CoreEngine.cpp revision fdf6a279c9a75c778eba382d9a156697092982a1 
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/CoreEngine.h"
18#include "clang/AST/Expr.h"
19#include "clang/AST/StmtCXX.h"
20#include "clang/StaticAnalyzer/Core/PathSensitive/AnalysisManager.h"
21#include "clang/StaticAnalyzer/Core/PathSensitive/ExprEngine.h"
22#include "llvm/ADT/DenseMap.h"
23#include "llvm/ADT/Statistic.h"
24#include "llvm/Support/Casting.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_back(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      SubEng.processCallEnter(CEnter, Pred);
247      break;
248    }
249
250    case ProgramPoint::CallExitBeginKind:
251      SubEng.processCallExit(Pred);
252      break;
253
254    case ProgramPoint::EpsilonKind: {
255      assert(Pred->hasSinglePred() &&
256             "Assume epsilon has exactly one predecessor by construction");
257      ExplodedNode *PNode = Pred->getFirstPred();
258      dispatchWorkItem(Pred, PNode->getLocation(), WU);
259      break;
260    }
261    default:
262      assert(isa<PostStmt>(Loc) ||
263             isa<PostInitializer>(Loc) ||
264             isa<PostImplicitCall>(Loc) ||
265             isa<CallExitEnd>(Loc));
266      HandlePostStmt(WU.getBlock(), WU.getIndex(), Pred);
267      break;
268  }
269}
270
271bool CoreEngine::ExecuteWorkListWithInitialState(const LocationContext *L,
272                                                 unsigned Steps,
273                                                 ProgramStateRef InitState,
274                                                 ExplodedNodeSet &Dst) {
275  bool DidNotFinish = ExecuteWorkList(L, Steps, InitState);
276  for (ExplodedGraph::eop_iterator I = G->eop_begin(),
277                                   E = G->eop_end(); I != E; ++I) {
278    Dst.Add(*I);
279  }
280  return DidNotFinish;
281}
282
283void CoreEngine::HandleBlockEdge(const BlockEdge &L, ExplodedNode *Pred) {
284
285  const CFGBlock *Blk = L.getDst();
286  NodeBuilderContext BuilderCtx(*this, Blk, Pred);
287
288  // Mark this block as visited.
289  const LocationContext *LC = Pred->getLocationContext();
290  FunctionSummaries->markVisitedBasicBlock(Blk->getBlockID(),
291                                           LC->getDecl(),
292                                           LC->getCFG()->getNumBlockIDs());
293
294  // Check if we are entering the EXIT block.
295  if (Blk == &(L.getLocationContext()->getCFG()->getExit())) {
296
297    assert (L.getLocationContext()->getCFG()->getExit().size() == 0
298            && "EXIT block cannot contain Stmts.");
299
300    // Process the final state transition.
301    SubEng.processEndOfFunction(BuilderCtx, Pred);
302
303    // This path is done. Don't enqueue any more nodes.
304    return;
305  }
306
307  // Call into the SubEngine to process entering the CFGBlock.
308  ExplodedNodeSet dstNodes;
309  BlockEntrance BE(Blk, Pred->getLocationContext());
310  NodeBuilderWithSinks nodeBuilder(Pred, dstNodes, BuilderCtx, BE);
311  SubEng.processCFGBlockEntrance(L, nodeBuilder, Pred);
312
313  // Auto-generate a node.
314  if (!nodeBuilder.hasGeneratedNodes()) {
315    nodeBuilder.generateNode(Pred->State, Pred);
316  }
317
318  // Enqueue nodes onto the worklist.
319  enqueue(dstNodes);
320}
321
322void CoreEngine::HandleBlockEntrance(const BlockEntrance &L,
323                                       ExplodedNode *Pred) {
324
325  // Increment the block counter.
326  const LocationContext *LC = Pred->getLocationContext();
327  unsigned BlockId = L.getBlock()->getBlockID();
328  BlockCounter Counter = WList->getBlockCounter();
329  Counter = BCounterFactory.IncrementCount(Counter, LC->getCurrentStackFrame(),
330                                           BlockId);
331  WList->setBlockCounter(Counter);
332
333  // Process the entrance of the block.
