
422                                            RegionBindingsRef B,
423                                            InvalidatedRegions *Invalidated);
424 
425   StoreRef invalidateRegions(Store store,
426                              ArrayRef<SVal> Values,
427                              const Expr *E, unsigned Count,
428                              const LocationContext *LCtx,
429                              const CallEvent *Call,
430                              InvalidatedSymbols &IS,
431                              RegionAndSymbolInvalidationTraits &ITraits,
432                              InvalidatedRegions *Invalidated,
433                              InvalidatedRegions *InvalidatedTopLevel) override;
434 
435   bool scanReachableSymbols(Store S, const MemRegion *R,
436                             ScanReachableSymbols &Callbacks) override;
437 
438   RegionBindingsRef removeSubRegionBindings(RegionBindingsConstRef B,
439                                             const SubRegion *R);
440 
441 public: // Part of public interface to class.
442 
443   StoreRef Bind(Store store, Loc LV, SVal V) override {
444     return StoreRef(bind(getRegionBindings(store), LV, V).asStore(), *this);
445   }
446 
447   RegionBindingsRef bind(RegionBindingsConstRef B, Loc LV, SVal V);
448 
449   // BindDefaultInitial is only used to initialize a region with
450   // a default value.
451   StoreRef BindDefaultInitial(Store store, const MemRegion *R,
452                               SVal V) override {
453     RegionBindingsRef B = getRegionBindings(store);
454     // Use other APIs when you have to wipe the region that was initialized
455     // earlier.
456     assert(!(B.getDefaultBinding(R) || B.getDirectBinding(R)) &&
457            "Double initialization!");
458     B = B.addBinding(BindingKey::Make(R, BindingKey::Default), V);
459     return StoreRef(B.asImmutableMap().getRootWithoutRetain(), *this);
460   }
461 
462   // BindDefaultZero is used for zeroing constructors that may accidentally
463   // overwrite existing bindings.
464   StoreRef BindDefaultZero(Store store, const MemRegion *R) override {
465     // FIXME: The offsets of empty bases can be tricky because of
466     // of the so called "empty base class optimization".
467     // If a base class has been optimized out
468     // we should not try to create a binding, otherwise we should.
469     // Unfortunately, at the moment ASTRecordLayout doesn't expose
470     // the actual sizes of the empty bases
471     // and trying to infer them from offsets/alignments
472     // seems to be error-prone and non-trivial because of the trailing padding.
473     // As a temporary mitigation we don't create bindings for empty bases.
474     if (const auto *BR = dyn_cast<CXXBaseObjectRegion>(R))
475       if (BR->getDecl()->isEmpty())
476         return StoreRef(store, *this);
477 
478     RegionBindingsRef B = getRegionBindings(store);
479     SVal V = svalBuilder.makeZeroVal(Ctx.CharTy);
480     B = removeSubRegionBindings(B, cast<SubRegion>(R));
481     B = B.addBinding(BindingKey::Make(R, BindingKey::Default), V);
482     return StoreRef(B.asImmutableMap().getRootWithoutRetain(), *this);
483   }
484 
485   /// Attempt to extract the fields of \p LCV and bind them to the struct region
486   /// \p R.
487   ///
488   /// This path is used when it seems advantageous to "force" loading the values
489   /// within a LazyCompoundVal to bind memberwise to the struct region, rather
490   /// than using a Default binding at the base of the entire region. This is a
491   /// heuristic attempting to avoid building long chains of LazyCompoundVals.
492   ///
493   /// \returns The updated store bindings, or \c None if binding non-lazily
494   ///          would be too expensive.
495   Optional<RegionBindingsRef> tryBindSmallStruct(RegionBindingsConstRef B,
496                                                  const TypedValueRegion *R,
497                                                  const RecordDecl *RD,
498                                                  nonloc::LazyCompoundVal LCV);
499 
500   /// BindStruct - Bind a compound value to a structure.
501   RegionBindingsRef bindStruct(RegionBindingsConstRef B,
502                                const TypedValueRegion* R, SVal V);
503 
504   /// BindVector - Bind a compound value to a vector.
505   RegionBindingsRef bindVector(RegionBindingsConstRef B,
506                                const TypedValueRegion* R, SVal V);
507 
508   RegionBindingsRef bindArray(RegionBindingsConstRef B,
509                               const TypedValueRegion* R,
510                               SVal V);
511 
512   /// Clears out all bindings in the given region and assigns a new value
513   /// as a Default binding.
514   RegionBindingsRef bindAggregate(RegionBindingsConstRef B,
515                                   const TypedRegion *R,
516                                   SVal DefaultVal);
517 
518   /// Create a new store with the specified binding removed.
519   /// \param ST the original store, that is the basis for the new store.
520   /// \param L the location whose binding should be removed.
521   StoreRef killBinding(Store ST, Loc L) override;
522 
523   void incrementReferenceCount(Store store) override {
524     getRegionBindings(store).manualRetain();
525   }
526 
527   /// If the StoreManager supports it, decrement the reference count of
528   /// the specified Store object.  If the reference count hits 0, the memory
529   /// associated with the object is recycled.
530   void decrementReferenceCount(Store store) override {
531     getRegionBindings(store).manualRelease();
532   }
533 
534   bool includedInBindings(Store store, const MemRegion *region) const override;
535 
536   /// Return the value bound to specified location in a given state.
537   ///
538   /// The high level logic for this method is this:
539   /// getBinding (L)
540   ///   if L has binding
541   ///     return L's binding
542   ///   else if L is in killset
543   ///     return unknown
544   ///   else
545   ///     if L is on stack or heap
546   ///       return undefined
547   ///     else
548   ///       return symbolic
549   SVal getBinding(Store S, Loc L, QualType T) override {
550     return getBinding(getRegionBindings(S), L, T);
551   }
552 
553   Optional<SVal> getDefaultBinding(Store S, const MemRegion *R) override {
554     RegionBindingsRef B = getRegionBindings(S);
555     // Default bindings are always applied over a base region so look up the
556     // base region's default binding, otherwise the lookup will fail when R
557     // is at an offset from R->getBaseRegion().
558     return B.getDefaultBinding(R->getBaseRegion());
559   }
560 
561   SVal getBinding(RegionBindingsConstRef B, Loc L, QualType T = QualType());
562 
563   SVal getBindingForElement(RegionBindingsConstRef B, const ElementRegion *R);
564 
565   SVal getBindingForField(RegionBindingsConstRef B, const FieldRegion *R);
566 
567   SVal getBindingForObjCIvar(RegionBindingsConstRef B, const ObjCIvarRegion *R);
568 
569   SVal getBindingForVar(RegionBindingsConstRef B, const VarRegion *R);
570 
571   SVal getBindingForLazySymbol(const TypedValueRegion *R);
572 
573   SVal getBindingForFieldOrElementCommon(RegionBindingsConstRef B,
574                                          const TypedValueRegion *R,
575                                          QualType Ty);
576 
577   SVal getLazyBinding(const SubRegion *LazyBindingRegion,
578                       RegionBindingsRef LazyBinding);
579 
580   /// Get bindings for the values in a struct and return a CompoundVal, used
581   /// when doing struct copy:
582   /// struct s x, y;
583   /// x = y;
584   /// y's value is retrieved by this method.
585   SVal getBindingForStruct(RegionBindingsConstRef B, const TypedValueRegion *R);
586   SVal getBindingForArray(RegionBindingsConstRef B, const TypedValueRegion *R);
587   NonLoc createLazyBinding(RegionBindingsConstRef B, const TypedValueRegion *R);
588 
589   /// Used to lazily generate derived symbols for bindings that are defined
590   /// implicitly by default bindings in a super region.
591   ///
592   /// Note that callers may need to specially handle LazyCompoundVals, which
593   /// are returned as is in case the caller needs to treat them differently.
594   Optional<SVal> getBindingForDerivedDefaultValue(RegionBindingsConstRef B,
595                                                   const MemRegion *superR,
596                                                   const TypedValueRegion *R,
597                                                   QualType Ty);
598 
599   /// Get the state and region whose binding this region \p R corresponds to.
600   ///
601   /// If there is no lazy binding for \p R, the returned value will have a null
602   /// \c second. Note that a null pointer can represents a valid Store.
603   std::pair<Store, const SubRegion *>
604   findLazyBinding(RegionBindingsConstRef B, const SubRegion *R,
605                   const SubRegion *originalRegion);
606 
607   /// Returns the cached set of interesting SVals contained within a lazy
608   /// binding.
609   ///
610   /// The precise value of "interesting" is determined for the purposes of
611   /// RegionStore's internal analysis. It must always contain all regions and
612   /// symbols, but may omit constants and other kinds of SVal.
613   const SValListTy &getInterestingValues(nonloc::LazyCompoundVal LCV);
614 
615   //===------------------------------------------------------------------===//
616   // State pruning.
617   //===------------------------------------------------------------------===//
618 
619   /// removeDeadBindings - Scans the RegionStore of 'state' for dead values.
620   ///  It returns a new Store with these values removed.
621   StoreRef removeDeadBindings(Store store, const StackFrameContext *LCtx,
622                               SymbolReaper& SymReaper) override;
623 
624   //===------------------------------------------------------------------===//
625   // Utility methods.
626   //===------------------------------------------------------------------===//
627 
628   RegionBindingsRef getRegionBindings(Store store) const {
629     llvm::PointerIntPair<Store, 1, bool> Ptr;
630     Ptr.setFromOpaqueValue(const_cast<void *>(store));
631     return RegionBindingsRef(
632         CBFactory,
633         static_cast<const RegionBindings::TreeTy *>(Ptr.getPointer()),
634         RBFactory.getTreeFactory(),
635         Ptr.getInt());
636   }
637 
638   void printJson(raw_ostream &Out, Store S, const char *NL = "\n",
639                  unsigned int Space = 0, bool IsDot = false) const override;
640 
641   void iterBindings(Store store, BindingsHandler& f) override {
642     RegionBindingsRef B = getRegionBindings(store);
643     for (RegionBindingsRef::iterator I = B.begin(), E = B.end(); I != E; ++I) {
644       const ClusterBindings &Cluster = I.getData();
645       for (ClusterBindings::iterator CI = Cluster.begin(), CE = Cluster.end();
646            CI != CE; ++CI) {
647         const BindingKey &K = CI.getKey();
648         if (!K.isDirect())
649           continue;
650         if (const SubRegion *R = dyn_cast<SubRegion>(K.getRegion())) {
651           // FIXME: Possibly incorporate the offset?
652           if (!f.HandleBinding(*this, store, R, CI.getData()))
653             return;
654         }
655       }
656     }
657   }
658 };
659 
660 } // end anonymous namespace
661 
662 //===----------------------------------------------------------------------===//
663 // RegionStore creation.
664 //===----------------------------------------------------------------------===//
665 
666 std::unique_ptr<StoreManager>
667 ento::CreateRegionStoreManager(ProgramStateManager &StMgr) {
668   RegionStoreFeatures F = maximal_features_tag();
669   return std::make_unique<RegionStoreManager>(StMgr, F);
670 }
671 
672 std::unique_ptr<StoreManager>
673 ento::CreateFieldsOnlyRegionStoreManager(ProgramStateManager &StMgr) {
674   RegionStoreFeatures F = minimal_features_tag();
675   F.enableFields(true);
676   return std::make_unique<RegionStoreManager>(StMgr, F);
677 }
678 
679 
680 //===----------------------------------------------------------------------===//
681 // Region Cluster analysis.
682 //===----------------------------------------------------------------------===//
683 
684 namespace {
685 /// Used to determine which global regions are automatically included in the
686 /// initial worklist of a ClusterAnalysis.
687 enum GlobalsFilterKind {
688   /// Don't include any global regions.
689   GFK_None,
690   /// Only include system globals.
691   GFK_SystemOnly,
692   /// Include all global regions.
693   GFK_All
694 };
695 
696 template <typename DERIVED>
697 class ClusterAnalysis  {
698 protected:
699   typedef llvm::DenseMap<const MemRegion *, const ClusterBindings *> ClusterMap;
700   typedef const MemRegion * WorkListElement;
701   typedef SmallVector<WorkListElement, 10> WorkList;
702 
703   llvm::SmallPtrSet<const ClusterBindings *, 16> Visited;
704 
705   WorkList WL;
706 
707   RegionStoreManager &RM;
708   ASTContext &Ctx;
709   SValBuilder &svalBuilder;
710 
711   RegionBindingsRef B;
712 
713 
714 protected:
715   const ClusterBindings *getCluster(const MemRegion *R) {
716     return B.lookup(R);
717   }
718 
719   /// Returns true if all clusters in the given memspace should be initially
720   /// included in the cluster analysis. Subclasses may provide their
721   /// own implementation.
