CoreEngine.cpp revision df19fe7cafcb02859efeb6963cddeafef4350ddf 
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#define DEBUG_TYPE "CoreEngine"
16
17#include "clang/StaticAnalyzer/Core/PathSensitive/AnalysisManager.h"
18#include "clang/StaticAnalyzer/Core/PathSensitive/CoreEngine.h"
19#include "clang/StaticAnalyzer/Core/PathSensitive/ExprEngine.h"
20#include "clang/AST/Expr.h"
21#include "clang/AST/StmtCXX.h"
22#include "llvm/Support/Casting.h"
23#include "llvm/ADT/DenseMap.h"
24#include "llvm/ADT/Statistic.h"
25
26using namespace clang;
27using namespace ento;
28
29STATISTIC(NumSteps,
30            "The # of steps executed.");
31STATISTIC(NumReachedMaxSteps,
32            "The # of times we reached the max number of steps.");
33STATISTIC(NumPathsExplored,
34            "The # of paths explored by the analyzer.");
35
36//===----------------------------------------------------------------------===//
37// Worklist classes for exploration of reachable states.
38//===----------------------------------------------------------------------===//
39
40WorkList::Visitor::~Visitor() {}
41
42namespace {
43class DFS : public WorkList {
44  SmallVector<WorkListUnit,20> Stack;
45public:
46  virtual bool hasWork() const {
47    return !Stack.empty();
48  }
49
50  virtual void enqueue(const WorkListUnit& U) {
51    Stack.push_back(U);
52  }
53
54  virtual WorkListUnit dequeue() {
55    assert (!Stack.empty());
56    const WorkListUnit& U = Stack.back();
57    Stack.pop_back(); // This technically "invalidates" U, but we are fine.
58    return U;
59  }
60
61  virtual bool visitItemsInWorkList(Visitor &V) {
62    for (SmallVectorImpl<WorkListUnit>::iterator
63         I = Stack.begin(), E = Stack.end(); I != E; ++I) {
64      if (V.visit(*I))
65        return true;
66    }
67    return false;
68  }
69};
70
71class BFS : public WorkList {
72  std::deque<WorkListUnit> Queue;
73public:
74  virtual bool hasWork() const {
75    return !Queue.empty();
76  }
77
78  virtual void enqueue(const WorkListUnit& U) {
79    Queue.push_front(U);
80  }
81
82  virtual WorkListUnit dequeue() {
83    WorkListUnit U = Queue.front();
84    Queue.pop_front();
85    return U;
86  }
87
88  virtual bool visitItemsInWorkList(Visitor &V) {
89    for (std::deque<WorkListUnit>::iterator
90         I = Queue.begin(), E = Queue.end(); I != E; ++I) {
91      if (V.visit(*I))
92        return true;
93    }
94    return false;
95  }
96};
97
98} // end anonymous namespace
99
100// Place the dstor for WorkList here because it contains virtual member
101// functions, and we the code for the dstor generated in one compilation unit.
102WorkList::~WorkList() {}
103
104WorkList *WorkList::makeDFS() { return new DFS(); }
105WorkList *WorkList::makeBFS() { return new BFS(); }
106
107namespace {
108  class BFSBlockDFSContents : public WorkList {
109    std::deque<WorkListUnit> Queue;
110    SmallVector<WorkListUnit,20> Stack;
111  public:
112    virtual bool hasWork() const {
113      return !Queue.empty() || !Stack.empty();
114    }
115
116    virtual void enqueue(const WorkListUnit& U) {
117      if (isa<BlockEntrance>(U.getNode()->getLocation()))
118        Queue.push_front(U);
119      else
120        Stack.push_back(U);
121    }
122
123    virtual WorkListUnit dequeue() {
124      // Process all basic blocks to completion.
125      if (!Stack.empty()) {
126        const WorkListUnit& U = Stack.back();
127        Stack.pop_back(); // This technically "invalidates" U, but we are fine.
128        return U;
129      }
130
131      assert(!Queue.empty());
132      // Don't use const reference.  The subsequent pop_back() might make it
133      // unsafe.
