C++ BranchInst::getSuccessor方法代码示例

//.........这里部分代码省略.........
            LoadInst* load = new LoadInst(liveIn, "", false, 4, term); 

            std::vector<Value *> produce_args;
            CastInst* cast = CastInst::CreateIntegerCast(load, int64Ty, true);
            cast->insertAfter(load);
            produce_args.push_back(cast);
            ConstantInt* val = ConstantInt::get(int32Ty, (uint32_t) 3);
            produce_args.push_back(val);
            CallInst::Create(pro, ArrayRef<Value*>(produce_args), "", term);

            produce_args.pop_back();
            val = ConstantInt::get(int32Ty, (uint32_t) 2);
            produce_args.push_back(val);
            CallInst::Create(pro, ArrayRef<Value*>(produce_args), "", term);
        }
    
    }

    // set branch instructions to restructure the CFG in created function
    BasicBlock* func1EndBlock = BasicBlock::Create(*context, "if.end.func1", func1); 
    BasicBlock* garbageBB = BasicBlock::Create(*context, "garbage", func1);
    ConstantInt* retVal_g = ConstantInt::get(int32Ty, (uint32_t) 0);
    ReturnInst* ret_g = ReturnInst::Create(*context, retVal_g, garbageBB);

    
    for (Function::iterator fit = func1->begin(), fite = func1->end(); fit != fite; ++fit) {
        if (fit->getTerminator() == NULL || fit->getName() == "garbage")
            continue;
      
        BranchInst* br = dyn_cast<BranchInst>(fit->getTerminator());
        int numSuccessors = br->getNumSuccessors();
        
        for (int i = 0; i < numSuccessors; i++) {
            BasicBlock* successor = br->getSuccessor(i);
            
            if (BlockMap[successor] != NULL) {
                
                br->setSuccessor(i, BlockMap[successor]);
            } 
            else {
                br->setSuccessor(i, func1EndBlock);
            }
            
        }
/*
        if (fit->getName() == "for.body.func1") {
            for (int i = 0; i < 4; i++) {
                BasicBlock::iterator it = fit->begin();
                it->moveBefore(ret_g);
            }
        }
        */
    }
    garbageBB->eraseFromParent();

    BranchInst* br = dyn_cast<BranchInst>(forBody->getTerminator());
    br->setSuccessor(0, forCond);
    forInc->eraseFromParent();


    // Create return instruction for func1EndBlock and set a branch from loop header to func1EndBlock
    ConstantInt* retVal = ConstantInt::get(int32Ty, (uint32_t) 0);
    ReturnInst* ret1 = ReturnInst::Create(*context, retVal, func1EndBlock);
    BasicBlock* loopHeader = BlockMap.at(L->getHeader());
    BranchInst* brInst = BranchInst::Create(loopHeader, func1EntryBlock);
    

//.........这里部分代码省略.........
  // Branch to either remainder (extra iterations) loop or unrolling loop.
  B.CreateCondBr(BranchVal, RemainderLoop, UnrollingLoop);
  PreHeaderBR->eraseFromParent();
  Function *F = Header->getParent();
  // Get an ordered list of blocks in the loop to help with the ordering of the
  // cloned blocks in the prolog/epilog code
  LoopBlocksDFS LoopBlocks(L);
  LoopBlocks.perform(LI);

  //
  // For each extra loop iteration, create a copy of the loop's basic blocks
  // and generate a condition that branches to the copy depending on the
  // number of 'left over' iterations.
  //
  std::vector<BasicBlock *> NewBlocks;
  ValueToValueMapTy VMap;

  // For unroll factor 2 remainder loop will have 1 iterations.
  // Do not create 1 iteration loop.
  bool CreateRemainderLoop = (Count != 2);

  // Clone all the basic blocks in the loop. If Count is 2, we don't clone
  // the loop, otherwise we create a cloned loop to execute the extra
  // iterations. This function adds the appropriate CFG connections.
  BasicBlock *InsertBot = UseEpilogRemainder ? Exit : PrologExit;
  BasicBlock *InsertTop = UseEpilogRemainder ? EpilogPreHeader : PrologPreHeader;
  CloneLoopBlocks(L, ModVal, CreateRemainderLoop, UseEpilogRemainder, InsertTop,
                  InsertBot, NewPreHeader, NewBlocks, LoopBlocks, VMap, LI);