722   bool includeEntireMemorySpace(const MemRegion *Base) {
723     return false;
724   }
725 
726 public:
727   ClusterAnalysis(RegionStoreManager &rm, ProgramStateManager &StateMgr,
728                   RegionBindingsRef b)
729       : RM(rm), Ctx(StateMgr.getContext()),
730         svalBuilder(StateMgr.getSValBuilder()), B(std::move(b)) {}
731 
732   RegionBindingsRef getRegionBindings() const { return B; }
733 
734   bool isVisited(const MemRegion *R) {
735     return Visited.count(getCluster(R));
736   }
737 
738   void GenerateClusters() {
739     // Scan the entire set of bindings and record the region clusters.
740     for (RegionBindingsRef::iterator RI = B.begin(), RE = B.end();
741          RI != RE; ++RI){
742       const MemRegion *Base = RI.getKey();
743 
744       const ClusterBindings &Cluster = RI.getData();
745       assert(!Cluster.isEmpty() && "Empty clusters should be removed");
746       static_cast<DERIVED*>(this)->VisitAddedToCluster(Base, Cluster);
747 
748       // If the base's memspace should be entirely invalidated, add the cluster
749       // to the workspace up front.
750       if (static_cast<DERIVED*>(this)->includeEntireMemorySpace(Base))
751         AddToWorkList(WorkListElement(Base), &Cluster);
752     }
753   }
754 
755   bool AddToWorkList(WorkListElement E, const ClusterBindings *C) {
756     if (C && !Visited.insert(C).second)
757       return false;
758     WL.push_back(E);
759     return true;
760   }
761 
762   bool AddToWorkList(const MemRegion *R) {
763     return static_cast<DERIVED*>(this)->AddToWorkList(R);
764   }
765 
766   void RunWorkList() {
767     while (!WL.empty()) {
768       WorkListElement E = WL.pop_back_val();
769       const MemRegion *BaseR = E;
770 
771       static_cast<DERIVED*>(this)->VisitCluster(BaseR, getCluster(BaseR));
772     }
773   }
774 
775   void VisitAddedToCluster(const MemRegion *baseR, const ClusterBindings &C) {}
776   void VisitCluster(const MemRegion *baseR, const ClusterBindings *C) {}
777 
778   void VisitCluster(const MemRegion *BaseR, const ClusterBindings *C,
779                     bool Flag) {
780     static_cast<DERIVED*>(this)->VisitCluster(BaseR, C);
781   }
782 };
783 }
784 
785 //===----------------------------------------------------------------------===//
786 // Binding invalidation.
787 //===----------------------------------------------------------------------===//
788 
789 bool RegionStoreManager::scanReachableSymbols(Store S, const MemRegion *R,
790                                               ScanReachableSymbols &Callbacks) {
791   assert(R == R->getBaseRegion() && "Should only be called for base regions");
792   RegionBindingsRef B = getRegionBindings(S);
793   const ClusterBindings *Cluster = B.lookup(R);
794 
795   if (!Cluster)
796     return true;
797 
798   for (ClusterBindings::iterator RI = Cluster->begin(), RE = Cluster->end();
799        RI != RE; ++RI) {
800     if (!Callbacks.scan(RI.getData()))
801       return false;
802   }
803 
804   return true;
805 }
806 
807 static inline bool isUnionField(const FieldRegion *FR) {
808   return FR->getDecl()->getParent()->isUnion();
809 }
810 
811 typedef SmallVector<const FieldDecl *, 8> FieldVector;
812 
813 static void getSymbolicOffsetFields(BindingKey K, FieldVector &Fields) {
814   assert(K.hasSymbolicOffset() && "Not implemented for concrete offset keys");
815 
816   const MemRegion *Base = K.getConcreteOffsetRegion();
817   const MemRegion *R = K.getRegion();
818 
819   while (R != Base) {
820     if (const FieldRegion *FR = dyn_cast<FieldRegion>(R))
821       if (!isUnionField(FR))
822         Fields.push_back(FR->getDecl());
823 
824     R = cast<SubRegion>(R)->getSuperRegion();
825   }
826 }
827 
828 static bool isCompatibleWithFields(BindingKey K, const FieldVector &Fields) {
829   assert(K.hasSymbolicOffset() && "Not implemented for concrete offset keys");
830 
831   if (Fields.empty())
832     return true;
833 
834   FieldVector FieldsInBindingKey;
835   getSymbolicOffsetFields(K, FieldsInBindingKey);
836 
837   ptrdiff_t Delta = FieldsInBindingKey.size() - Fields.size();
838   if (Delta >= 0)
839     return std::equal(FieldsInBindingKey.begin() + Delta,
840                       FieldsInBindingKey.end(),
841                       Fields.begin());
842   else
843     return std::equal(FieldsInBindingKey.begin(), FieldsInBindingKey.end(),
844                       Fields.begin() - Delta);
845 }
846 
847 /// Collects all bindings in \p Cluster that may refer to bindings within
848 /// \p Top.
849 ///
850 /// Each binding is a pair whose \c first is the key (a BindingKey) and whose
851 /// \c second is the value (an SVal).
852 ///
853 /// The \p IncludeAllDefaultBindings parameter specifies whether to include
854 /// default bindings that may extend beyond \p Top itself, e.g. if \p Top is
855 /// an aggregate within a larger aggregate with a default binding.
856 static void
857 collectSubRegionBindings(SmallVectorImpl<BindingPair> &Bindings,
858                          SValBuilder &SVB, const ClusterBindings &Cluster,
859                          const SubRegion *Top, BindingKey TopKey,
860                          bool IncludeAllDefaultBindings) {
861   FieldVector FieldsInSymbolicSubregions;
862   if (TopKey.hasSymbolicOffset()) {
863     getSymbolicOffsetFields(TopKey, FieldsInSymbolicSubregions);
864     Top = TopKey.getConcreteOffsetRegion();
865     TopKey = BindingKey::Make(Top, BindingKey::Default);
866   }
867 
868   // Find the length (in bits) of the region being invalidated.
869   uint64_t Length = UINT64_MAX;
870   SVal Extent = Top->getMemRegionManager().getStaticSize(Top, SVB);
871   if (Optional<nonloc::ConcreteInt> ExtentCI =
872           Extent.getAs<nonloc::ConcreteInt>()) {
873     const llvm::APSInt &ExtentInt = ExtentCI->getValue();
874     assert(ExtentInt.isNonNegative() || ExtentInt.isUnsigned());
875     // Extents are in bytes but region offsets are in bits. Be careful!
876     Length = ExtentInt.getLimitedValue() * SVB.getContext().getCharWidth();
877   } else if (const FieldRegion *FR = dyn_cast<FieldRegion>(Top)) {
878     if (FR->getDecl()->isBitField())
879       Length = FR->getDecl()->getBitWidthValue(SVB.getContext());
880   }
881 
882   for (ClusterBindings::iterator I = Cluster.begin(), E = Cluster.end();
883        I != E; ++I) {
884     BindingKey NextKey = I.getKey();
885     if (NextKey.getRegion() == TopKey.getRegion()) {
886       // FIXME: This doesn't catch the case where we're really invalidating a
887       // region with a symbolic offset. Example:
888       //      R: points[i].y
889       //   Next: points[0].x
890 
891       if (NextKey.getOffset() > TopKey.getOffset() &&
892           NextKey.getOffset() - TopKey.getOffset() < Length) {
893         // Case 1: The next binding is inside the region we're invalidating.
894         // Include it.
895         Bindings.push_back(*I);
896 
897       } else if (NextKey.getOffset() == TopKey.getOffset()) {
898         // Case 2: The next binding is at the same offset as the region we're
899         // invalidating. In this case, we need to leave default bindings alone,
900         // since they may be providing a default value for a regions beyond what
901         // we're invalidating.
902         // FIXME: This is probably incorrect; consider invalidating an outer
903         // struct whose first field is bound to a LazyCompoundVal.
904         if (IncludeAllDefaultBindings || NextKey.isDirect())
905           Bindings.push_back(*I);
906       }
907 
908     } else if (NextKey.hasSymbolicOffset()) {
909       const MemRegion *Base = NextKey.getConcreteOffsetRegion();
910       if (Top->isSubRegionOf(Base) && Top != Base) {
911         // Case 3: The next key is symbolic and we just changed something within
912         // its concrete region. We don't know if the binding is still valid, so
913         // we'll be conservative and include it.
914         if (IncludeAllDefaultBindings || NextKey.isDirect())
915           if (isCompatibleWithFields(NextKey, FieldsInSymbolicSubregions))
916             Bindings.push_back(*I);
917       } else if (const SubRegion *BaseSR = dyn_cast<SubRegion>(Base)) {
918         // Case 4: The next key is symbolic, but we changed a known
919         // super-region. In this case the binding is certainly included.
920         if (BaseSR->isSubRegionOf(Top))
921           if (isCompatibleWithFields(NextKey, FieldsInSymbolicSubregions))
922             Bindings.push_back(*I);
923       }
924     }
925   }
926 }
927 
928 static void
929 collectSubRegionBindings(SmallVectorImpl<BindingPair> &Bindings,
930                          SValBuilder &SVB, const ClusterBindings &Cluster,
931                          const SubRegion *Top, bool IncludeAllDefaultBindings) {
932   collectSubRegionBindings(Bindings, SVB, Cluster, Top,
933                            BindingKey::Make(Top, BindingKey::Default),
934                            IncludeAllDefaultBindings);
935 }
936 
937 RegionBindingsRef
938 RegionStoreManager::removeSubRegionBindings(RegionBindingsConstRef B,
939                                             const SubRegion *Top) {
940   BindingKey TopKey = BindingKey::Make(Top, BindingKey::Default);
941   const MemRegion *ClusterHead = TopKey.getBaseRegion();
942 
943   if (Top == ClusterHead) {
944     // We can remove an entire cluster's bindings all in one go.
945     return B.remove(Top);
946   }
947 
948   const ClusterBindings *Cluster = B.lookup(ClusterHead);
949   if (!Cluster) {
950     // If we're invalidating a region with a symbolic offset, we need to make
951     // sure we don't treat the base region as uninitialized anymore.
952     if (TopKey.hasSymbolicOffset()) {
953       const SubRegion *Concrete = TopKey.getConcreteOffsetRegion();
954       return B.addBinding(Concrete, BindingKey::Default, UnknownVal());
955     }
956     return B;
957   }
958 
959   SmallVector<BindingPair, 32> Bindings;
960   collectSubRegionBindings(Bindings, svalBuilder, *Cluster, Top, TopKey,
961                            /*IncludeAllDefaultBindings=*/false);
962 
963   ClusterBindingsRef Result(*Cluster, CBFactory);
964   for (SmallVectorImpl<BindingPair>::const_iterator I = Bindings.begin(),
965                                                     E = Bindings.end();
966        I != E; ++I)
967     Result = Result.remove(I->first);
968 
969   // If we're invalidating a region with a symbolic offset, we need to make sure
970   // we don't treat the base region as uninitialized anymore.
971   // FIXME: This isn't very precise; see the example in
972   // collectSubRegionBindings.
973   if (TopKey.hasSymbolicOffset()) {
974     const SubRegion *Concrete = TopKey.getConcreteOffsetRegion();
975     Result = Result.add(BindingKey::Make(Concrete, BindingKey::Default),
976                         UnknownVal());
977   }
978 
979   if (Result.isEmpty())
980     return B.remove(ClusterHead);
981   return B.add(ClusterHead, Result.asImmutableMap());
982 }
983 
984 namespace {
985 class InvalidateRegionsWorker : public ClusterAnalysis<InvalidateRegionsWorker>
986 {
987   const Expr *Ex;
988   unsigned Count;
989   const LocationContext *LCtx;
990   InvalidatedSymbols &IS;
991   RegionAndSymbolInvalidationTraits &ITraits;
992   StoreManager::InvalidatedRegions *Regions;
993   GlobalsFilterKind GlobalsFilter;
994 public:
995   InvalidateRegionsWorker(RegionStoreManager &rm,
996                           ProgramStateManager &stateMgr,
997                           RegionBindingsRef b,
998                           const Expr *ex, unsigned count,
999                           const LocationContext *lctx,
1000                           InvalidatedSymbols &is,
1001                           RegionAndSymbolInvalidationTraits &ITraitsIn,
1002                           StoreManager::InvalidatedRegions *r,
1003                           GlobalsFilterKind GFK)
1004      : ClusterAnalysis<InvalidateRegionsWorker>(rm, stateMgr, b),
1005        Ex(ex), Count(count), LCtx(lctx), IS(is), ITraits(ITraitsIn), Regions(r),
1006        GlobalsFilter(GFK) {}
1007 
1008   void VisitCluster(const MemRegion *baseR, const ClusterBindings *C);
1009   void VisitBinding(SVal V);
1010 
1011   using ClusterAnalysis::AddToWorkList;
1012 
1013   bool AddToWorkList(const MemRegion *R);
1014 
1015   /// Returns true if all clusters in the memory space for \p Base should be
1016   /// be invalidated.