134      WorkListUnit U = Queue.front();
135      Queue.pop_front();
136      return U;
137    }
138    virtual bool visitItemsInWorkList(Visitor &V) {
139      for (SmallVectorImpl<WorkListUnit>::iterator
140           I = Stack.begin(), E = Stack.end(); I != E; ++I) {
141        if (V.visit(*I))
142          return true;
143      }
144      for (std::deque<WorkListUnit>::iterator
145           I = Queue.begin(), E = Queue.end(); I != E; ++I) {
146        if (V.visit(*I))
147          return true;
148      }
149      return false;
150    }
151
152  };
153} // end anonymous namespace
154
155WorkList* WorkList::makeBFSBlockDFSContents() {
156  return new BFSBlockDFSContents();
157}
158
159//===----------------------------------------------------------------------===//
160// Core analysis engine.
161//===----------------------------------------------------------------------===//
162
163/// ExecuteWorkList - Run the worklist algorithm for a maximum number of steps.
164bool CoreEngine::ExecuteWorkList(const LocationContext *L, unsigned Steps,
165                                   ProgramStateRef InitState) {
166
167  if (G->num_roots() == 0) { // Initialize the analysis by constructing
168    // the root if none exists.
169
170    const CFGBlock *Entry = &(L->getCFG()->getEntry());
171
172    assert (Entry->empty() &&
173            "Entry block must be empty.");
174
175    assert (Entry->succ_size() == 1 &&
176            "Entry block must have 1 successor.");
177
178    // Mark the entry block as visited.
179    FunctionSummaries->markVisitedBasicBlock(Entry->getBlockID(),
180                                             L->getDecl(),
181                                             L->getCFG()->getNumBlockIDs());
182
183    // Get the solitary successor.
184    const CFGBlock *Succ = *(Entry->succ_begin());
185
186    // Construct an edge representing the
187    // starting location in the function.
188    BlockEdge StartLoc(Entry, Succ, L);
189
190    // Set the current block counter to being empty.
191    WList->setBlockCounter(BCounterFactory.GetEmptyCounter());
192
193    if (!InitState)
194      // Generate the root.
195      generateNode(StartLoc, SubEng.getInitialState(L), 0);
196    else
197      generateNode(StartLoc, InitState, 0);
198  }
199
200  // Check if we have a steps limit
201  bool UnlimitedSteps = Steps == 0;
202
203  while (WList->hasWork()) {
204    if (!UnlimitedSteps) {
205      if (Steps == 0) {
206        NumReachedMaxSteps++;
207        break;
208      }
209      --Steps;
210    }
211
212    NumSteps++;
213
214    const WorkListUnit& WU = WList->dequeue();
215
216    // Set the current block counter.
217    WList->setBlockCounter(WU.getBlockCounter());
218
219    // Retrieve the node.
220    ExplodedNode *Node = WU.getNode();
221
222    dispatchWorkItem(Node, Node->getLocation(), WU);
223  }
224  SubEng.processEndWorklist(hasWorkRemaining());
225  return WList->hasWork();
226}
227
228void CoreEngine::dispatchWorkItem(ExplodedNode* Pred, ProgramPoint Loc,
229                                  const WorkListUnit& WU) {
230  // Dispatch on the location type.
