CoreEngine.cpp revision 437ee81e54f39c2363d5fe0ea155604c28adc615 
1//==- CoreEngine.cpp - Path-Sensitive Dataflow Engine ------------*- C++ -*-//
2//
3//                     The LLVM Compiler Infrastructure
4//
5// This file is distributed under the University of Illinois Open Source
6// License. See LICENSE.TXT for details.
7//
8//===----------------------------------------------------------------------===//
9//
10//  This file defines a generic engine for intraprocedural, path-sensitive,
11//  dataflow analysis via graph reachability engine.
12//
13//===----------------------------------------------------------------------===//
14
15#include "clang/StaticAnalyzer/Core/PathSensitive/AnalysisManager.h"
16#include "clang/StaticAnalyzer/Core/PathSensitive/CoreEngine.h"
17#include "clang/StaticAnalyzer/Core/PathSensitive/ExprEngine.h"
18#include "clang/Index/TranslationUnit.h"
19#include "clang/AST/Expr.h"
20#include "clang/AST/StmtCXX.h"
21#include "llvm/Support/Casting.h"
22#include "llvm/ADT/DenseMap.h"
23using namespace clang;
24using namespace ento;
25
26//===----------------------------------------------------------------------===//
27// Worklist classes for exploration of reachable states.
28//===----------------------------------------------------------------------===//
29
30WorkList::Visitor::~Visitor() {}
31
32namespace {
33class DFS : public WorkList {
34  SmallVector<WorkListUnit,20> Stack;
35public:
36  virtual bool hasWork() const {
37    return !Stack.empty();
38  }
39
40  virtual void enqueue(const WorkListUnit& U) {
41    Stack.push_back(U);
42  }
43
44  virtual WorkListUnit dequeue() {
45    assert (!Stack.empty());
46    const WorkListUnit& U = Stack.back();
47    Stack.pop_back(); // This technically "invalidates" U, but we are fine.
48    return U;
49  }
50
51  virtual bool visitItemsInWorkList(Visitor &V) {
52    for (SmallVectorImpl<WorkListUnit>::iterator
53         I = Stack.begin(), E = Stack.end(); I != E; ++I) {
54      if (V.visit(*I))
55        return true;
56    }
57    return false;
58  }
59};
60
61class BFS : public WorkList {
62  std::deque<WorkListUnit> Queue;
63public:
64  virtual bool hasWork() const {
65    return !Queue.empty();
66  }
67
68  virtual void enqueue(const WorkListUnit& U) {
69    Queue.push_front(U);
70  }
71
72  virtual WorkListUnit dequeue() {
73    WorkListUnit U = Queue.front();
74    Queue.pop_front();
75    return U;
76  }
77
78  virtual bool visitItemsInWorkList(Visitor &V) {
79    for (std::deque<WorkListUnit>::iterator
80         I = Queue.begin(), E = Queue.end(); I != E; ++I) {
81      if (V.visit(*I))
82        return true;
83    }
84    return false;
85  }
86};
87
88} // end anonymous namespace
89
90// Place the dstor for WorkList here because it contains virtual member
91// functions, and we the code for the dstor generated in one compilation unit.
92WorkList::~WorkList() {}
93
94WorkList *WorkList::makeDFS() { return new DFS(); }
95WorkList *WorkList::makeBFS() { return new BFS(); }
96
97namespace {
98  class BFSBlockDFSContents : public WorkList {
99    std::deque<WorkListUnit> Queue;
100    SmallVector<WorkListUnit,20> Stack;
101  public:
102    virtual bool hasWork() const {
103      return !Queue.empty() || !Stack.empty();
104    }
105
106    virtual void enqueue(const WorkListUnit& U) {
107      if (isa<BlockEntrance>(U.getNode()->getLocation()))
108        Queue.push_front(U);
109      else
110        Stack.push_back(U);
111    }
112
113    virtual WorkListUnit dequeue() {
114      // Process all basic blocks to completion.