334  if (CFGElement E = L.getFirstElement()) {
335    NodeBuilderContext Ctx(*this, L.getBlock(), Pred);
336    SubEng.processCFGElement(E, Pred, 0, &Ctx);
337  }
338  else
339    HandleBlockExit(L.getBlock(), Pred);
340}
341
342void CoreEngine::HandleBlockExit(const CFGBlock * B, ExplodedNode *Pred) {
343
344  if (const Stmt *Term = B->getTerminator()) {
345    switch (Term->getStmtClass()) {
346      default:
347        llvm_unreachable("Analysis for this terminator not implemented.");
348
349      case Stmt::BinaryOperatorClass: // '&&' and '||'
350        HandleBranch(cast<BinaryOperator>(Term)->getLHS(), Term, B, Pred);
351        return;
352
353      case Stmt::BinaryConditionalOperatorClass:
354      case Stmt::ConditionalOperatorClass:
355        HandleBranch(cast<AbstractConditionalOperator>(Term)->getCond(),
356                     Term, B, Pred);
357        return;
358
359        // FIXME: Use constant-folding in CFG construction to simplify this
360        // case.
361
362      case Stmt::ChooseExprClass:
363        HandleBranch(cast<ChooseExpr>(Term)->getCond(), Term, B, Pred);
364        return;
365
366      case Stmt::CXXTryStmtClass: {
367        // Generate a node for each of the successors.
368        // Our logic for EH analysis can certainly be improved.
369        for (CFGBlock::const_succ_iterator it = B->succ_begin(),
370             et = B->succ_end(); it != et; ++it) {
371          if (const CFGBlock *succ = *it) {
372            generateNode(BlockEdge(B, succ, Pred->getLocationContext()),
373                         Pred->State, Pred);
374          }
375        }
376        return;
377      }
378
379      case Stmt::DoStmtClass:
380        HandleBranch(cast<DoStmt>(Term)->getCond(), Term, B, Pred);
381        return;
382
383      case Stmt::CXXForRangeStmtClass:
384        HandleBranch(cast<CXXForRangeStmt>(Term)->getCond(), Term, B, Pred);
385        return;
386
387      case Stmt::ForStmtClass:
388        HandleBranch(cast<ForStmt>(Term)->getCond(), Term, B, Pred);
389        return;
390
391      case Stmt::ContinueStmtClass:
392      case Stmt::BreakStmtClass:
393      case Stmt::GotoStmtClass:
394        break;
395
396      case Stmt::IfStmtClass:
397        HandleBranch(cast<IfStmt>(Term)->getCond(), Term, B, Pred);
398        return;
399
400      case Stmt::IndirectGotoStmtClass: {
401        // Only 1 successor: the indirect goto dispatch block.
402        assert (B->succ_size() == 1);
403
404        IndirectGotoNodeBuilder
405           builder(Pred, B, cast<IndirectGotoStmt>(Term)->getTarget(),
406                   *(B->succ_begin()), this);
407
408        SubEng.processIndirectGoto(builder);
409        return;
410      }
411
412      case Stmt::ObjCForCollectionStmtClass: {
413        // In the case of ObjCForCollectionStmt, it appears twice in a CFG:
414        //
415        //  (1) inside a basic block, which represents the binding of the
416        //      'element' variable to a value.
417        //  (2) in a terminator, which represents the branch.
418        //
419        // For (1), subengines will bind a value (i.e., 0 or 1) indicating
420        // whether or not collection contains any more elements.  We cannot
421        // just test to see if the element is nil because a container can
422        // contain nil elements.
423        HandleBranch(Term, Term, B, Pred);
424        return;
425      }
426
427      case Stmt::SwitchStmtClass: {
428        SwitchNodeBuilder builder(Pred, B, cast<SwitchStmt>(Term)->getCond(),
429                                    this);
430
431        SubEng.processSwitch(builder);
432        return;
433      }
434
435      case Stmt::WhileStmtClass:
436        HandleBranch(cast<WhileStmt>(Term)->getCond(), Term, B, Pred);
437        return;
438    }
439  }
440
441  assert (B->succ_size() == 1 &&
442          "Blocks with no terminator should have at most 1 successor.");
443
444  generateNode(BlockEdge(B, *(B->succ_begin()), Pred->getLocationContext()),
445               Pred->State, Pred);
446}
447
448void CoreEngine::HandleBranch(const Stmt *Cond, const Stmt *Term,
449                                const CFGBlock * B, ExplodedNode *Pred) {
450  assert(B->succ_size() == 2);
451  NodeBuilderContext Ctx(*this, B, Pred);
452  ExplodedNodeSet Dst;
453  SubEng.processBranch(Cond, Term, Ctx, Pred, Dst,
454                       *(B->succ_begin()), *(B->succ_begin()+1));
455  // Enqueue the new frontier onto the worklist.