1017   bool includeEntireMemorySpace(const MemRegion *Base);
1018 
1019   /// Returns true if the memory space of the given region is one of the global
1020   /// regions specially included at the start of invalidation.
1021   bool isInitiallyIncludedGlobalRegion(const MemRegion *R);
1022 };
1023 }
1024 
1025 bool InvalidateRegionsWorker::AddToWorkList(const MemRegion *R) {
1026   bool doNotInvalidateSuperRegion = ITraits.hasTrait(
1027       R, RegionAndSymbolInvalidationTraits::TK_DoNotInvalidateSuperRegion);
1028   const MemRegion *BaseR = doNotInvalidateSuperRegion ? R : R->getBaseRegion();
1029   return AddToWorkList(WorkListElement(BaseR), getCluster(BaseR));
1030 }
1031 
1032 void InvalidateRegionsWorker::VisitBinding(SVal V) {
1033   // A symbol?  Mark it touched by the invalidation.
1034   if (SymbolRef Sym = V.getAsSymbol())
1035     IS.insert(Sym);
1036 
1037   if (const MemRegion *R = V.getAsRegion()) {
1038     AddToWorkList(R);
1039     return;
1040   }
1041 
1042   // Is it a LazyCompoundVal?  All references get invalidated as well.
1043   if (Optional<nonloc::LazyCompoundVal> LCS =
1044           V.getAs<nonloc::LazyCompoundVal>()) {
1045 
1046     const RegionStoreManager::SValListTy &Vals = RM.getInterestingValues(*LCS);
1047 
1048     for (RegionStoreManager::SValListTy::const_iterator I = Vals.begin(),
1049                                                         E = Vals.end();
1050          I != E; ++I)
1051       VisitBinding(*I);
1052 
1053     return;
1054   }
1055 }
1056 
1057 void InvalidateRegionsWorker::VisitCluster(const MemRegion *baseR,
1058                                            const ClusterBindings *C) {
1059 
1060   bool PreserveRegionsContents =
1061       ITraits.hasTrait(baseR,
1062                        RegionAndSymbolInvalidationTraits::TK_PreserveContents);
1063 
1064   if (C) {
1065     for (ClusterBindings::iterator I = C->begin(), E = C->end(); I != E; ++I)
1066       VisitBinding(I.getData());
1067 
1068     // Invalidate regions contents.
1069     if (!PreserveRegionsContents)
1070       B = B.remove(baseR);
1071   }
1072 
1073   if (const auto *TO = dyn_cast<TypedValueRegion>(baseR)) {
1074     if (const auto *RD = TO->getValueType()->getAsCXXRecordDecl()) {
1075 
1076       // Lambdas can affect all static local variables without explicitly
1077       // capturing those.
1078       // We invalidate all static locals referenced inside the lambda body.
1079       if (RD->isLambda() && RD->getLambdaCallOperator()->getBody()) {
1080         using namespace ast_matchers;
1081 
1082         const char *DeclBind = "DeclBind";
1083         StatementMatcher RefToStatic = stmt(hasDescendant(declRefExpr(
1084               to(varDecl(hasStaticStorageDuration()).bind(DeclBind)))));
1085         auto Matches =
1086             match(RefToStatic, *RD->getLambdaCallOperator()->getBody(),
1087                   RD->getASTContext());
1088 
1089         for (BoundNodes &Match : Matches) {
1090           auto *VD = Match.getNodeAs<VarDecl>(DeclBind);
1091           const VarRegion *ToInvalidate =
1092               RM.getRegionManager().getVarRegion(VD, LCtx);
1093           AddToWorkList(ToInvalidate);
1094         }
1095       }
1096     }
1097   }
1098 
1099   // BlockDataRegion?  If so, invalidate captured variables that are passed
1100   // by reference.
1101   if (const BlockDataRegion *BR = dyn_cast<BlockDataRegion>(baseR)) {
1102     for (BlockDataRegion::referenced_vars_iterator
1103          BI = BR->referenced_vars_begin(), BE = BR->referenced_vars_end() ;
1104          BI != BE; ++BI) {
1105       const VarRegion *VR = BI.getCapturedRegion();
1106       const VarDecl *VD = VR->getDecl();
1107       if (VD->hasAttr<BlocksAttr>() || !VD->hasLocalStorage()) {
1108         AddToWorkList(VR);
1109       }
1110       else if (Loc::isLocType(VR->getValueType())) {
1111         // Map the current bindings to a Store to retrieve the value
1112         // of the binding.  If that binding itself is a region, we should
1113         // invalidate that region.  This is because a block may capture
1114         // a pointer value, but the thing pointed by that pointer may
1115         // get invalidated.
1116         SVal V = RM.getBinding(B, loc::MemRegionVal(VR));
1117         if (Optional<Loc> L = V.getAs<Loc>()) {
1118           if (const MemRegion *LR = L->getAsRegion())
1119             AddToWorkList(LR);
1120         }
1121       }
1122     }
1123     return;
1124   }
1125 
1126   // Symbolic region?
1127   if (const SymbolicRegion *SR = dyn_cast<SymbolicRegion>(baseR))
1128     IS.insert(SR->getSymbol());
1129 
1130   // Nothing else should be done in the case when we preserve regions context.
1131   if (PreserveRegionsContents)
1132     return;
1133 
1134   // Otherwise, we have a normal data region. Record that we touched the region.
1135   if (Regions)
1136     Regions->push_back(baseR);
1137 
1138   if (isa<AllocaRegion>(baseR) || isa<SymbolicRegion>(baseR)) {
1139     // Invalidate the region by setting its default value to
1140     // conjured symbol. The type of the symbol is irrelevant.
1141     DefinedOrUnknownSVal V =
1142       svalBuilder.conjureSymbolVal(baseR, Ex, LCtx, Ctx.IntTy, Count);
1143     B = B.addBinding(baseR, BindingKey::Default, V);
1144     return;
1145   }
1146 
1147   if (!baseR->isBoundable())
1148     return;
1149 
1150   const TypedValueRegion *TR = cast<TypedValueRegion>(baseR);
1151   QualType T = TR->getValueType();
1152 
1153   if (isInitiallyIncludedGlobalRegion(baseR)) {
1154     // If the region is a global and we are invalidating all globals,
1155     // erasing the entry is good enough.  This causes all globals to be lazily
1156     // symbolicated from the same base symbol.
1157     return;
1158   }
1159 
1160   if (T->isRecordType()) {
1161     // Invalidate the region by setting its default value to
1162     // conjured symbol. The type of the symbol is irrelevant.
1163     DefinedOrUnknownSVal V = svalBuilder.conjureSymbolVal(baseR, Ex, LCtx,
1164                                                           Ctx.IntTy, Count);
1165     B = B.addBinding(baseR, BindingKey::Default, V);
1166     return;
1167   }
1168 
1169   if (const ArrayType *AT = Ctx.getAsArrayType(T)) {
1170     bool doNotInvalidateSuperRegion = ITraits.hasTrait(
1171         baseR,
1172         RegionAndSymbolInvalidationTraits::TK_DoNotInvalidateSuperRegion);
1173 
1174     if (doNotInvalidateSuperRegion) {
1175       // We are not doing blank invalidation of the whole array region so we
1176       // have to manually invalidate each elements.
1177       Optional<uint64_t> NumElements;
1178 
1179       // Compute lower and upper offsets for region within array.
1180       if (const ConstantArrayType *CAT = dyn_cast<ConstantArrayType>(AT))
1181         NumElements = CAT->getSize().getZExtValue();
1182       if (!NumElements) // We are not dealing with a constant size array
1183         goto conjure_default;
1184       QualType ElementTy = AT->getElementType();
1185       uint64_t ElemSize = Ctx.getTypeSize(ElementTy);
1186       const RegionOffset &RO = baseR->getAsOffset();
1187       const MemRegion *SuperR = baseR->getBaseRegion();
1188       if (RO.hasSymbolicOffset()) {
1189         // If base region has a symbolic offset,
1190         // we revert to invalidating the super region.
1191         if (SuperR)
1192           AddToWorkList(SuperR);
1193         goto conjure_default;
1194       }
1195 
1196       uint64_t LowerOffset = RO.getOffset();
1197       uint64_t UpperOffset = LowerOffset + *NumElements * ElemSize;
1198       bool UpperOverflow = UpperOffset < LowerOffset;
1199 
1200       // Invalidate regions which are within array boundaries,
1201       // or have a symbolic offset.
1202       if (!SuperR)
1203         goto conjure_default;
1204 
1205       const ClusterBindings *C = B.lookup(SuperR);
1206       if (!C)
1207         goto conjure_default;
1208 
1209       for (ClusterBindings::iterator I = C->begin(), E = C->end(); I != E;
1210            ++I) {
1211         const BindingKey &BK = I.getKey();
1212         Optional<uint64_t> ROffset =
1213             BK.hasSymbolicOffset() ? Optional<uint64_t>() : BK.getOffset();
1214 
1215         // Check offset is not symbolic and within array's boundaries.
1216         // Handles arrays of 0 elements and of 0-sized elements as well.
1217         if (!ROffset ||
1218             ((*ROffset >= LowerOffset && *ROffset < UpperOffset) ||
1219              (UpperOverflow &&
1220               (*ROffset >= LowerOffset || *ROffset < UpperOffset)) ||
1221              (LowerOffset == UpperOffset && *ROffset == LowerOffset))) {
1222           B = B.removeBinding(I.getKey());
1223           // Bound symbolic regions need to be invalidated for dead symbol
1224           // detection.
1225           SVal V = I.getData();
1226           const MemRegion *R = V.getAsRegion();
1227           if (R && isa<SymbolicRegion>(R))
1228             VisitBinding(V);
1229         }
1230       }
1231     }
1232   conjure_default:
1233       // Set the default value of the array to conjured symbol.
1234     DefinedOrUnknownSVal V =
1235     svalBuilder.conjureSymbolVal(baseR, Ex, LCtx,
1236                                      AT->getElementType(), Count);
1237     B = B.addBinding(baseR, BindingKey::Default, V);
1238     return;
1239   }
1240 
1241   DefinedOrUnknownSVal V = svalBuilder.conjureSymbolVal(baseR, Ex, LCtx,
1242                                                         T,Count);
1243   assert(SymbolManager::canSymbolicate(T) || V.isUnknown());
1244   B = B.addBinding(baseR, BindingKey::Direct, V);
1245 }
1246 
1247 bool InvalidateRegionsWorker::isInitiallyIncludedGlobalRegion(
1248     const MemRegion *R) {
1249   switch (GlobalsFilter) {
1250   case GFK_None:
1251     return false;
1252   case GFK_SystemOnly:
1253     return isa<GlobalSystemSpaceRegion>(R->getMemorySpace());
1254   case GFK_All:
1255     return isa<NonStaticGlobalSpaceRegion>(R->getMemorySpace());
1256   }
1257 
1258   llvm_unreachable("unknown globals filter");
1259 }
1260 
1261 bool InvalidateRegionsWorker::includeEntireMemorySpace(const MemRegion *Base) {
1262   if (isInitiallyIncludedGlobalRegion(Base))
1263     return true;
1264 
1265   const MemSpaceRegion *MemSpace = Base->getMemorySpace();
1266   return ITraits.hasTrait(MemSpace,
1267                           RegionAndSymbolInvalidationTraits::TK_EntireMemSpace);
1268 }
1269 
1270 RegionBindingsRef
1271 RegionStoreManager::invalidateGlobalRegion(MemRegion::Kind K,
1272                                            const Expr *Ex,
1273                                            unsigned Count,
1274                                            const LocationContext *LCtx,
1275                                            RegionBindingsRef B,
1276                                            InvalidatedRegions *Invalidated) {
1277   // Bind the globals memory space to a new symbol that we will use to derive
1278   // the bindings for all globals.