115      if (!Stack.empty()) {
116        const WorkListUnit& U = Stack.back();
117        Stack.pop_back(); // This technically "invalidates" U, but we are fine.
118        return U;
119      }
120
121      assert(!Queue.empty());
122      // Don't use const reference.  The subsequent pop_back() might make it
123      // unsafe.
124      WorkListUnit U = Queue.front();
125      Queue.pop_front();
126      return U;
127    }
128    virtual bool visitItemsInWorkList(Visitor &V) {
129      for (SmallVectorImpl<WorkListUnit>::iterator
130           I = Stack.begin(), E = Stack.end(); I != E; ++I) {
131        if (V.visit(*I))
132          return true;
133      }
134      for (std::deque<WorkListUnit>::iterator
135           I = Queue.begin(), E = Queue.end(); I != E; ++I) {
136        if (V.visit(*I))
137          return true;
138      }
139      return false;
140    }
141
142  };
143} // end anonymous namespace
144
145WorkList* WorkList::makeBFSBlockDFSContents() {
146  return new BFSBlockDFSContents();
147}
148
149//===----------------------------------------------------------------------===//
150// Core analysis engine.
151//===----------------------------------------------------------------------===//
152
153/// ExecuteWorkList - Run the worklist algorithm for a maximum number of steps.
154bool CoreEngine::ExecuteWorkList(const LocationContext *L, unsigned Steps,
155                                   ProgramStateRef InitState) {
156
157  if (G->num_roots() == 0) { // Initialize the analysis by constructing
158    // the root if none exists.
159
160    const CFGBlock *Entry = &(L->getCFG()->getEntry());
161
162    assert (Entry->empty() &&
163            "Entry block must be empty.");
164
165    assert (Entry->succ_size() == 1 &&
166            "Entry block must have 1 successor.");
167
168    // Get the solitary successor.
169    const CFGBlock *Succ = *(Entry->succ_begin());
170
171    // Construct an edge representing the
172    // starting location in the function.
173    BlockEdge StartLoc(Entry, Succ, L);
174
175    // Set the current block counter to being empty.
176    WList->setBlockCounter(BCounterFactory.GetEmptyCounter());
177
178    if (!InitState)
179      // Generate the root.
180      generateNode(StartLoc, SubEng.getInitialState(L), 0);
181    else
182      generateNode(StartLoc, InitState, 0);
183  }
184
185  // Check if we have a steps limit
186  bool UnlimitedSteps = Steps == 0;
187
188  while (WList->hasWork()) {
189    if (!UnlimitedSteps) {
190      if (Steps == 0)
191        break;
192      --Steps;
193    }
194
195    getGraph().reclaimChangedNodes();
196    const WorkListUnit& WU = WList->dequeue();
197
198    // Set the current block counter.
199    WList->setBlockCounter(WU.getBlockCounter());
200
201    // Retrieve the node.
202    ExplodedNode *Node = WU.getNode();
203
204    // Dispatch on the location type.