  // Insert the cloned blocks into the function.
  F->getBasicBlockList().splice(InsertBot->getIterator(),
                                F->getBasicBlockList(),
                                NewBlocks[0]->getIterator(),
                                F->end());

  // Loop structure should be the following:
  //  Epilog             Prolog
  //
  // PreHeader         PreHeader
  // NewPreHeader      PrologPreHeader
  //   Header            PrologHeader
  //   ...               ...
  //   Latch             PrologLatch
  // NewExit           PrologExit
  // EpilogPreHeader   NewPreHeader
  //   EpilogHeader      Header
  //   ...               ...
  //   EpilogLatch       Latch
  // Exit              Exit

  // Rewrite the cloned instruction operands to use the values created when the
  // clone is created.
  for (BasicBlock *BB : NewBlocks) {
    for (Instruction &I : *BB) {
      RemapInstruction(&I, VMap,
                       RF_NoModuleLevelChanges | RF_IgnoreMissingLocals);
    }
  }

  if (UseEpilogRemainder) {
    // Connect the epilog code to the original loop and update the
    // PHI functions.
    ConnectEpilog(L, ModVal, NewExit, Exit, PreHeader,
                  EpilogPreHeader, NewPreHeader, VMap, DT, LI,
                  PreserveLCSSA);

    // Update counter in loop for unrolling.
    // I should be multiply of Count.
    IRBuilder<> B2(NewPreHeader->getTerminator());
    Value *TestVal = B2.CreateSub(TripCount, ModVal, "unroll_iter");
    BranchInst *LatchBR = cast<BranchInst>(Latch->getTerminator());
    B2.SetInsertPoint(LatchBR);
    PHINode *NewIdx = PHINode::Create(TestVal->getType(), 2, "niter",
                                      Header->getFirstNonPHI());
    Value *IdxSub =
        B2.CreateSub(NewIdx, ConstantInt::get(NewIdx->getType(), 1),
                     NewIdx->getName() + ".nsub");
    Value *IdxCmp;
    if (LatchBR->getSuccessor(0) == Header)
      IdxCmp = B2.CreateIsNotNull(IdxSub, NewIdx->getName() + ".ncmp");
    else
      IdxCmp = B2.CreateIsNull(IdxSub, NewIdx->getName() + ".ncmp");
    NewIdx->addIncoming(TestVal, NewPreHeader);
    NewIdx->addIncoming(IdxSub, Latch);
    LatchBR->setCondition(IdxCmp);
  } else {
    // Connect the prolog code to the original loop and update the
    // PHI functions.
    ConnectProlog(L, BECount, Count, PrologExit, PreHeader, NewPreHeader,
                  VMap, DT, LI, PreserveLCSSA);
  }

  // If this loop is nested, then the loop unroller changes the code in the
  // parent loop, so the Scalar Evolution pass needs to be run again.
  if (Loop *ParentLoop = L->getParentLoop())
    SE->forgetLoop(ParentLoop);

  NumRuntimeUnrolled++;
  return true;
}

/// HandleFloatingPointIV - If the loop has floating induction variable
/// then insert corresponding integer induction variable if possible.
/// For example,
/// for(double i = 0; i < 10000; ++i)
///   bar(i)
/// is converted into
/// for(int i = 0; i < 10000; ++i)
///   bar((double)i);
///
void IndVarSimplify::HandleFloatingPointIV(Loop *L, PHINode *PN) {
  unsigned IncomingEdge = L->contains(PN->getIncomingBlock(0));
  unsigned BackEdge     = IncomingEdge^1;

  // Check incoming value.
  ConstantFP *InitValueVal =
    dyn_cast<ConstantFP>(PN->getIncomingValue(IncomingEdge));

  int64_t InitValue;
  if (!InitValueVal || !ConvertToSInt(InitValueVal->getValueAPF(), InitValue))
    return;