231  switch (Loc.getKind()) {
232    case ProgramPoint::BlockEdgeKind:
233      HandleBlockEdge(cast<BlockEdge>(Loc), Pred);
234      break;
235
236    case ProgramPoint::BlockEntranceKind:
237      HandleBlockEntrance(cast<BlockEntrance>(Loc), Pred);
238      break;
239
240    case ProgramPoint::BlockExitKind:
241      assert (false && "BlockExit location never occur in forward analysis.");
242      break;
243
244    case ProgramPoint::CallEnterKind: {
245      CallEnter CEnter = cast<CallEnter>(Loc);
246      if (AnalyzedCallees)
247        if (const CallExpr* CE =
248            dyn_cast_or_null<CallExpr>(CEnter.getCallExpr()))
249          if (const Decl *CD = CE->getCalleeDecl())
250            AnalyzedCallees->insert(CD);
251      SubEng.processCallEnter(CEnter, Pred);
252      break;
253    }
254
255    case ProgramPoint::CallExitBeginKind:
256      SubEng.processCallExit(Pred);
257      break;
258
259    case ProgramPoint::EpsilonKind: {
260      assert(Pred->hasSinglePred() &&
261             "Assume epsilon has exactly one predecessor by construction");
262      ExplodedNode *PNode = Pred->getFirstPred();
263      dispatchWorkItem(Pred, PNode->getLocation(), WU);
264      break;
265    }
266    default:
267      assert(isa<PostStmt>(Loc) ||
268             isa<PostInitializer>(Loc));
269      HandlePostStmt(WU.getBlock(), WU.getIndex(), Pred);
270      break;
271  }
272}
273
274bool CoreEngine::ExecuteWorkListWithInitialState(const LocationContext *L,
275                                                 unsigned Steps,
276                                                 ProgramStateRef InitState,
277                                                 ExplodedNodeSet &Dst) {
278  bool DidNotFinish = ExecuteWorkList(L, Steps, InitState);
279  for (ExplodedGraph::eop_iterator I = G->eop_begin(),
280                                   E = G->eop_end(); I != E; ++I) {
281    Dst.Add(*I);
282  }
283  return DidNotFinish;
284}
285
286void CoreEngine::HandleBlockEdge(const BlockEdge &L, ExplodedNode *Pred) {
287
288  const CFGBlock *Blk = L.getDst();
289  NodeBuilderContext BuilderCtx(*this, Blk, Pred);
290
291  // Mark this block as visited.
292  const LocationContext *LC = Pred->getLocationContext();
293  FunctionSummaries->markVisitedBasicBlock(Blk->getBlockID(),
294                                           LC->getDecl(),
295                                           LC->getCFG()->getNumBlockIDs());
296
297  // Check if we are entering the EXIT block.
298  if (Blk == &(L.getLocationContext()->getCFG()->getExit())) {
299
300    assert (L.getLocationContext()->getCFG()->getExit().size() == 0
301            && "EXIT block cannot contain Stmts.");
302
303    // Process the final state transition.
304    SubEng.processEndOfFunction(BuilderCtx);
305
306    // This path is done. Don't enqueue any more nodes.
307    return;
308  }
309
310  // Call into the SubEngine to process entering the CFGBlock.
311  ExplodedNodeSet dstNodes;
312  BlockEntrance BE(Blk, Pred->getLocationContext());
313  NodeBuilderWithSinks nodeBuilder(Pred, dstNodes, BuilderCtx, BE);
314  SubEng.processCFGBlockEntrance(L, nodeBuilder);
315
316  // Auto-generate a node.
317  if (!nodeBuilder.hasGeneratedNodes()) {
318    nodeBuilder.generateNode(Pred->State, Pred);
319  }
320
321  // Enqueue nodes onto the worklist.
322  enqueue(dstNodes);
323}
324
325void CoreEngine::HandleBlockEntrance(const BlockEntrance &L,
326                                       ExplodedNode *Pred) {
327
328  // Increment the block counter.
329  const LocationContext *LC = Pred->getLocationContext();
330  unsigned BlockId = L.getBlock()->getBlockID();
331  BlockCounter Counter = WList->getBlockCounter();
332  Counter = BCounterFactory.IncrementCount(Counter, LC->getCurrentStackFrame(),
333                                           BlockId);
334  WList->setBlockCounter(Counter);
335
336  // Process the entrance of the block.
337  if (CFGElement E = L.getFirstElement()) {
338    NodeBuilderContext Ctx(*this, L.getBlock(), Pred);
339    SubEng.processCFGElement(E, Pred, 0, &Ctx);
340  }
341  else
342    HandleBlockExit(L.getBlock(), Pred);
343}
344
345void CoreEngine::HandleBlockExit(const CFGBlock * B, ExplodedNode *Pred) {
346
347  if (const Stmt *Term = B->getTerminator()) {
348    switch (Term->getStmtClass()) {
349      default:
350        llvm_unreachable("Analysis for this terminator not implemented.");
351
352      case Stmt::BinaryOperatorClass: // '&&' and '||'
353        HandleBranch(cast<BinaryOperator>(Term)->getLHS(), Term, B, Pred);
354        return;
355
356      case Stmt::BinaryConditionalOperatorClass:
357      case Stmt::ConditionalOperatorClass:
358        HandleBranch(cast<AbstractConditionalOperator>(Term)->getCond(),
359                     Term, B, Pred);
360        return;
361
362        // FIXME: Use constant-folding in CFG construction to simplify this
363        // case.