205    switch (Node->getLocation().getKind()) {
206      case ProgramPoint::BlockEdgeKind:
207        HandleBlockEdge(cast<BlockEdge>(Node->getLocation()), Node);
208        break;
209
210      case ProgramPoint::BlockEntranceKind:
211        HandleBlockEntrance(cast<BlockEntrance>(Node->getLocation()), Node);
212        break;
213
214      case ProgramPoint::BlockExitKind:
215        assert (false && "BlockExit location never occur in forward analysis.");
216        break;
217
218      case ProgramPoint::CallEnterKind:
219        SubEng.processCallEnter(cast<CallEnter>(Node->getLocation()), Node);
220        break;
221
222      case ProgramPoint::CallExitKind:
223        SubEng.processCallExit(Node);
224        break;
225
226      default:
227        assert(isa<PostStmt>(Node->getLocation()) ||
228               isa<PostInitializer>(Node->getLocation()));
229        HandlePostStmt(WU.getBlock(), WU.getIndex(), Node);
230        break;
231    }
232  }
233
234  SubEng.processEndWorklist(hasWorkRemaining());
235  return WList->hasWork();
236}
237
238void CoreEngine::ExecuteWorkListWithInitialState(const LocationContext *L,
239                                                 unsigned Steps,
240                                                 ProgramStateRef InitState,
241                                                 ExplodedNodeSet &Dst) {
242  ExecuteWorkList(L, Steps, InitState);
243  for (ExplodedGraph::eop_iterator I = G->eop_begin(),
244                                   E = G->eop_end(); I != E; ++I) {
245    Dst.Add(*I);
246  }
247}
248
249void CoreEngine::HandleBlockEdge(const BlockEdge &L, ExplodedNode *Pred) {
250
251  const CFGBlock *Blk = L.getDst();
252  NodeBuilderContext BuilderCtx(*this, Blk, Pred);
253
254  // Check if we are entering the EXIT block.
255  if (Blk == &(L.getLocationContext()->getCFG()->getExit())) {
256
257    assert (L.getLocationContext()->getCFG()->getExit().size() == 0
258            && "EXIT block cannot contain Stmts.");
259
260    // Process the final state transition.
261    SubEng.processEndOfFunction(BuilderCtx);
262
263    // This path is done. Don't enqueue any more nodes.
264    return;
265  }
266
267  // Call into the SubEngine to process entering the CFGBlock.
268  ExplodedNodeSet dstNodes;
269  BlockEntrance BE(Blk, Pred->getLocationContext());
270  NodeBuilderWithSinks nodeBuilder(Pred, dstNodes, BuilderCtx, BE);
271  SubEng.processCFGBlockEntrance(nodeBuilder);
272
273  // Auto-generate a node.
274  if (!nodeBuilder.hasGeneratedNodes()) {
275    nodeBuilder.generateNode(Pred->State, Pred);
276  }
277
278  // Enqueue nodes onto the worklist.
279  enqueue(dstNodes);
280
281  // Make sink nodes as exhausted.
282  const SmallVectorImpl<ExplodedNode*> &Sinks =  nodeBuilder.getSinks();
283  for (SmallVectorImpl<ExplodedNode*>::const_iterator
284         I =Sinks.begin(), E = Sinks.end(); I != E; ++I) {
285    blocksExhausted.push_back(std::make_pair(L, *I));
286  }
287}
288
289void CoreEngine::HandleBlockEntrance(const BlockEntrance &L,
290                                       ExplodedNode *Pred) {
291
292  // Increment the block counter.
293  BlockCounter Counter = WList->getBlockCounter();
294  Counter = BCounterFactory.IncrementCount(Counter,
295                             Pred->getLocationContext()->getCurrentStackFrame(),
296                                           L.getBlock()->getBlockID());
297  WList->setBlockCounter(Counter);
298
299  // Process the entrance of the block.
300  if (CFGElement E = L.getFirstElement()) {
301    NodeBuilderContext Ctx(*this, L.getBlock(), Pred);
302    SubEng.processCFGElement(E, Pred, 0, &Ctx);
303  }
304  else
305    HandleBlockExit(L.getBlock(), Pred);
306}
307
308void CoreEngine::HandleBlockExit(const CFGBlock * B, ExplodedNode *Pred) {
309
310  if (const Stmt *Term = B->getTerminator()) {
311    switch (Term->getStmtClass()) {
312      default:
313        llvm_unreachable("Analysis for this terminator not implemented.");
314
315      case Stmt::BinaryOperatorClass: // '&&' and '||'
316        HandleBranch(cast<BinaryOperator>(Term)->getLHS(), Term, B, Pred);
317        return;
318
319      case Stmt::BinaryConditionalOperatorClass:
320      case Stmt::ConditionalOperatorClass:
321        HandleBranch(cast<AbstractConditionalOperator>(Term)->getCond(),
322                     Term, B, Pred);
323        return;
324
325        // FIXME: Use constant-folding in CFG construction to simplify this
326        // case.