1279   const GlobalsSpaceRegion *GS = MRMgr.getGlobalsRegion(K);
1280   SVal V = svalBuilder.conjureSymbolVal(/* symbolTag = */ (const void*) GS, Ex, LCtx,
1281                                         /* type does not matter */ Ctx.IntTy,
1282                                         Count);
1283 
1284   B = B.removeBinding(GS)
1285        .addBinding(BindingKey::Make(GS, BindingKey::Default), V);
1286 
1287   // Even if there are no bindings in the global scope, we still need to
1288   // record that we touched it.
1289   if (Invalidated)
1290     Invalidated->push_back(GS);
1291 
1292   return B;
1293 }
1294 
1295 void RegionStoreManager::populateWorkList(InvalidateRegionsWorker &W,
1296                                           ArrayRef<SVal> Values,
1297                                           InvalidatedRegions *TopLevelRegions) {
1298   for (ArrayRef<SVal>::iterator I = Values.begin(),
1299                                 E = Values.end(); I != E; ++I) {
1300     SVal V = *I;
1301     if (Optional<nonloc::LazyCompoundVal> LCS =
1302         V.getAs<nonloc::LazyCompoundVal>()) {
1303 
1304       const SValListTy &Vals = getInterestingValues(*LCS);
1305 
1306       for (SValListTy::const_iterator I = Vals.begin(),
1307                                       E = Vals.end(); I != E; ++I) {
1308         // Note: the last argument is false here because these are
1309         // non-top-level regions.
1310         if (const MemRegion *R = (*I).getAsRegion())
1311           W.AddToWorkList(R);
1312       }
1313       continue;
1314     }
1315 
1316     if (const MemRegion *R = V.getAsRegion()) {
1317       if (TopLevelRegions)
1318         TopLevelRegions->push_back(R);
1319       W.AddToWorkList(R);
1320       continue;
1321     }
1322   }
1323 }
1324 
1325 StoreRef
1326 RegionStoreManager::invalidateRegions(Store store,
1327                                      ArrayRef<SVal> Values,
1328                                      const Expr *Ex, unsigned Count,
1329                                      const LocationContext *LCtx,
1330                                      const CallEvent *Call,
1331                                      InvalidatedSymbols &IS,
1332                                      RegionAndSymbolInvalidationTraits &ITraits,
1333                                      InvalidatedRegions *TopLevelRegions,
1334                                      InvalidatedRegions *Invalidated) {
1335   GlobalsFilterKind GlobalsFilter;
1336   if (Call) {
1337     if (Call->isInSystemHeader())
1338       GlobalsFilter = GFK_SystemOnly;
1339     else
1340       GlobalsFilter = GFK_All;
1341   } else {
1342     GlobalsFilter = GFK_None;
1343   }
1344 
1345   RegionBindingsRef B = getRegionBindings(store);
1346   InvalidateRegionsWorker W(*this, StateMgr, B, Ex, Count, LCtx, IS, ITraits,
1347                             Invalidated, GlobalsFilter);
1348 
1349   // Scan the bindings and generate the clusters.
1350   W.GenerateClusters();
1351 
1352   // Add the regions to the worklist.
1353   populateWorkList(W, Values, TopLevelRegions);
1354 
1355   W.RunWorkList();
1356 
1357   // Return the new bindings.
1358   B = W.getRegionBindings();
1359 
1360   // For calls, determine which global regions should be invalidated and
1361   // invalidate them. (Note that function-static and immutable globals are never
1362   // invalidated by this.)
1363   // TODO: This could possibly be more precise with modules.
1364   switch (GlobalsFilter) {
1365   case GFK_All:
1366     B = invalidateGlobalRegion(MemRegion::GlobalInternalSpaceRegionKind,
1367                                Ex, Count, LCtx, B, Invalidated);
1368     LLVM_FALLTHROUGH;
1369   case GFK_SystemOnly:
1370     B = invalidateGlobalRegion(MemRegion::GlobalSystemSpaceRegionKind,
1371                                Ex, Count, LCtx, B, Invalidated);
1372     LLVM_FALLTHROUGH;
1373   case GFK_None:
1374     break;
1375   }
1376 
1377   return StoreRef(B.asStore(), *this);
1378 }
1379 
1380 //===----------------------------------------------------------------------===//
1381 // Location and region casting.
1382 //===----------------------------------------------------------------------===//
1383 
1384 /// ArrayToPointer - Emulates the "decay" of an array to a pointer
1385 ///  type.  'Array' represents the lvalue of the array being decayed
1386 ///  to a pointer, and the returned SVal represents the decayed
1387 ///  version of that lvalue (i.e., a pointer to the first element of
1388 ///  the array).  This is called by ExprEngine when evaluating casts
1389 ///  from arrays to pointers.
1390 SVal RegionStoreManager::ArrayToPointer(Loc Array, QualType T) {
1391   if (Array.getAs<loc::ConcreteInt>())
1392     return Array;
1393 
1394   if (!Array.getAs<loc::MemRegionVal>())
1395     return UnknownVal();
1396 
1397   const SubRegion *R =
1398       cast<SubRegion>(Array.castAs<loc::MemRegionVal>().getRegion());
1399   NonLoc ZeroIdx = svalBuilder.makeZeroArrayIndex();
1400   return loc::MemRegionVal(MRMgr.getElementRegion(T, ZeroIdx, R, Ctx));
1401 }
1402 
1403 //===----------------------------------------------------------------------===//
1404 // Loading values from regions.
1405 //===----------------------------------------------------------------------===//
1406 
1407 SVal RegionStoreManager::getBinding(RegionBindingsConstRef B, Loc L, QualType T) {
1408   assert(!L.getAs<UnknownVal>() && "location unknown");
1409   assert(!L.getAs<UndefinedVal>() && "location undefined");
1410 
1411   // For access to concrete addresses, return UnknownVal.  Checks
1412   // for null dereferences (and similar errors) are done by checkers, not
1413   // the Store.
1414   // FIXME: We can consider lazily symbolicating such memory, but we really
1415   // should defer this when we can reason easily about symbolicating arrays
1416   // of bytes.
1417   if (L.getAs<loc::ConcreteInt>()) {
1418     return UnknownVal();
1419   }
1420   if (!L.getAs<loc::MemRegionVal>()) {
1421     return UnknownVal();
1422   }
1423 
1424   const MemRegion *MR = L.castAs<loc::MemRegionVal>().getRegion();
1425 
1426   if (isa<BlockDataRegion>(MR)) {
1427     return UnknownVal();
1428   }
1429 
1430   if (!isa<TypedValueRegion>(MR)) {
1431     if (T.isNull()) {
1432       if (const TypedRegion *TR = dyn_cast<TypedRegion>(MR))
1433         T = TR->getLocationType()->getPointeeType();
1434       else if (const SymbolicRegion *SR = dyn_cast<SymbolicRegion>(MR))
1435         T = SR->getSymbol()->getType()->getPointeeType();
1436     }
1437     assert(!T.isNull() && "Unable to auto-detect binding type!");
1438     assert(!T->isVoidType() && "Attempting to dereference a void pointer!");
1439     MR = GetElementZeroRegion(cast<SubRegion>(MR), T);
1440   } else {
1441     T = cast<TypedValueRegion>(MR)->getValueType();
1442   }
1443 
1444   // FIXME: Perhaps this method should just take a 'const MemRegion*' argument
1445   //  instead of 'Loc', and have the other Loc cases handled at a higher level.
1446   const TypedValueRegion *R = cast<TypedValueRegion>(MR);
1447   QualType RTy = R->getValueType();
1448 
1449   // FIXME: we do not yet model the parts of a complex type, so treat the
1450   // whole thing as "unknown".
1451   if (RTy->isAnyComplexType())
1452     return UnknownVal();
1453 
1454   // FIXME: We should eventually handle funny addressing.  e.g.:
1455   //
1456   //   int x = ...;
1457   //   int *p = &x;
1458   //   char *q = (char*) p;
1459   //   char c = *q;  // returns the first byte of 'x'.
1460   //
1461   // Such funny addressing will occur due to layering of regions.
1462   if (RTy->isStructureOrClassType())
1463     return getBindingForStruct(B, R);
1464 
1465   // FIXME: Handle unions.
1466   if (RTy->isUnionType())
1467     return createLazyBinding(B, R);
1468 
1469   if (RTy->isArrayType()) {
1470     if (RTy->isConstantArrayType())
1471       return getBindingForArray(B, R);
1472     else
1473       return UnknownVal();
1474   }
1475 
1476   // FIXME: handle Vector types.
1477   if (RTy->isVectorType())
1478     return UnknownVal();
1479 
1480   if (const FieldRegion* FR = dyn_cast<FieldRegion>(R))
1481     return svalBuilder.evalCast(getBindingForField(B, FR), T, QualType{});
1482 
1483   if (const ElementRegion* ER = dyn_cast<ElementRegion>(R)) {
1484     // FIXME: Here we actually perform an implicit conversion from the loaded
1485     // value to the element type.  Eventually we want to compose these values
1486     // more intelligently.  For example, an 'element' can encompass multiple
1487     // bound regions (e.g., several bound bytes), or could be a subset of
1488     // a larger value.
1489     return svalBuilder.evalCast(getBindingForElement(B, ER), T, QualType{});
1490   }
1491 
1492   if (const ObjCIvarRegion *IVR = dyn_cast<ObjCIvarRegion>(R)) {
1493     // FIXME: Here we actually perform an implicit conversion from the loaded
1494     // value to the ivar type.  What we should model is stores to ivars
1495     // that blow past the extent of the ivar.  If the address of the ivar is
1496     // reinterpretted, it is possible we stored a different value that could
1497     // fit within the ivar.  Either we need to cast these when storing them
1498     // or reinterpret them lazily (as we do here).
1499     return svalBuilder.evalCast(getBindingForObjCIvar(B, IVR), T, QualType{});
1500   }
1501 
1502   if (const VarRegion *VR = dyn_cast<VarRegion>(R)) {
1503     // FIXME: Here we actually perform an implicit conversion from the loaded
1504     // value to the variable type.  What we should model is stores to variables
1505     // that blow past the extent of the variable.  If the address of the
1506     // variable is reinterpretted, it is possible we stored a different value
1507     // that could fit within the variable.  Either we need to cast these when
1508     // storing them or reinterpret them lazily (as we do here).
1509     return svalBuilder.evalCast(getBindingForVar(B, VR), T, QualType{});
1510   }
1511 
1512   const SVal *V = B.lookup(R, BindingKey::Direct);
1513 
1514   // Check if the region has a binding.
1515   if (V)
1516     return *V;
1517 
1518   // The location does not have a bound value.  This means that it has
1519   // the value it had upon its creation and/or entry to the analyzed
1520   // function/method.  These are either symbolic values or 'undefined'.
1521   if (R->hasStackNonParametersStorage()) {
1522     // All stack variables are considered to have undefined values
1523     // upon creation.  All heap allocated blocks are considered to
1524     // have undefined values as well unless they are explicitly bound
1525     // to specific values.
1526     return UndefinedVal();
1527   }
1528 
1529   // All other values are symbolic.
1530   return svalBuilder.getRegionValueSymbolVal(R);
1531 }
1532 
1533 static QualType getUnderlyingType(const SubRegion *R) {
1534   QualType RegionTy;
1535   if (const TypedValueRegion *TVR = dyn_cast<TypedValueRegion>(R))
1536     RegionTy = TVR->getValueType();
1537 
1538   if (const SymbolicRegion *SR = dyn_cast<SymbolicRegion>(R))
1539     RegionTy = SR->getSymbol()->getType();
1540 
1541   return RegionTy;
1542 }
1543 
1544 /// Checks to see if store \p B has a lazy binding for region \p R.
1545 ///
1546 /// If \p AllowSubregionBindings is \c false, a lazy binding will be rejected
1547 /// if there are additional bindings within \p R.
1548 ///
1549 /// Note that unlike RegionStoreManager::findLazyBinding, this will not search
1550 /// for lazy bindings for super-regions of \p R.