364
365      case Stmt::ChooseExprClass:
366        HandleBranch(cast<ChooseExpr>(Term)->getCond(), Term, B, Pred);
367        return;
368
369      case Stmt::CXXTryStmtClass: {
370        // Generate a node for each of the successors.
371        // Our logic for EH analysis can certainly be improved.
372        for (CFGBlock::const_succ_iterator it = B->succ_begin(),
373             et = B->succ_end(); it != et; ++it) {
374          if (const CFGBlock *succ = *it) {
375            generateNode(BlockEdge(B, succ, Pred->getLocationContext()),
376                         Pred->State, Pred);
377          }
378        }
379        return;
380      }
381
382      case Stmt::DoStmtClass:
383        HandleBranch(cast<DoStmt>(Term)->getCond(), Term, B, Pred);
384        return;
385
386      case Stmt::CXXForRangeStmtClass:
387        HandleBranch(cast<CXXForRangeStmt>(Term)->getCond(), Term, B, Pred);
388        return;
389
390      case Stmt::ForStmtClass:
391        HandleBranch(cast<ForStmt>(Term)->getCond(), Term, B, Pred);
392        return;
393
394      case Stmt::ContinueStmtClass:
395      case Stmt::BreakStmtClass:
396      case Stmt::GotoStmtClass:
397        break;
398
399      case Stmt::IfStmtClass:
400        HandleBranch(cast<IfStmt>(Term)->getCond(), Term, B, Pred);
401        return;
402
403      case Stmt::IndirectGotoStmtClass: {
404        // Only 1 successor: the indirect goto dispatch block.
405        assert (B->succ_size() == 1);
406
407        IndirectGotoNodeBuilder
408           builder(Pred, B, cast<IndirectGotoStmt>(Term)->getTarget(),
409                   *(B->succ_begin()), this);
410
411        SubEng.processIndirectGoto(builder);
412        return;
413      }
414
415      case Stmt::ObjCForCollectionStmtClass: {
416        // In the case of ObjCForCollectionStmt, it appears twice in a CFG:
417        //
418        //  (1) inside a basic block, which represents the binding of the
419        //      'element' variable to a value.
420        //  (2) in a terminator, which represents the branch.
421        //
422        // For (1), subengines will bind a value (i.e., 0 or 1) indicating
423        // whether or not collection contains any more elements.  We cannot
424        // just test to see if the element is nil because a container can
425        // contain nil elements.
426        HandleBranch(Term, Term, B, Pred);
427        return;
428      }
429
430      case Stmt::SwitchStmtClass: {
431        SwitchNodeBuilder builder(Pred, B, cast<SwitchStmt>(Term)->getCond(),
432                                    this);
433
434        SubEng.processSwitch(builder);
435        return;
436      }
437
438      case Stmt::WhileStmtClass:
439        HandleBranch(cast<WhileStmt>(Term)->getCond(), Term, B, Pred);
440        return;
441    }
442  }
443
444  assert (B->succ_size() == 1 &&
445          "Blocks with no terminator should have at most 1 successor.");
446
447  generateNode(BlockEdge(B, *(B->succ_begin()), Pred->getLocationContext()),
448               Pred->State, Pred);
449}
450
451void CoreEngine::HandleBranch(const Stmt *Cond, const Stmt *Term,
452                                const CFGBlock * B, ExplodedNode *Pred) {
453  assert(B->succ_size() == 2);
454  NodeBuilderContext Ctx(*this, B, Pred);
455  ExplodedNodeSet Dst;
456  SubEng.processBranch(Cond, Term, Ctx, Pred, Dst,
457                       *(B->succ_begin()), *(B->succ_begin()+1));
458  // Enqueue the new frontier onto the worklist.