327
328      case Stmt::ChooseExprClass:
329        HandleBranch(cast<ChooseExpr>(Term)->getCond(), Term, B, Pred);
330        return;
331
332      case Stmt::DoStmtClass:
333        HandleBranch(cast<DoStmt>(Term)->getCond(), Term, B, Pred);
334        return;
335
336      case Stmt::CXXForRangeStmtClass:
337        HandleBranch(cast<CXXForRangeStmt>(Term)->getCond(), Term, B, Pred);
338        return;
339
340      case Stmt::ForStmtClass:
341        HandleBranch(cast<ForStmt>(Term)->getCond(), Term, B, Pred);
342        return;
343
344      case Stmt::ContinueStmtClass:
345      case Stmt::BreakStmtClass:
346      case Stmt::GotoStmtClass:
347        break;
348
349      case Stmt::IfStmtClass:
350        HandleBranch(cast<IfStmt>(Term)->getCond(), Term, B, Pred);
351        return;
352
353      case Stmt::IndirectGotoStmtClass: {
354        // Only 1 successor: the indirect goto dispatch block.
355        assert (B->succ_size() == 1);
356
357        IndirectGotoNodeBuilder
358           builder(Pred, B, cast<IndirectGotoStmt>(Term)->getTarget(),
359                   *(B->succ_begin()), this);
360
361        SubEng.processIndirectGoto(builder);
362        return;
363      }
364
365      case Stmt::ObjCForCollectionStmtClass: {
366        // In the case of ObjCForCollectionStmt, it appears twice in a CFG:
367        //
368        //  (1) inside a basic block, which represents the binding of the
369        //      'element' variable to a value.
370        //  (2) in a terminator, which represents the branch.
371        //
372        // For (1), subengines will bind a value (i.e., 0 or 1) indicating
373        // whether or not collection contains any more elements.  We cannot
374        // just test to see if the element is nil because a container can
375        // contain nil elements.
376        HandleBranch(Term, Term, B, Pred);
377        return;
378      }
379
380      case Stmt::SwitchStmtClass: {
381        SwitchNodeBuilder builder(Pred, B, cast<SwitchStmt>(Term)->getCond(),
382                                    this);
383
384        SubEng.processSwitch(builder);
385        return;
386      }
387
388      case Stmt::WhileStmtClass:
389        HandleBranch(cast<WhileStmt>(Term)->getCond(), Term, B, Pred);
390        return;
391    }
392  }
393
394  assert (B->succ_size() == 1 &&
395          "Blocks with no terminator should have at most 1 successor.");
396
397  generateNode(BlockEdge(B, *(B->succ_begin()), Pred->getLocationContext()),
398               Pred->State, Pred);
399}
400
401void CoreEngine::HandleBranch(const Stmt *Cond, const Stmt *Term,
402                                const CFGBlock * B, ExplodedNode *Pred) {
403  assert(B->succ_size() == 2);
404  NodeBuilderContext Ctx(*this, B, Pred);
405  ExplodedNodeSet Dst;
406  SubEng.processBranch(Cond, Term, Ctx, Pred, Dst,
407                       *(B->succ_begin()), *(B->succ_begin()+1));
408  // Enqueue the new frontier onto the worklist.
409  enqueue(Dst);
410}
411
412void CoreEngine::HandlePostStmt(const CFGBlock *B, unsigned StmtIdx,
413                                  ExplodedNode *Pred) {
414  assert(B);
415  assert(!B->empty());
416
417  if (StmtIdx == B->size())
418    HandleBlockExit(B, Pred);
419  else {
420    NodeBuilderContext Ctx(*this, B, Pred);
421    SubEng.processCFGElement((*B)[StmtIdx], Pred, StmtIdx, &Ctx);
422  }
423}
424
425/// generateNode - Utility method to generate nodes, hook up successors,
426///  and add nodes to the worklist.