1551 static Optional<nonloc::LazyCompoundVal>
1552 getExistingLazyBinding(SValBuilder &SVB, RegionBindingsConstRef B,
1553                        const SubRegion *R, bool AllowSubregionBindings) {
1554   Optional<SVal> V = B.getDefaultBinding(R);
1555   if (!V)
1556     return None;
1557 
1558   Optional<nonloc::LazyCompoundVal> LCV = V->getAs<nonloc::LazyCompoundVal>();
1559   if (!LCV)
1560     return None;
1561 
1562   // If the LCV is for a subregion, the types might not match, and we shouldn't
1563   // reuse the binding.
1564   QualType RegionTy = getUnderlyingType(R);
1565   if (!RegionTy.isNull() &&
1566       !RegionTy->isVoidPointerType()) {
1567     QualType SourceRegionTy = LCV->getRegion()->getValueType();
1568     if (!SVB.getContext().hasSameUnqualifiedType(RegionTy, SourceRegionTy))
1569       return None;
1570   }
1571 
1572   if (!AllowSubregionBindings) {
1573     // If there are any other bindings within this region, we shouldn't reuse
1574     // the top-level binding.
1575     SmallVector<BindingPair, 16> Bindings;
1576     collectSubRegionBindings(Bindings, SVB, *B.lookup(R->getBaseRegion()), R,
1577                              /*IncludeAllDefaultBindings=*/true);
1578     if (Bindings.size() > 1)
1579       return None;
1580   }
1581 
1582   return *LCV;
1583 }
1584 
1585 
1586 std::pair<Store, const SubRegion *>
1587 RegionStoreManager::findLazyBinding(RegionBindingsConstRef B,
1588                                    const SubRegion *R,
1589                                    const SubRegion *originalRegion) {
1590   if (originalRegion != R) {
1591     if (Optional<nonloc::LazyCompoundVal> V =
1592           getExistingLazyBinding(svalBuilder, B, R, true))
1593       return std::make_pair(V->getStore(), V->getRegion());
1594   }
1595 
1596   typedef std::pair<Store, const SubRegion *> StoreRegionPair;
1597   StoreRegionPair Result = StoreRegionPair();
1598 
1599   if (const ElementRegion *ER = dyn_cast<ElementRegion>(R)) {
1600     Result = findLazyBinding(B, cast<SubRegion>(ER->getSuperRegion()),
1601                              originalRegion);
1602 
1603     if (Result.second)
1604       Result.second = MRMgr.getElementRegionWithSuper(ER, Result.second);
1605 
1606   } else if (const FieldRegion *FR = dyn_cast<FieldRegion>(R)) {
1607     Result = findLazyBinding(B, cast<SubRegion>(FR->getSuperRegion()),
1608                                        originalRegion);
1609 
1610     if (Result.second)
1611       Result.second = MRMgr.getFieldRegionWithSuper(FR, Result.second);
1612 
1613   } else if (const CXXBaseObjectRegion *BaseReg =
1614                dyn_cast<CXXBaseObjectRegion>(R)) {
1615     // C++ base object region is another kind of region that we should blast
1616     // through to look for lazy compound value. It is like a field region.
1617     Result = findLazyBinding(B, cast<SubRegion>(BaseReg->getSuperRegion()),
1618                              originalRegion);
1619 
1620     if (Result.second)
1621       Result.second = MRMgr.getCXXBaseObjectRegionWithSuper(BaseReg,
1622                                                             Result.second);
1623   }
1624 
1625   return Result;
1626 }
1627 
1628 SVal RegionStoreManager::getBindingForElement(RegionBindingsConstRef B,
1629                                               const ElementRegion* R) {
1630   // Check if the region has a binding.
1631   if (const Optional<SVal> &V = B.getDirectBinding(R))
1632     return *V;
1633 
1634   const MemRegion* superR = R->getSuperRegion();
1635 
1636   // Check if the region is an element region of a string literal.
1637   if (const StringRegion *StrR = dyn_cast<StringRegion>(superR)) {
1638     // FIXME: Handle loads from strings where the literal is treated as
1639     // an integer, e.g., *((unsigned int*)"hello")
1640     QualType T = Ctx.getAsArrayType(StrR->getValueType())->getElementType();
1641     if (!Ctx.hasSameUnqualifiedType(T, R->getElementType()))
1642       return UnknownVal();
1643 
1644     const StringLiteral *Str = StrR->getStringLiteral();
1645     SVal Idx = R->getIndex();
1646     if (Optional<nonloc::ConcreteInt> CI = Idx.getAs<nonloc::ConcreteInt>()) {
1647       int64_t i = CI->getValue().getSExtValue();
1648       // Abort on string underrun.  This can be possible by arbitrary
1649       // clients of getBindingForElement().
1650       if (i < 0)
1651         return UndefinedVal();
1652       int64_t length = Str->getLength();
1653       // Technically, only i == length is guaranteed to be null.
1654       // However, such overflows should be caught before reaching this point;
1655       // the only time such an access would be made is if a string literal was
1656       // used to initialize a larger array.
1657       char c = (i >= length) ? '\0' : Str->getCodeUnit(i);
1658       return svalBuilder.makeIntVal(c, T);
1659     }
1660   } else if (const VarRegion *VR = dyn_cast<VarRegion>(superR)) {
1661     // Check if the containing array has an initialized value that we can trust.
1662     // We can trust a const value or a value of a global initializer in main().
1663     const VarDecl *VD = VR->getDecl();
1664     if (VD->getType().isConstQualified() ||
1665         R->getElementType().isConstQualified() ||
1666         (B.isMainAnalysis() && VD->hasGlobalStorage())) {
1667       if (const Expr *Init = VD->getAnyInitializer()) {
1668         if (const auto *InitList = dyn_cast<InitListExpr>(Init)) {
1669           // The array index has to be known.
1670           if (auto CI = R->getIndex().getAs<nonloc::ConcreteInt>()) {
1671             // If it is not an array, return Undef.
1672             QualType T = VD->getType();
1673             const ConstantArrayType *CAT = Ctx.getAsConstantArrayType(T);
1674             if (!CAT)
1675               return UndefinedVal();
1676 
1677             // Support one-dimensional array.
1678             // C++20 [expr.add] 7.6.6.4 (excerpt):
1679             //   If P points to an array element i of an array object x with n
1680             //   elements, where i < 0 or i > n, the behavior is undefined.
1681             //   Dereferencing is not allowed on the "one past the last
1682             //   element", when i == n.
1683             // Example:
1684             //   const int arr[4] = {1, 2};
1685             //   const int *ptr = arr;
1686             //   int x0 = ptr[0]; // 1
1687             //   int x1 = ptr[1]; // 2
1688             //   int x2 = ptr[2]; // 0
1689             //   int x3 = ptr[3]; // 0
1690             //   int x4 = ptr[4]; // UB
1691             // TODO: Support multidimensional array.
1692             if (!isa<ConstantArrayType>(CAT->getElementType())) {
1693               // One-dimensional array.
1694               const llvm::APSInt &Idx = CI->getValue();
1695               const auto I = static_cast<uint64_t>(Idx.getExtValue());
1696               // Use `getZExtValue` because array extent can not be negative.
1697               const uint64_t Extent = CAT->getSize().getZExtValue();
1698               // Check for `Idx < 0`, NOT for `I < 0`, because `Idx` CAN be
1699               // negative, but `I` can NOT.
1700               if (Idx < 0 || I >= Extent)
1701                 return UndefinedVal();
1702 
1703               // C++20 [expr.add] 9.4.17.5 (excerpt):
1704               //   i-th array element is value-initialized for each k < i ≤ n,
1705               //   where k is an expression-list size and n is an array extent.
1706               if (I >= InitList->getNumInits())
1707                 return svalBuilder.makeZeroVal(R->getElementType());
1708 
1709               // Return a constant value, if it is presented.
1710               // FIXME: Support other SVals.
1711               const Expr *E = InitList->getInit(I);
1712               if (Optional<SVal> V = svalBuilder.getConstantVal(E))
1713                 return *V;
1714             }
1715           }
1716         }
1717       }
1718     }
1719   }
1720 
1721   // Check for loads from a code text region.  For such loads, just give up.
1722   if (isa<CodeTextRegion>(superR))
1723     return UnknownVal();
1724 
1725   // Handle the case where we are indexing into a larger scalar object.
1726   // For example, this handles:
1727   //   int x = ...
1728   //   char *y = &x;
1729   //   return *y;
1730   // FIXME: This is a hack, and doesn't do anything really intelligent yet.
1731   const RegionRawOffset &O = R->getAsArrayOffset();
1732 
1733   // If we cannot reason about the offset, return an unknown value.
1734   if (!O.getRegion())
1735     return UnknownVal();
1736 
1737   if (const TypedValueRegion *baseR =
1738         dyn_cast_or_null<TypedValueRegion>(O.getRegion())) {
1739     QualType baseT = baseR->getValueType();
1740     if (baseT->isScalarType()) {
1741       QualType elemT = R->getElementType();
1742       if (elemT->isScalarType()) {
1743         if (Ctx.getTypeSizeInChars(baseT) >= Ctx.getTypeSizeInChars(elemT)) {
1744           if (const Optional<SVal> &V = B.getDirectBinding(superR)) {
1745             if (SymbolRef parentSym = V->getAsSymbol())
1746               return svalBuilder.getDerivedRegionValueSymbolVal(parentSym, R);
1747 
1748             if (V->isUnknownOrUndef())
1749               return *V;
1750             // Other cases: give up.  We are indexing into a larger object
1751             // that has some value, but we don't know how to handle that yet.
1752             return UnknownVal();
1753           }
1754         }
1755       }
1756     }
1757   }
1758   return getBindingForFieldOrElementCommon(B, R, R->getElementType());
1759 }
1760 
1761 SVal RegionStoreManager::getBindingForField(RegionBindingsConstRef B,
1762                                             const FieldRegion* R) {
1763 
1764   // Check if the region has a binding.
1765   if (const Optional<SVal> &V = B.getDirectBinding(R))
1766     return *V;
1767 
1768   // Is the field declared constant and has an in-class initializer?
1769   const FieldDecl *FD = R->getDecl();
1770   QualType Ty = FD->getType();
1771   if (Ty.isConstQualified())
1772     if (const Expr *Init = FD->getInClassInitializer())
1773       if (Optional<SVal> V = svalBuilder.getConstantVal(Init))
1774         return *V;
1775 
1776   // If the containing record was initialized, try to get its constant value.
1777   const MemRegion* superR = R->getSuperRegion();
1778   if (const auto *VR = dyn_cast<VarRegion>(superR)) {
1779     const VarDecl *VD = VR->getDecl();
1780     QualType RecordVarTy = VD->getType();
1781     unsigned Index = FD->getFieldIndex();
1782     // Either the record variable or the field has an initializer that we can
1783     // trust. We trust initializers of constants and, additionally, respect
1784     // initializers of globals when analyzing main().
1785     if (RecordVarTy.isConstQualified() || Ty.isConstQualified() ||
1786         (B.isMainAnalysis() && VD->hasGlobalStorage()))
1787       if (const Expr *Init = VD->getAnyInitializer())
1788         if (const auto *InitList = dyn_cast<InitListExpr>(Init)) {
1789           if (Index < InitList->getNumInits()) {
1790             if (const Expr *FieldInit = InitList->getInit(Index))
1791               if (Optional<SVal> V = svalBuilder.getConstantVal(FieldInit))
1792                 return *V;
1793           } else {
1794             return svalBuilder.makeZeroVal(Ty);
1795           }
1796         }
1797   }
1798 
1799   return getBindingForFieldOrElementCommon(B, R, Ty);
1800 }
1801 
1802 Optional<SVal>
1803 RegionStoreManager::getBindingForDerivedDefaultValue(RegionBindingsConstRef B,
1804                                                      const MemRegion *superR,
1805                                                      const TypedValueRegion *R,
1806                                                      QualType Ty) {
1807 
1808   if (const Optional<SVal> &D = B.getDefaultBinding(superR)) {
1809     const SVal &val = D.getValue();
1810     if (SymbolRef parentSym = val.getAsSymbol())
1811       return svalBuilder.getDerivedRegionValueSymbolVal(parentSym, R);
1812 
1813     if (val.isZeroConstant())
1814       return svalBuilder.makeZeroVal(Ty);
1815 
1816     if (val.isUnknownOrUndef())
1817       return val;
1818 
1819     // Lazy bindings are usually handled through getExistingLazyBinding().
1820     // We should unify these two code paths at some point.