456  enqueue(Dst);
457}
458
459void CoreEngine::HandlePostStmt(const CFGBlock *B, unsigned StmtIdx,
460                                  ExplodedNode *Pred) {
461  assert(B);
462  assert(!B->empty());
463
464  if (StmtIdx == B->size())
465    HandleBlockExit(B, Pred);
466  else {
467    NodeBuilderContext Ctx(*this, B, Pred);
468    SubEng.processCFGElement((*B)[StmtIdx], Pred, StmtIdx, &Ctx);
469  }
470}
471
472/// generateNode - Utility method to generate nodes, hook up successors,
473///  and add nodes to the worklist.
474void CoreEngine::generateNode(const ProgramPoint &Loc,
475                              ProgramStateRef State,
476                              ExplodedNode *Pred) {
477
478  bool IsNew;
479  ExplodedNode *Node = G->getNode(Loc, State, false, &IsNew);
480
481  if (Pred)
482    Node->addPredecessor(Pred, *G);  // Link 'Node' with its predecessor.
483  else {
484    assert (IsNew);
485    G->addRoot(Node);  // 'Node' has no predecessor.  Make it a root.
486  }
487
488  // Only add 'Node' to the worklist if it was freshly generated.
489  if (IsNew) WList->enqueue(Node);
490}
491
492void CoreEngine::enqueueStmtNode(ExplodedNode *N,
493                                 const CFGBlock *Block, unsigned Idx) {
494  assert(Block);
495  assert (!N->isSink());
496
497  // Check if this node entered a callee.
498  if (isa<CallEnter>(N->getLocation())) {
499    // Still use the index of the CallExpr. It's needed to create the callee
500    // StackFrameContext.
501    WList->enqueue(N, Block, Idx);
502    return;
503  }
504
505  // Do not create extra nodes. Move to the next CFG element.
506  if (isa<PostInitializer>(N->getLocation()) ||
507      isa<PostImplicitCall>(N->getLocation())) {
508    WList->enqueue(N, Block, Idx+1);
509    return;
510  }
511
512  if (isa<EpsilonPoint>(N->getLocation())) {
513    WList->enqueue(N, Block, Idx);
514    return;
515  }
516
517  // At this point, we know we're processing a normal statement.
518  CFGStmt CS = (*Block)[Idx].castAs<CFGStmt>();
519  PostStmt Loc(CS.getStmt(), N->getLocationContext());
520
521  if (Loc == N->getLocation()) {
522    // Note: 'N' should be a fresh node because otherwise it shouldn't be
523    // a member of Deferred.
524    WList->enqueue(N, Block, Idx+1);
525    return;
526  }
527
528  bool IsNew;
529  ExplodedNode *Succ = G->getNode(Loc, N->getState(), false, &IsNew);
530  Succ->addPredecessor(N, *G);
531
532  if (IsNew)
533    WList->enqueue(Succ, Block, Idx+1);
534}
535
536ExplodedNode *CoreEngine::generateCallExitBeginNode(ExplodedNode *N) {
537  // Create a CallExitBegin node and enqueue it.
538  const StackFrameContext *LocCtx
539                         = cast<StackFrameContext>(N->getLocationContext());
540
541  // Use the callee location context.
542  CallExitBegin Loc(LocCtx);
543
544  bool isNew;
545  ExplodedNode *Node = G->getNode(Loc, N->getState(), false, &isNew);
546  Node->addPredecessor(N, *G);
547  return isNew ? Node : 0;
548}
549
550
551void CoreEngine::enqueue(ExplodedNodeSet &Set) {
552  for (ExplodedNodeSet::iterator I = Set.begin(),
553                                 E = Set.end(); I != E; ++I) {
554    WList->enqueue(*I);
555  }
556}
557
558void CoreEngine::enqueue(ExplodedNodeSet &Set,
559                         const CFGBlock *Block, unsigned Idx) {
560  for (ExplodedNodeSet::iterator I = Set.begin(),
561                                 E = Set.end(); I != E; ++I) {
562    enqueueStmtNode(*I, Block, Idx);
563  }
564}
565
566void CoreEngine::enqueueEndOfFunction(ExplodedNodeSet &Set) {
567  for (ExplodedNodeSet::iterator I = Set.begin(), E = Set.end(); I != E; ++I) {
568    ExplodedNode *N = *I;
569    // If we are in an inlined call, generate CallExitBegin node.