427void CoreEngine::generateNode(const ProgramPoint &Loc,
428                              ProgramStateRef State,
429                              ExplodedNode *Pred) {
430
431  bool IsNew;
432  ExplodedNode *Node = G->getNode(Loc, State, false, &IsNew);
433
434  if (Pred)
435    Node->addPredecessor(Pred, *G);  // Link 'Node' with its predecessor.
436  else {
437    assert (IsNew);
438    G->addRoot(Node);  // 'Node' has no predecessor.  Make it a root.
439  }
440
441  // Only add 'Node' to the worklist if it was freshly generated.
442  if (IsNew) WList->enqueue(Node);
443}
444
445void CoreEngine::enqueueStmtNode(ExplodedNode *N,
446                                 const CFGBlock *Block, unsigned Idx) {
447  assert(Block);
448  assert (!N->isSink());
449
450  // Check if this node entered a callee.
451  if (isa<CallEnter>(N->getLocation())) {
452    // Still use the index of the CallExpr. It's needed to create the callee
453    // StackFrameContext.
454    WList->enqueue(N, Block, Idx);
455    return;
456  }
457
458  // Do not create extra nodes. Move to the next CFG element.
459  if (isa<PostInitializer>(N->getLocation())) {
460    WList->enqueue(N, Block, Idx+1);
461    return;
462  }
463
464  const CFGStmt *CS = (*Block)[Idx].getAs<CFGStmt>();
465  const Stmt *St = CS ? CS->getStmt() : 0;
466  PostStmt Loc(St, N->getLocationContext());
467
468  if (Loc == N->getLocation()) {
469    // Note: 'N' should be a fresh node because otherwise it shouldn't be
470    // a member of Deferred.
471    WList->enqueue(N, Block, Idx+1);
472    return;
473  }
474
475  bool IsNew;
476  ExplodedNode *Succ = G->getNode(Loc, N->getState(), false, &IsNew);
477  Succ->addPredecessor(N, *G);
478
479  if (IsNew)
480    WList->enqueue(Succ, Block, Idx+1);
481}
482
483ExplodedNode *CoreEngine::generateCallExitNode(ExplodedNode *N) {
484  // Create a CallExit node and enqueue it.
485  const StackFrameContext *LocCtx
486                         = cast<StackFrameContext>(N->getLocationContext());
487  const Stmt *CE = LocCtx->getCallSite();
488
489  // Use the the callee location context.
490  CallExit Loc(CE, LocCtx);
491
492  bool isNew;
493  ExplodedNode *Node = G->getNode(Loc, N->getState(), false, &isNew);
494  Node->addPredecessor(N, *G);
495  return isNew ? Node : 0;
496}
497
498
499void CoreEngine::enqueue(ExplodedNodeSet &Set) {
500  for (ExplodedNodeSet::iterator I = Set.begin(),
501                                 E = Set.end(); I != E; ++I) {
502    WList->enqueue(*I);
503  }
504}
505
506void CoreEngine::enqueue(ExplodedNodeSet &Set,
507                         const CFGBlock *Block, unsigned Idx) {
508  for (ExplodedNodeSet::iterator I = Set.begin(),
509                                 E = Set.end(); I != E; ++I) {
510    enqueueStmtNode(*I, Block, Idx);
511  }
512}
513
514void CoreEngine::enqueueEndOfFunction(ExplodedNodeSet &Set) {
515  for (ExplodedNodeSet::iterator I = Set.begin(), E = Set.end(); I != E; ++I) {
516    ExplodedNode *N = *I;
517    // If we are in an inlined call, generate CallExit node.