1821     if (val.getAs<nonloc::LazyCompoundVal>() ||
1822         val.getAs<nonloc::CompoundVal>())
1823       return val;
1824 
1825     llvm_unreachable("Unknown default value");
1826   }
1827 
1828   return None;
1829 }
1830 
1831 SVal RegionStoreManager::getLazyBinding(const SubRegion *LazyBindingRegion,
1832                                         RegionBindingsRef LazyBinding) {
1833   SVal Result;
1834   if (const ElementRegion *ER = dyn_cast<ElementRegion>(LazyBindingRegion))
1835     Result = getBindingForElement(LazyBinding, ER);
1836   else
1837     Result = getBindingForField(LazyBinding,
1838                                 cast<FieldRegion>(LazyBindingRegion));
1839 
1840   // FIXME: This is a hack to deal with RegionStore's inability to distinguish a
1841   // default value for /part/ of an aggregate from a default value for the
1842   // /entire/ aggregate. The most common case of this is when struct Outer
1843   // has as its first member a struct Inner, which is copied in from a stack
1844   // variable. In this case, even if the Outer's default value is symbolic, 0,
1845   // or unknown, it gets overridden by the Inner's default value of undefined.
1846   //
1847   // This is a general problem -- if the Inner is zero-initialized, the Outer
1848   // will now look zero-initialized. The proper way to solve this is with a
1849   // new version of RegionStore that tracks the extent of a binding as well
1850   // as the offset.
1851   //
1852   // This hack only takes care of the undefined case because that can very
1853   // quickly result in a warning.
1854   if (Result.isUndef())
1855     Result = UnknownVal();
1856 
1857   return Result;
1858 }
1859 
1860 SVal
1861 RegionStoreManager::getBindingForFieldOrElementCommon(RegionBindingsConstRef B,
1862                                                       const TypedValueRegion *R,
1863                                                       QualType Ty) {
1864 
1865   // At this point we have already checked in either getBindingForElement or
1866   // getBindingForField if 'R' has a direct binding.
1867 
1868   // Lazy binding?
1869   Store lazyBindingStore = nullptr;
1870   const SubRegion *lazyBindingRegion = nullptr;
1871   std::tie(lazyBindingStore, lazyBindingRegion) = findLazyBinding(B, R, R);
1872   if (lazyBindingRegion)
1873     return getLazyBinding(lazyBindingRegion,
1874                           getRegionBindings(lazyBindingStore));
1875 
1876   // Record whether or not we see a symbolic index.  That can completely
1877   // be out of scope of our lookup.
1878   bool hasSymbolicIndex = false;
1879 
1880   // FIXME: This is a hack to deal with RegionStore's inability to distinguish a
1881   // default value for /part/ of an aggregate from a default value for the
1882   // /entire/ aggregate. The most common case of this is when struct Outer
1883   // has as its first member a struct Inner, which is copied in from a stack
1884   // variable. In this case, even if the Outer's default value is symbolic, 0,
1885   // or unknown, it gets overridden by the Inner's default value of undefined.
1886   //
1887   // This is a general problem -- if the Inner is zero-initialized, the Outer
1888   // will now look zero-initialized. The proper way to solve this is with a
1889   // new version of RegionStore that tracks the extent of a binding as well
1890   // as the offset.
1891   //
1892   // This hack only takes care of the undefined case because that can very
1893   // quickly result in a warning.
1894   bool hasPartialLazyBinding = false;
1895 
1896   const SubRegion *SR = R;
1897   while (SR) {
1898     const MemRegion *Base = SR->getSuperRegion();
1899     if (Optional<SVal> D = getBindingForDerivedDefaultValue(B, Base, R, Ty)) {
1900       if (D->getAs<nonloc::LazyCompoundVal>()) {
1901         hasPartialLazyBinding = true;
1902         break;
1903       }
1904 
1905       return *D;
1906     }
1907 
1908     if (const ElementRegion *ER = dyn_cast<ElementRegion>(Base)) {
1909       NonLoc index = ER->getIndex();
1910       if (!index.isConstant())
1911         hasSymbolicIndex = true;
1912     }
1913 
1914     // If our super region is a field or element itself, walk up the region
1915     // hierarchy to see if there is a default value installed in an ancestor.
1916     SR = dyn_cast<SubRegion>(Base);
1917   }
1918 
1919   if (R->hasStackNonParametersStorage()) {
1920     if (isa<ElementRegion>(R)) {
1921       // Currently we don't reason specially about Clang-style vectors.  Check
1922       // if superR is a vector and if so return Unknown.
1923       if (const TypedValueRegion *typedSuperR =
1924             dyn_cast<TypedValueRegion>(R->getSuperRegion())) {
1925         if (typedSuperR->getValueType()->isVectorType())
1926           return UnknownVal();
1927       }
1928     }
1929 
1930     // FIXME: We also need to take ElementRegions with symbolic indexes into
1931     // account.  This case handles both directly accessing an ElementRegion
1932     // with a symbolic offset, but also fields within an element with
1933     // a symbolic offset.
1934     if (hasSymbolicIndex)
1935       return UnknownVal();
1936 
1937     // Additionally allow introspection of a block's internal layout.
1938     if (!hasPartialLazyBinding && !isa<BlockDataRegion>(R->getBaseRegion()))
1939       return UndefinedVal();
1940   }
1941 
1942   // All other values are symbolic.
1943   return svalBuilder.getRegionValueSymbolVal(R);
1944 }
1945 
1946 SVal RegionStoreManager::getBindingForObjCIvar(RegionBindingsConstRef B,
1947                                                const ObjCIvarRegion* R) {
1948   // Check if the region has a binding.
1949   if (const Optional<SVal> &V = B.getDirectBinding(R))
1950     return *V;
1951 
1952   const MemRegion *superR = R->getSuperRegion();
1953 
1954   // Check if the super region has a default binding.
1955   if (const Optional<SVal> &V = B.getDefaultBinding(superR)) {
1956     if (SymbolRef parentSym = V->getAsSymbol())
1957       return svalBuilder.getDerivedRegionValueSymbolVal(parentSym, R);
1958 
1959     // Other cases: give up.
1960     return UnknownVal();
1961   }
1962 
1963   return getBindingForLazySymbol(R);
1964 }
1965 
1966 SVal RegionStoreManager::getBindingForVar(RegionBindingsConstRef B,
1967                                           const VarRegion *R) {
1968 
1969   // Check if the region has a binding.
1970   if (Optional<SVal> V = B.getDirectBinding(R))
1971     return *V;
1972 
1973   if (Optional<SVal> V = B.getDefaultBinding(R))
1974     return *V;
1975 
1976   // Lazily derive a value for the VarRegion.
1977   const VarDecl *VD = R->getDecl();
1978   const MemSpaceRegion *MS = R->getMemorySpace();
1979 
1980   // Arguments are always symbolic.
1981   if (isa<StackArgumentsSpaceRegion>(MS))
1982     return svalBuilder.getRegionValueSymbolVal(R);
1983 
1984   // Is 'VD' declared constant?  If so, retrieve the constant value.
1985   if (VD->getType().isConstQualified()) {
1986     if (const Expr *Init = VD->getAnyInitializer()) {
1987       if (Optional<SVal> V = svalBuilder.getConstantVal(Init))
1988         return *V;
1989 
1990       // If the variable is const qualified and has an initializer but
1991       // we couldn't evaluate initializer to a value, treat the value as
1992       // unknown.
1993       return UnknownVal();
1994     }
1995   }
1996 
1997   // This must come after the check for constants because closure-captured
1998   // constant variables may appear in UnknownSpaceRegion.
1999   if (isa<UnknownSpaceRegion>(MS))
2000     return svalBuilder.getRegionValueSymbolVal(R);
2001 
2002   if (isa<GlobalsSpaceRegion>(MS)) {
2003     QualType T = VD->getType();
2004 
2005     // If we're in main(), then global initializers have not become stale yet.
2006     if (B.isMainAnalysis())
2007       if (const Expr *Init = VD->getAnyInitializer())
2008         if (Optional<SVal> V = svalBuilder.getConstantVal(Init))
2009           return *V;
2010 
2011     // Function-scoped static variables are default-initialized to 0; if they
2012     // have an initializer, it would have been processed by now.
2013     // FIXME: This is only true when we're starting analysis from main().
2014     // We're losing a lot of coverage here.
2015     if (isa<StaticGlobalSpaceRegion>(MS))
2016       return svalBuilder.makeZeroVal(T);
2017 
2018     if (Optional<SVal> V = getBindingForDerivedDefaultValue(B, MS, R, T)) {
2019       assert(!V->getAs<nonloc::LazyCompoundVal>());
2020       return V.getValue();
2021     }
2022 
2023     return svalBuilder.getRegionValueSymbolVal(R);
2024   }
2025 
2026   return UndefinedVal();
2027 }
2028 
2029 SVal RegionStoreManager::getBindingForLazySymbol(const TypedValueRegion *R) {
2030   // All other values are symbolic.
2031   return svalBuilder.getRegionValueSymbolVal(R);
2032 }
2033 
2034 const RegionStoreManager::SValListTy &
2035 RegionStoreManager::getInterestingValues(nonloc::LazyCompoundVal LCV) {
2036   // First, check the cache.
2037   LazyBindingsMapTy::iterator I = LazyBindingsMap.find(LCV.getCVData());
2038   if (I != LazyBindingsMap.end())
2039     return I->second;
2040 
2041   // If we don't have a list of values cached, start constructing it.
2042   SValListTy List;
2043 
2044   const SubRegion *LazyR = LCV.getRegion();
2045   RegionBindingsRef B = getRegionBindings(LCV.getStore());
2046 
2047   // If this region had /no/ bindings at the time, there are no interesting
2048   // values to return.
2049   const ClusterBindings *Cluster = B.lookup(LazyR->getBaseRegion());
2050   if (!Cluster)
2051     return (LazyBindingsMap[LCV.getCVData()] = std::move(List));
2052 
2053   SmallVector<BindingPair, 32> Bindings;
2054   collectSubRegionBindings(Bindings, svalBuilder, *Cluster, LazyR,
2055                            /*IncludeAllDefaultBindings=*/true);
2056   for (SmallVectorImpl<BindingPair>::const_iterator I = Bindings.begin(),
2057                                                     E = Bindings.end();
2058        I != E; ++I) {
2059     SVal V = I->second;
2060     if (V.isUnknownOrUndef() || V.isConstant())
2061       continue;
2062 
2063     if (Optional<nonloc::LazyCompoundVal> InnerLCV =
2064             V.getAs<nonloc::LazyCompoundVal>()) {
2065       const SValListTy &InnerList = getInterestingValues(*InnerLCV);
2066       List.insert(List.end(), InnerList.begin(), InnerList.end());
2067       continue;
2068     }
2069 
2070     List.push_back(V);
2071   }
2072 
2073   return (LazyBindingsMap[LCV.getCVData()] = std::move(List));
2074 }
2075 
2076 NonLoc RegionStoreManager::createLazyBinding(RegionBindingsConstRef B,
2077                                              const TypedValueRegion *R) {
2078   if (Optional<nonloc::LazyCompoundVal> V =
2079         getExistingLazyBinding(svalBuilder, B, R, false))
2080     return *V;
2081 
2082   return svalBuilder.makeLazyCompoundVal(StoreRef(B.asStore(), *this), R);
2083 }
2084 
2085 static bool isRecordEmpty(const RecordDecl *RD) {
2086   if (!RD->field_empty())
2087     return false;
2088   if (const CXXRecordDecl *CRD = dyn_cast<CXXRecordDecl>(RD))
2089     return CRD->getNumBases() == 0;
2090   return true;
2091 }
2092 
2093 SVal RegionStoreManager::getBindingForStruct(RegionBindingsConstRef B,
2094                                              const TypedValueRegion *R) {
2095   const RecordDecl *RD = R->getValueType()->castAs<RecordType>()->getDecl();
2096   if (!RD->getDefinition() || isRecordEmpty(RD))
2097     return UnknownVal();
2098 
2099   return createLazyBinding(B, R);
2100 }
2101 
2102 SVal RegionStoreManager::getBindingForArray(RegionBindingsConstRef B,
2103                                             const TypedValueRegion *R) {
2104   assert(Ctx.getAsConstantArrayType(R->getValueType()) &&
2105          "Only constant array types can have compound bindings.");
2106 
2107   return createLazyBinding(B, R);
2108 }
2109 
2110 bool RegionStoreManager::includedInBindings(Store store,
2111                                             const MemRegion *region) const {
2112   RegionBindingsRef B = getRegionBindings(store);
2113   region = region->getBaseRegion();
2114 
2115   // Quick path: if the base is the head of a cluster, the region is live.