570    if (N->getLocationContext()->getParent()) {
571      N = generateCallExitBeginNode(N);
572      if (N)
573        WList->enqueue(N);
574    } else {
575      // TODO: We should run remove dead bindings here.
576      G->addEndOfPath(N);
577      NumPathsExplored++;
578    }
579  }
580}
581
582
583void NodeBuilder::anchor() { }
584
585ExplodedNode* NodeBuilder::generateNodeImpl(const ProgramPoint &Loc,
586                                            ProgramStateRef State,
587                                            ExplodedNode *FromN,
588                                            bool MarkAsSink) {
589  HasGeneratedNodes = true;
590  bool IsNew;
591  ExplodedNode *N = C.Eng.G->getNode(Loc, State, MarkAsSink, &IsNew);
592  N->addPredecessor(FromN, *C.Eng.G);
593  Frontier.erase(FromN);
594
595  if (!IsNew)
596    return 0;
597
598  if (!MarkAsSink)
599    Frontier.Add(N);
600
601  return N;
602}
603
604void NodeBuilderWithSinks::anchor() { }
605
606StmtNodeBuilder::~StmtNodeBuilder() {
607  if (EnclosingBldr)
608    for (ExplodedNodeSet::iterator I = Frontier.begin(),
609                                   E = Frontier.end(); I != E; ++I )
610      EnclosingBldr->addNodes(*I);
611}
612
613void BranchNodeBuilder::anchor() { }
614
615ExplodedNode *BranchNodeBuilder::generateNode(ProgramStateRef State,
616                                              bool branch,
617                                              ExplodedNode *NodePred) {
618  // If the branch has been marked infeasible we should not generate a node.
619  if (!isFeasible(branch))
620    return NULL;
621
622  ProgramPoint Loc = BlockEdge(C.Block, branch ? DstT:DstF,
623                               NodePred->getLocationContext());
624  ExplodedNode *Succ = generateNodeImpl(Loc, State, NodePred);
625  return Succ;
626}
627
628ExplodedNode*
629IndirectGotoNodeBuilder::generateNode(const iterator &I,
630                                      ProgramStateRef St,
631                                      bool IsSink) {
632  bool IsNew;
633  ExplodedNode *Succ = Eng.G->getNode(BlockEdge(Src, I.getBlock(),
634                                      Pred->getLocationContext()), St,
635                                      IsSink, &IsNew);
636  Succ->addPredecessor(Pred, *Eng.G);
637
638  if (!IsNew)
639    return 0;
640
641  if (!IsSink)
642    Eng.WList->enqueue(Succ);
643
644  return Succ;
645}
646
647
648ExplodedNode*
649SwitchNodeBuilder::generateCaseStmtNode(const iterator &I,
650                                        ProgramStateRef St) {
651
652  bool IsNew;
653  ExplodedNode *Succ = Eng.G->getNode(BlockEdge(Src, I.getBlock(),
654                                      Pred->getLocationContext()), St,
655                                      false, &IsNew);
656  Succ->addPredecessor(Pred, *Eng.G);
657  if (!IsNew)
658    return 0;
659
660  Eng.WList->enqueue(Succ);
661  return Succ;
662}
663
664
665ExplodedNode*
666SwitchNodeBuilder::generateDefaultCaseNode(ProgramStateRef St,
667                                           bool IsSink) {
668  // Get the block for the default case.
669  assert(Src->succ_rbegin() != Src->succ_rend());
670  CFGBlock *DefaultBlock = *Src->succ_rbegin();
671
672  // Sanity check for default blocks that are unreachable and not caught
673  // by earlier stages.
674  if (!DefaultBlock)
675    return NULL;
676
677  bool IsNew;
678  ExplodedNode *Succ = Eng.G->getNode(BlockEdge(Src, DefaultBlock,
679                                      Pred->getLocationContext()), St,
680                                      IsSink, &IsNew);
681  Succ->addPredecessor(Pred, *Eng.G);
682
683  if (!IsNew)
684    return 0;
685
686  if (!IsSink)
687    Eng.WList->enqueue(Succ);
688
689  return Succ;
690}
691