459  enqueue(Dst);
460}
461
462void CoreEngine::HandlePostStmt(const CFGBlock *B, unsigned StmtIdx,
463                                  ExplodedNode *Pred) {
464  assert(B);
465  assert(!B->empty());
466
467  if (StmtIdx == B->size())
468    HandleBlockExit(B, Pred);
469  else {
470    NodeBuilderContext Ctx(*this, B, Pred);
471    SubEng.processCFGElement((*B)[StmtIdx], Pred, StmtIdx, &Ctx);
472  }
473}
474
475/// generateNode - Utility method to generate nodes, hook up successors,
476///  and add nodes to the worklist.
477void CoreEngine::generateNode(const ProgramPoint &Loc,
478                              ProgramStateRef State,
479                              ExplodedNode *Pred) {
480
481  bool IsNew;
482  ExplodedNode *Node = G->getNode(Loc, State, false, &IsNew);
483
484  if (Pred)
485    Node->addPredecessor(Pred, *G);  // Link 'Node' with its predecessor.
486  else {
487    assert (IsNew);
488    G->addRoot(Node);  // 'Node' has no predecessor.  Make it a root.
489  }
490
491  // Only add 'Node' to the worklist if it was freshly generated.
492  if (IsNew) WList->enqueue(Node);
493}
494
495void CoreEngine::enqueueStmtNode(ExplodedNode *N,
496                                 const CFGBlock *Block, unsigned Idx) {
497  assert(Block);
498  assert (!N->isSink());
499
500  // Check if this node entered a callee.
501  if (isa<CallEnter>(N->getLocation())) {
502    // Still use the index of the CallExpr. It's needed to create the callee
503    // StackFrameContext.
504    WList->enqueue(N, Block, Idx);
505    return;
506  }
507
508  // Do not create extra nodes. Move to the next CFG element.
509  if (isa<PostInitializer>(N->getLocation())) {
510    WList->enqueue(N, Block, Idx+1);
511    return;
512  }
513
514  if (isa<EpsilonPoint>(N->getLocation())) {
515    WList->enqueue(N, Block, Idx);
516    return;
517  }
518
519  const CFGStmt *CS = (*Block)[Idx].getAs<CFGStmt>();
520  const Stmt *St = CS ? CS->getStmt() : 0;
521  PostStmt Loc(St, N->getLocationContext());
522
523  if (Loc == N->getLocation()) {
524    // Note: 'N' should be a fresh node because otherwise it shouldn't be
525    // a member of Deferred.
526    WList->enqueue(N, Block, Idx+1);
527    return;
528  }
529
530  bool IsNew;
531  ExplodedNode *Succ = G->getNode(Loc, N->getState(), false, &IsNew);
532  Succ->addPredecessor(N, *G);
533
534  if (IsNew)
535    WList->enqueue(Succ, Block, Idx+1);
536}
537
538ExplodedNode *CoreEngine::generateCallExitBeginNode(ExplodedNode *N) {
539  // Create a CallExitBegin node and enqueue it.
540  const StackFrameContext *LocCtx
541                         = cast<StackFrameContext>(N->getLocationContext());
542  const Stmt *CE = LocCtx->getCallSite();
543
544  // Use the the callee location context.
545  CallExitBegin Loc(CE, LocCtx);
546
547  bool isNew;
548  ExplodedNode *Node = G->getNode(Loc, N->getState(), false, &isNew);
549  Node->addPredecessor(N, *G);
550  return isNew ? Node : 0;
551}
552
553
554void CoreEngine::enqueue(ExplodedNodeSet &Set) {
555  for (ExplodedNodeSet::iterator I = Set.begin(),
556                                 E = Set.end(); I != E; ++I) {
557    WList->enqueue(*I);
558  }
559}
560
561void CoreEngine::enqueue(ExplodedNodeSet &Set,
562                         const CFGBlock *Block, unsigned Idx) {
563  for (ExplodedNodeSet::iterator I = Set.begin(),
564                                 E = Set.end(); I != E; ++I) {
565    enqueueStmtNode(*I, Block, Idx);
566  }
567}
568
569void CoreEngine::enqueueEndOfFunction(ExplodedNodeSet &Set) {
570  for (ExplodedNodeSet::iterator I = Set.begin(), E = Set.end(); I != E; ++I) {
571    ExplodedNode *N = *I;
572    // If we are in an inlined call, generate CallExitBegin node.