2116   if (B.lookup(region))
2117     return true;
2118 
2119   // Slow path: if the region is the VALUE of any binding, it is live.
2120   for (RegionBindingsRef::iterator RI = B.begin(), RE = B.end(); RI != RE; ++RI) {
2121     const ClusterBindings &Cluster = RI.getData();
2122     for (ClusterBindings::iterator CI = Cluster.begin(), CE = Cluster.end();
2123          CI != CE; ++CI) {
2124       const SVal &D = CI.getData();
2125       if (const MemRegion *R = D.getAsRegion())
2126         if (R->getBaseRegion() == region)
2127           return true;
2128     }
2129   }
2130 
2131   return false;
2132 }
2133 
2134 //===----------------------------------------------------------------------===//
2135 // Binding values to regions.
2136 //===----------------------------------------------------------------------===//
2137 
2138 StoreRef RegionStoreManager::killBinding(Store ST, Loc L) {
2139   if (Optional<loc::MemRegionVal> LV = L.getAs<loc::MemRegionVal>())
2140     if (const MemRegion* R = LV->getRegion())
2141       return StoreRef(getRegionBindings(ST).removeBinding(R)
2142                                            .asImmutableMap()
2143                                            .getRootWithoutRetain(),
2144                       *this);
2145 
2146   return StoreRef(ST, *this);
2147 }
2148 
2149 RegionBindingsRef
2150 RegionStoreManager::bind(RegionBindingsConstRef B, Loc L, SVal V) {
2151   if (L.getAs<loc::ConcreteInt>())
2152     return B;
2153 
2154   // If we get here, the location should be a region.
2155   const MemRegion *R = L.castAs<loc::MemRegionVal>().getRegion();
2156 
2157   // Check if the region is a struct region.
2158   if (const TypedValueRegion* TR = dyn_cast<TypedValueRegion>(R)) {
2159     QualType Ty = TR->getValueType();
2160     if (Ty->isArrayType())
2161       return bindArray(B, TR, V);
2162     if (Ty->isStructureOrClassType())
2163       return bindStruct(B, TR, V);
2164     if (Ty->isVectorType())
2165       return bindVector(B, TR, V);
2166     if (Ty->isUnionType())
2167       return bindAggregate(B, TR, V);
2168   }
2169 
2170   if (const SymbolicRegion *SR = dyn_cast<SymbolicRegion>(R)) {
2171     // Binding directly to a symbolic region should be treated as binding
2172     // to element 0.
2173     QualType T = SR->getSymbol()->getType();
2174     if (T->isAnyPointerType() || T->isReferenceType())
2175       T = T->getPointeeType();
2176 
2177     R = GetElementZeroRegion(SR, T);
2178   }
2179 
2180   assert((!isa<CXXThisRegion>(R) || !B.lookup(R)) &&
2181          "'this' pointer is not an l-value and is not assignable");
2182 
2183   // Clear out bindings that may overlap with this binding.
2184   RegionBindingsRef NewB = removeSubRegionBindings(B, cast<SubRegion>(R));
2185   return NewB.addBinding(BindingKey::Make(R, BindingKey::Direct), V);
2186 }
2187 
2188 RegionBindingsRef
2189 RegionStoreManager::setImplicitDefaultValue(RegionBindingsConstRef B,
2190                                             const MemRegion *R,
2191                                             QualType T) {
2192   SVal V;
2193 
2194   if (Loc::isLocType(T))
2195     V = svalBuilder.makeNull();
2196   else if (T->isIntegralOrEnumerationType())
2197     V = svalBuilder.makeZeroVal(T);
2198   else if (T->isStructureOrClassType() || T->isArrayType()) {
2199     // Set the default value to a zero constant when it is a structure
2200     // or array.  The type doesn't really matter.
2201     V = svalBuilder.makeZeroVal(Ctx.IntTy);
2202   }
2203   else {
2204     // We can't represent values of this type, but we still need to set a value
2205     // to record that the region has been initialized.
2206     // If this assertion ever fires, a new case should be added above -- we
2207     // should know how to default-initialize any value we can symbolicate.
2208     assert(!SymbolManager::canSymbolicate(T) && "This type is representable");
2209     V = UnknownVal();
2210   }
2211 
2212   return B.addBinding(R, BindingKey::Default, V);
2213 }
2214 
2215 RegionBindingsRef
2216 RegionStoreManager::bindArray(RegionBindingsConstRef B,
2217                               const TypedValueRegion* R,
2218                               SVal Init) {
2219 
2220   const ArrayType *AT =cast<ArrayType>(Ctx.getCanonicalType(R->getValueType()));
2221   QualType ElementTy = AT->getElementType();
2222   Optional<uint64_t> Size;
2223 
2224   if (const ConstantArrayType* CAT = dyn_cast<ConstantArrayType>(AT))
2225     Size = CAT->getSize().getZExtValue();
2226 
2227   // Check if the init expr is a literal. If so, bind the rvalue instead.
2228   // FIXME: It's not responsibility of the Store to transform this lvalue
2229   // to rvalue. ExprEngine or maybe even CFG should do this before binding.
2230   if (Optional<loc::MemRegionVal> MRV = Init.getAs<loc::MemRegionVal>()) {
2231     SVal V = getBinding(B.asStore(), *MRV, R->getValueType());
2232     return bindAggregate(B, R, V);
2233   }
2234 
2235   // Handle lazy compound values.
2236   if (Init.getAs<nonloc::LazyCompoundVal>())
2237     return bindAggregate(B, R, Init);
2238 
2239   if (Init.isUnknown())
2240     return bindAggregate(B, R, UnknownVal());
2241 
2242   // Remaining case: explicit compound values.
2243   const nonloc::CompoundVal& CV = Init.castAs<nonloc::CompoundVal>();
2244   nonloc::CompoundVal::iterator VI = CV.begin(), VE = CV.end();
2245   uint64_t i = 0;
2246 
2247   RegionBindingsRef NewB(B);
2248 
2249   for (; Size.hasValue() ? i < Size.getValue() : true ; ++i, ++VI) {
2250     // The init list might be shorter than the array length.
2251     if (VI == VE)
2252       break;
2253 
2254     const NonLoc &Idx = svalBuilder.makeArrayIndex(i);
2255     const ElementRegion *ER = MRMgr.getElementRegion(ElementTy, Idx, R, Ctx);
2256 
2257     if (ElementTy->isStructureOrClassType())
2258       NewB = bindStruct(NewB, ER, *VI);
2259     else if (ElementTy->isArrayType())
2260       NewB = bindArray(NewB, ER, *VI);
2261     else
2262       NewB = bind(NewB, loc::MemRegionVal(ER), *VI);
2263   }
2264 
2265   // If the init list is shorter than the array length (or the array has
2266   // variable length), set the array default value. Values that are already set
2267   // are not overwritten.
2268   if (!Size.hasValue() || i < Size.getValue())
2269     NewB = setImplicitDefaultValue(NewB, R, ElementTy);
2270 
2271   return NewB;
2272 }
2273 
2274 RegionBindingsRef RegionStoreManager::bindVector(RegionBindingsConstRef B,
2275                                                  const TypedValueRegion* R,
2276                                                  SVal V) {
2277   QualType T = R->getValueType();
2278   const VectorType *VT = T->castAs<VectorType>(); // Use castAs for typedefs.
2279 
2280   // Handle lazy compound values and symbolic values.
2281   if (V.getAs<nonloc::LazyCompoundVal>() || V.getAs<nonloc::SymbolVal>())
2282     return bindAggregate(B, R, V);
2283 
2284   // We may get non-CompoundVal accidentally due to imprecise cast logic or
2285   // that we are binding symbolic struct value. Kill the field values, and if
2286   // the value is symbolic go and bind it as a "default" binding.
2287   if (!V.getAs<nonloc::CompoundVal>()) {
2288     return bindAggregate(B, R, UnknownVal());
2289   }
2290 
2291   QualType ElemType = VT->getElementType();
2292   nonloc::CompoundVal CV = V.castAs<nonloc::CompoundVal>();
2293   nonloc::CompoundVal::iterator VI = CV.begin(), VE = CV.end();
2294   unsigned index = 0, numElements = VT->getNumElements();
2295   RegionBindingsRef NewB(B);
2296 
2297   for ( ; index != numElements ; ++index) {
2298     if (VI == VE)
2299       break;
2300 
2301     NonLoc Idx = svalBuilder.makeArrayIndex(index);
2302     const ElementRegion *ER = MRMgr.getElementRegion(ElemType, Idx, R, Ctx);
2303 
2304     if (ElemType->isArrayType())
2305       NewB = bindArray(NewB, ER, *VI);
2306     else if (ElemType->isStructureOrClassType())
2307       NewB = bindStruct(NewB, ER, *VI);
2308     else
2309       NewB = bind(NewB, loc::MemRegionVal(ER), *VI);
2310   }
2311   return NewB;
2312 }
2313 
2314 Optional<RegionBindingsRef>
2315 RegionStoreManager::tryBindSmallStruct(RegionBindingsConstRef B,
2316                                        const TypedValueRegion *R,
2317                                        const RecordDecl *RD,
2318                                        nonloc::LazyCompoundVal LCV) {
2319   FieldVector Fields;
2320 
2321   if (const CXXRecordDecl *Class = dyn_cast<CXXRecordDecl>(RD))
2322     if (Class->getNumBases() != 0 || Class->getNumVBases() != 0)
2323       return None;
2324 
2325   for (const auto *FD : RD->fields()) {
2326     if (FD->isUnnamedBitfield())
2327       continue;
2328 
2329     // If there are too many fields, or if any of the fields are aggregates,
2330     // just use the LCV as a default binding.
2331     if (Fields.size() == SmallStructLimit)
2332       return None;
2333 
2334     QualType Ty = FD->getType();
2335     if (!(Ty->isScalarType() || Ty->isReferenceType()))
2336       return None;
2337 
2338     Fields.push_back(FD);
2339   }
2340 
2341   RegionBindingsRef NewB = B;
2342 
2343   for (FieldVector::iterator I = Fields.begin(), E = Fields.end(); I != E; ++I){
2344     const FieldRegion *SourceFR = MRMgr.getFieldRegion(*I, LCV.getRegion());
2345     SVal V = getBindingForField(getRegionBindings(LCV.getStore()), SourceFR);
2346 
2347     const FieldRegion *DestFR = MRMgr.getFieldRegion(*I, R);
2348     NewB = bind(NewB, loc::MemRegionVal(DestFR), V);
2349   }
2350 
2351   return NewB;
2352 }
2353 
2354 RegionBindingsRef RegionStoreManager::bindStruct(RegionBindingsConstRef B,
2355                                                  const TypedValueRegion* R,
2356                                                  SVal V) {
2357   if (!Features.supportsFields())
2358     return B;
2359 
2360   QualType T = R->getValueType();
2361   assert(T->isStructureOrClassType());
2362 
2363   const RecordType* RT = T->castAs<RecordType>();
2364   const RecordDecl *RD = RT->getDecl();
2365 
2366   if (!RD->isCompleteDefinition())
2367     return B;
2368 
2369   // Handle lazy compound values and symbolic values.
2370   if (Optional<nonloc::LazyCompoundVal> LCV =
2371         V.getAs<nonloc::LazyCompoundVal>()) {
2372     if (Optional<RegionBindingsRef> NewB = tryBindSmallStruct(B, R, RD, *LCV))
2373       return *NewB;
2374     return bindAggregate(B, R, V);
2375   }
2376   if (V.getAs<nonloc::SymbolVal>())
2377     return bindAggregate(B, R, V);
2378 
2379   // We may get non-CompoundVal accidentally due to imprecise cast logic or
2380   // that we are binding symbolic struct value. Kill the field values, and if
2381   // the value is symbolic go and bind it as a "default" binding.
2382   if (V.isUnknown() || !V.getAs<nonloc::CompoundVal>())
2383     return bindAggregate(B, R, UnknownVal());
2384 
2385   // The raw CompoundVal is essentially a symbolic InitListExpr: an (immutable)
2386   // list of other values. It appears pretty much only when there's an actual
2387   // initializer list expression in the program, and the analyzer tries to
2388   // unwrap it as soon as possible.