518    if (N->getLocationContext()->getParent()) {
519      N = generateCallExitNode(N);
520      if (N)
521        WList->enqueue(N);
522    } else
523      G->addEndOfPath(N);
524  }
525}
526
527
528void NodeBuilder::anchor() { }
529
530ExplodedNode* NodeBuilder::generateNodeImpl(const ProgramPoint &Loc,
531                                            ProgramStateRef State,
532                                            ExplodedNode *FromN,
533                                            bool MarkAsSink) {
534  HasGeneratedNodes = true;
535  bool IsNew;
536  ExplodedNode *N = C.Eng.G->getNode(Loc, State, MarkAsSink, &IsNew);
537  N->addPredecessor(FromN, *C.Eng.G);
538  Frontier.erase(FromN);
539
540  if (!IsNew)
541    return 0;
542
543  if (!MarkAsSink)
544    Frontier.Add(N);
545
546  return N;
547}
548
549void NodeBuilderWithSinks::anchor() { }
550
551StmtNodeBuilder::~StmtNodeBuilder() {
552  if (EnclosingBldr)
553    for (ExplodedNodeSet::iterator I = Frontier.begin(),
554                                   E = Frontier.end(); I != E; ++I )
555      EnclosingBldr->addNodes(*I);
556}
557
558void BranchNodeBuilder::anchor() { }
559
560ExplodedNode *BranchNodeBuilder::generateNode(ProgramStateRef State,
561                                              bool branch,
562                                              ExplodedNode *NodePred) {
563  // If the branch has been marked infeasible we should not generate a node.
564  if (!isFeasible(branch))
565    return NULL;
566
567  ProgramPoint Loc = BlockEdge(C.Block, branch ? DstT:DstF,
568                               NodePred->getLocationContext());
569  ExplodedNode *Succ = generateNodeImpl(Loc, State, NodePred);
570  return Succ;
571}
572
573ExplodedNode*
574IndirectGotoNodeBuilder::generateNode(const iterator &I,
575                                      ProgramStateRef St,
576                                      bool IsSink) {
577  bool IsNew;
578  ExplodedNode *Succ = Eng.G->getNode(BlockEdge(Src, I.getBlock(),
579                                      Pred->getLocationContext()), St,
580                                      IsSink, &IsNew);
581  Succ->addPredecessor(Pred, *Eng.G);
582
583  if (!IsNew)
584    return 0;
585
586  if (!IsSink)
587    Eng.WList->enqueue(Succ);
588
589  return Succ;
590}
591
592
593ExplodedNode*
594SwitchNodeBuilder::generateCaseStmtNode(const iterator &I,
595                                        ProgramStateRef St) {
596
597  bool IsNew;
598  ExplodedNode *Succ = Eng.G->getNode(BlockEdge(Src, I.getBlock(),
599                                      Pred->getLocationContext()), St,
600                                      false, &IsNew);
601  Succ->addPredecessor(Pred, *Eng.G);
602  if (!IsNew)
603    return 0;
604
605  Eng.WList->enqueue(Succ);
606  return Succ;
607}
608
609
610ExplodedNode*
611SwitchNodeBuilder::generateDefaultCaseNode(ProgramStateRef St,
612                                           bool IsSink) {
613  // Get the block for the default case.
614  assert(Src->succ_rbegin() != Src->succ_rend());
615  CFGBlock *DefaultBlock = *Src->succ_rbegin();
616
617  // Sanity check for default blocks that are unreachable and not caught
618  // by earlier stages.
619  if (!DefaultBlock)
620    return NULL;
621
622  bool IsNew;
623  ExplodedNode *Succ = Eng.G->getNode(BlockEdge(Src, DefaultBlock,
624                                      Pred->getLocationContext()), St,
625                                      IsSink, &IsNew);
626  Succ->addPredecessor(Pred, *Eng.G);
627
628  if (!IsNew)
629    return 0;
630
631  if (!IsSink)
632    Eng.WList->enqueue(Succ);
633
634  return Succ;
635}
636