573    if (N->getLocationContext()->getParent()) {
574      N = generateCallExitBeginNode(N);
575      if (N)
576        WList->enqueue(N);
577    } else {
578      // TODO: We should run remove dead bindings here.
579      G->addEndOfPath(N);
580      NumPathsExplored++;
581    }
582  }
583}
584
585
586void NodeBuilder::anchor() { }
587
588ExplodedNode* NodeBuilder::generateNodeImpl(const ProgramPoint &Loc,
589                                            ProgramStateRef State,
590                                            ExplodedNode *FromN,
591                                            bool MarkAsSink) {
592  HasGeneratedNodes = true;
593  bool IsNew;
594  ExplodedNode *N = C.Eng.G->getNode(Loc, State, MarkAsSink, &IsNew);
595  N->addPredecessor(FromN, *C.Eng.G);
596  Frontier.erase(FromN);
597
598  if (!IsNew)
599    return 0;
600
601  if (!MarkAsSink)
602    Frontier.Add(N);
603
604  return N;
605}
606
607void NodeBuilderWithSinks::anchor() { }
608
609StmtNodeBuilder::~StmtNodeBuilder() {
610  if (EnclosingBldr)
611    for (ExplodedNodeSet::iterator I = Frontier.begin(),
612                                   E = Frontier.end(); I != E; ++I )
613      EnclosingBldr->addNodes(*I);
614}
615
616void BranchNodeBuilder::anchor() { }
617
618ExplodedNode *BranchNodeBuilder::generateNode(ProgramStateRef State,
619                                              bool branch,
620                                              ExplodedNode *NodePred) {
621  // If the branch has been marked infeasible we should not generate a node.
622  if (!isFeasible(branch))
623    return NULL;
624
625  ProgramPoint Loc = BlockEdge(C.Block, branch ? DstT:DstF,
626                               NodePred->getLocationContext());
627  ExplodedNode *Succ = generateNodeImpl(Loc, State, NodePred);
628  return Succ;
629}
630
631ExplodedNode*
632IndirectGotoNodeBuilder::generateNode(const iterator &I,
633                                      ProgramStateRef St,
634                                      bool IsSink) {
635  bool IsNew;
636  ExplodedNode *Succ = Eng.G->getNode(BlockEdge(Src, I.getBlock(),
637                                      Pred->getLocationContext()), St,
638                                      IsSink, &IsNew);
639  Succ->addPredecessor(Pred, *Eng.G);
640
641  if (!IsNew)
642    return 0;
643
644  if (!IsSink)
645    Eng.WList->enqueue(Succ);
646
647  return Succ;
648}
649
650
651ExplodedNode*
652SwitchNodeBuilder::generateCaseStmtNode(const iterator &I,
653                                        ProgramStateRef St) {
654
655  bool IsNew;
656  ExplodedNode *Succ = Eng.G->getNode(BlockEdge(Src, I.getBlock(),
657                                      Pred->getLocationContext()), St,
658                                      false, &IsNew);
659  Succ->addPredecessor(Pred, *Eng.G);
660  if (!IsNew)
661    return 0;
662
663  Eng.WList->enqueue(Succ);
664  return Succ;
665}
666
667
668ExplodedNode*
669SwitchNodeBuilder::generateDefaultCaseNode(ProgramStateRef St,
670                                           bool IsSink) {
671  // Get the block for the default case.
672  assert(Src->succ_rbegin() != Src->succ_rend());
673  CFGBlock *DefaultBlock = *Src->succ_rbegin();
674
675  // Sanity check for default blocks that are unreachable and not caught
676  // by earlier stages.
677  if (!DefaultBlock)
678    return NULL;
679
680  bool IsNew;
681  ExplodedNode *Succ = Eng.G->getNode(BlockEdge(Src, DefaultBlock,
682                                      Pred->getLocationContext()), St,
683                                      IsSink, &IsNew);
684  Succ->addPredecessor(Pred, *Eng.G);
685
686  if (!IsNew)
687    return 0;
688
689  if (!IsSink)
690    Eng.WList->enqueue(Succ);
691
692  return Succ;
693}
694