2389   // This code is where such unwrap happens: when the compound value is put into
2390   // the object that it was supposed to initialize (it's an *initializer* list,
2391   // after all), instead of binding the whole value to the whole object, we bind
2392   // sub-values to sub-objects. Sub-values may themselves be compound values,
2393   // and in this case the procedure becomes recursive.
2394   // FIXME: The annoying part about compound values is that they don't carry
2395   // any sort of information about which value corresponds to which sub-object.
2396   // It's simply a list of values in the middle of nowhere; we expect to match
2397   // them to sub-objects, essentially, "by index": first value binds to
2398   // the first field, second value binds to the second field, etc.
2399   // It would have been much safer to organize non-lazy compound values as
2400   // a mapping from fields/bases to values.
2401   const nonloc::CompoundVal& CV = V.castAs<nonloc::CompoundVal>();
2402   nonloc::CompoundVal::iterator VI = CV.begin(), VE = CV.end();
2403 
2404   RegionBindingsRef NewB(B);
2405 
2406   // In C++17 aggregates may have base classes, handle those as well.
2407   // They appear before fields in the initializer list / compound value.
2408   if (const auto *CRD = dyn_cast<CXXRecordDecl>(RD)) {
2409     // If the object was constructed with a constructor, its value is a
2410     // LazyCompoundVal. If it's a raw CompoundVal, it means that we're
2411     // performing aggregate initialization. The only exception from this
2412     // rule is sending an Objective-C++ message that returns a C++ object
2413     // to a nil receiver; in this case the semantics is to return a
2414     // zero-initialized object even if it's a C++ object that doesn't have
2415     // this sort of constructor; the CompoundVal is empty in this case.
2416     assert((CRD->isAggregate() || (Ctx.getLangOpts().ObjC && VI == VE)) &&
2417            "Non-aggregates are constructed with a constructor!");
2418 
2419     for (const auto &B : CRD->bases()) {
2420       // (Multiple inheritance is fine though.)
2421       assert(!B.isVirtual() && "Aggregates cannot have virtual base classes!");
2422 
2423       if (VI == VE)
2424         break;
2425 
2426       QualType BTy = B.getType();
2427       assert(BTy->isStructureOrClassType() && "Base classes must be classes!");
2428 
2429       const CXXRecordDecl *BRD = BTy->getAsCXXRecordDecl();
2430       assert(BRD && "Base classes must be C++ classes!");
2431 
2432       const CXXBaseObjectRegion *BR =
2433           MRMgr.getCXXBaseObjectRegion(BRD, R, /*IsVirtual=*/false);
2434 
2435       NewB = bindStruct(NewB, BR, *VI);
2436 
2437       ++VI;
2438     }
2439   }
2440 
2441   RecordDecl::field_iterator FI, FE;
2442 
2443   for (FI = RD->field_begin(), FE = RD->field_end(); FI != FE; ++FI) {
2444 
2445     if (VI == VE)
2446       break;
2447 
2448     // Skip any unnamed bitfields to stay in sync with the initializers.
2449     if (FI->isUnnamedBitfield())
2450       continue;
2451 
2452     QualType FTy = FI->getType();
2453     const FieldRegion* FR = MRMgr.getFieldRegion(*FI, R);
2454 
2455     if (FTy->isArrayType())
2456       NewB = bindArray(NewB, FR, *VI);
2457     else if (FTy->isStructureOrClassType())
2458       NewB = bindStruct(NewB, FR, *VI);
2459     else
2460       NewB = bind(NewB, loc::MemRegionVal(FR), *VI);
2461     ++VI;
2462   }
2463 
2464   // There may be fewer values in the initialize list than the fields of struct.
2465   if (FI != FE) {
2466     NewB = NewB.addBinding(R, BindingKey::Default,
2467                            svalBuilder.makeIntVal(0, false));
2468   }
2469 
2470   return NewB;
2471 }
2472 
2473 RegionBindingsRef
2474 RegionStoreManager::bindAggregate(RegionBindingsConstRef B,
2475                                   const TypedRegion *R,
2476                                   SVal Val) {
2477   // Remove the old bindings, using 'R' as the root of all regions
2478   // we will invalidate. Then add the new binding.
2479   return removeSubRegionBindings(B, R).addBinding(R, BindingKey::Default, Val);
2480 }
2481 
2482 //===----------------------------------------------------------------------===//
2483 // State pruning.
2484 //===----------------------------------------------------------------------===//
2485 
2486 namespace {
2487 class RemoveDeadBindingsWorker
2488     : public ClusterAnalysis<RemoveDeadBindingsWorker> {
2489   SmallVector<const SymbolicRegion *, 12> Postponed;
2490   SymbolReaper &SymReaper;
2491   const StackFrameContext *CurrentLCtx;
2492 
2493 public:
2494   RemoveDeadBindingsWorker(RegionStoreManager &rm,
2495                            ProgramStateManager &stateMgr,
2496                            RegionBindingsRef b, SymbolReaper &symReaper,
2497                            const StackFrameContext *LCtx)
2498     : ClusterAnalysis<RemoveDeadBindingsWorker>(rm, stateMgr, b),
2499       SymReaper(symReaper), CurrentLCtx(LCtx) {}
2500 
2501   // Called by ClusterAnalysis.
2502   void VisitAddedToCluster(const MemRegion *baseR, const ClusterBindings &C);
2503   void VisitCluster(const MemRegion *baseR, const ClusterBindings *C);
2504   using ClusterAnalysis<RemoveDeadBindingsWorker>::VisitCluster;
2505 
2506   using ClusterAnalysis::AddToWorkList;
2507 
2508   bool AddToWorkList(const MemRegion *R);
2509 
2510   bool UpdatePostponed();
2511   void VisitBinding(SVal V);
2512 };
2513 }
2514 
2515 bool RemoveDeadBindingsWorker::AddToWorkList(const MemRegion *R) {
2516   const MemRegion *BaseR = R->getBaseRegion();
2517   return AddToWorkList(WorkListElement(BaseR), getCluster(BaseR));
2518 }
2519 
2520 void RemoveDeadBindingsWorker::VisitAddedToCluster(const MemRegion *baseR,
2521                                                    const ClusterBindings &C) {
2522 
2523   if (const VarRegion *VR = dyn_cast<VarRegion>(baseR)) {
2524     if (SymReaper.isLive(VR))
2525       AddToWorkList(baseR, &C);
2526 
2527     return;
2528   }
2529 
2530   if (const SymbolicRegion *SR = dyn_cast<SymbolicRegion>(baseR)) {
2531     if (SymReaper.isLive(SR->getSymbol()))
2532       AddToWorkList(SR, &C);
2533     else
2534       Postponed.push_back(SR);
2535 
2536     return;
2537   }
2538 
2539   if (isa<NonStaticGlobalSpaceRegion>(baseR)) {
2540     AddToWorkList(baseR, &C);
2541     return;
2542   }
2543 
2544   // CXXThisRegion in the current or parent location context is live.
2545   if (const CXXThisRegion *TR = dyn_cast<CXXThisRegion>(baseR)) {
2546     const auto *StackReg =
2547         cast<StackArgumentsSpaceRegion>(TR->getSuperRegion());
2548     const StackFrameContext *RegCtx = StackReg->getStackFrame();
2549     if (CurrentLCtx &&
2550         (RegCtx == CurrentLCtx || RegCtx->isParentOf(CurrentLCtx)))
2551       AddToWorkList(TR, &C);
2552   }
2553 }
2554 
2555 void RemoveDeadBindingsWorker::VisitCluster(const MemRegion *baseR,
2556                                             const ClusterBindings *C) {
2557   if (!C)
2558     return;
2559 
2560   // Mark the symbol for any SymbolicRegion with live bindings as live itself.
2561   // This means we should continue to track that symbol.
2562   if (const SymbolicRegion *SymR = dyn_cast<SymbolicRegion>(baseR))
2563     SymReaper.markLive(SymR->getSymbol());
2564 
2565   for (ClusterBindings::iterator I = C->begin(), E = C->end(); I != E; ++I) {
2566     // Element index of a binding key is live.
2567     SymReaper.markElementIndicesLive(I.getKey().getRegion());
2568 
2569     VisitBinding(I.getData());
2570   }
2571 }
2572 
2573 void RemoveDeadBindingsWorker::VisitBinding(SVal V) {
2574   // Is it a LazyCompoundVal?  All referenced regions are live as well.
2575   if (Optional<nonloc::LazyCompoundVal> LCS =
2576           V.getAs<nonloc::LazyCompoundVal>()) {
2577 
2578     const RegionStoreManager::SValListTy &Vals = RM.getInterestingValues(*LCS);
2579 
2580     for (RegionStoreManager::SValListTy::const_iterator I = Vals.begin(),
2581                                                         E = Vals.end();
2582          I != E; ++I)
2583       VisitBinding(*I);
2584 
2585     return;
2586   }
2587 
2588   // If V is a region, then add it to the worklist.
2589   if (const MemRegion *R = V.getAsRegion()) {
2590     AddToWorkList(R);
2591     SymReaper.markLive(R);
2592 
2593     // All regions captured by a block are also live.
2594     if (const BlockDataRegion *BR = dyn_cast<BlockDataRegion>(R)) {
2595       BlockDataRegion::referenced_vars_iterator I = BR->referenced_vars_begin(),
2596                                                 E = BR->referenced_vars_end();
2597       for ( ; I != E; ++I)
2598         AddToWorkList(I.getCapturedRegion());
2599     }
2600   }
2601 
2602 
2603   // Update the set of live symbols.
2604   for (auto SI = V.symbol_begin(), SE = V.symbol_end(); SI!=SE; ++SI)
2605     SymReaper.markLive(*SI);
2606 }
2607 
2608 bool RemoveDeadBindingsWorker::UpdatePostponed() {
2609   // See if any postponed SymbolicRegions are actually live now, after
2610   // having done a scan.
2611   bool Changed = false;
2612 
2613   for (auto I = Postponed.begin(), E = Postponed.end(); I != E; ++I) {
2614     if (const SymbolicRegion *SR = *I) {
2615       if (SymReaper.isLive(SR->getSymbol())) {
2616         Changed |= AddToWorkList(SR);
2617         *I = nullptr;
2618       }
2619     }
2620   }
2621 
2622   return Changed;
2623 }
2624 
2625 StoreRef RegionStoreManager::removeDeadBindings(Store store,
2626                                                 const StackFrameContext *LCtx,
2627                                                 SymbolReaper& SymReaper) {
2628   RegionBindingsRef B = getRegionBindings(store);
2629   RemoveDeadBindingsWorker W(*this, StateMgr, B, SymReaper, LCtx);
2630   W.GenerateClusters();
2631 
2632   // Enqueue the region roots onto the worklist.
2633   for (SymbolReaper::region_iterator I = SymReaper.region_begin(),
2634        E = SymReaper.region_end(); I != E; ++I) {
2635     W.AddToWorkList(*I);
2636   }
2637 
2638   do W.RunWorkList(); while (W.UpdatePostponed());
2639 
2640   // We have now scanned the store, marking reachable regions and symbols
2641   // as live.  We now remove all the regions that are dead from the store
2642   // as well as update DSymbols with the set symbols that are now dead.
2643   for (RegionBindingsRef::iterator I = B.begin(), E = B.end(); I != E; ++I) {
2644     const MemRegion *Base = I.getKey();
2645 
2646     // If the cluster has been visited, we know the region has been marked.
2647     // Otherwise, remove the dead entry.
2648     if (!W.isVisited(Base))
2649       B = B.remove(Base);
2650   }
2651 
2652   return StoreRef(B.asStore(), *this);
2653 }
2654 
2655 //===----------------------------------------------------------------------===//
2656 // Utility methods.
2657 //===----------------------------------------------------------------------===//
2658 
2659 void RegionStoreManager::printJson(raw_ostream &Out, Store S, const char *NL,
2660                                    unsigned int Space, bool IsDot) const {
2661   RegionBindingsRef Bindings = getRegionBindings(S);
2662 
2663   Indent(Out, Space, IsDot) << "\"store\": ";
2664 
2665   if (Bindings.isEmpty()) {
2666     Out << "null," << NL;
2667     return;
2668   }
2669 
2670   Out << "{ \"pointer\": \"" << Bindings.asStore() << "\", \"items\": [" << NL;
2671   Bindings.printJson(Out, NL, Space + 1, IsDot);
2672   Indent(Out, Space, IsDot) << "]}," << NL;
2673 }
2674