  // Check IV increment. Reject this PN if increment operation is not
  // an add or increment value can not be represented by an integer.
  BinaryOperator *Incr =
    dyn_cast<BinaryOperator>(PN->getIncomingValue(BackEdge));
  if (Incr == 0 || Incr->getOpcode() != Instruction::FAdd) return;

  // If this is not an add of the PHI with a constantfp, or if the constant fp
  // is not an integer, bail out.
  ConstantFP *IncValueVal = dyn_cast<ConstantFP>(Incr->getOperand(1));
  int64_t IncValue;
  if (IncValueVal == 0 || Incr->getOperand(0) != PN ||
      !ConvertToSInt(IncValueVal->getValueAPF(), IncValue))
    return;

  // Check Incr uses. One user is PN and the other user is an exit condition
  // used by the conditional terminator.
  Value::use_iterator IncrUse = Incr->use_begin();
  Instruction *U1 = cast<Instruction>(*IncrUse++);
  if (IncrUse == Incr->use_end()) return;
  Instruction *U2 = cast<Instruction>(*IncrUse++);
  if (IncrUse != Incr->use_end()) return;

  // Find exit condition, which is an fcmp.  If it doesn't exist, or if it isn't
  // only used by a branch, we can't transform it.
  FCmpInst *Compare = dyn_cast<FCmpInst>(U1);
  if (!Compare)
    Compare = dyn_cast<FCmpInst>(U2);
  if (Compare == 0 || !Compare->hasOneUse() ||
      !isa<BranchInst>(Compare->use_back()))
    return;

  BranchInst *TheBr = cast<BranchInst>(Compare->use_back());

  // We need to verify that the branch actually controls the iteration count
  // of the loop.  If not, the new IV can overflow and no one will notice.
  // The branch block must be in the loop and one of the successors must be out
  // of the loop.
  assert(TheBr->isConditional() && "Can't use fcmp if not conditional");
  if (!L->contains(TheBr->getParent()) ||
      (L->contains(TheBr->getSuccessor(0)) &&
       L->contains(TheBr->getSuccessor(1))))
    return;


  // If it isn't a comparison with an integer-as-fp (the exit value), we can't
  // transform it.
  ConstantFP *ExitValueVal = dyn_cast<ConstantFP>(Compare->getOperand(1));
  int64_t ExitValue;
  if (ExitValueVal == 0 ||
      !ConvertToSInt(ExitValueVal->getValueAPF(), ExitValue))
    return;

  // Find new predicate for integer comparison.
  CmpInst::Predicate NewPred = CmpInst::BAD_ICMP_PREDICATE;
  switch (Compare->getPredicate()) {
  default: return;  // Unknown comparison.
  case CmpInst::FCMP_OEQ:
  case CmpInst::FCMP_UEQ: NewPred = CmpInst::ICMP_EQ; break;
  case CmpInst::FCMP_ONE:
  case CmpInst::FCMP_UNE: NewPred = CmpInst::ICMP_NE; break;
  case CmpInst::FCMP_OGT:
  case CmpInst::FCMP_UGT: NewPred = CmpInst::ICMP_SGT; break;
  case CmpInst::FCMP_OGE:
  case CmpInst::FCMP_UGE: NewPred = CmpInst::ICMP_SGE; break;
  case CmpInst::FCMP_OLT:
  case CmpInst::FCMP_ULT: NewPred = CmpInst::ICMP_SLT; break;
  case CmpInst::FCMP_OLE:
  case CmpInst::FCMP_ULE: NewPred = CmpInst::ICMP_SLE; break;
  }

  // We convert the floating point induction variable to a signed i32 value if
  // we can.  This is only safe if the comparison will not overflow in a way
  // that won't be trapped by the integer equivalent operations.  Check for this
  // now.
  // TODO: We could use i64 if it is native and the range requires it.

  // The start/stride/exit values must all fit in signed i32.
  if (!isInt<32>(InitValue) || !isInt<32>(IncValue) || !isInt<32>(ExitValue))
    return;
//.........这里部分代码省略.........

//.........这里部分代码省略.........
  // Report the unrolling decision.
  DebugLoc LoopLoc = L->getStartLoc();
  Function *F = Header->getParent();
  LLVMContext &Ctx = F->getContext();

  if (CompletelyUnroll) {
    DEBUG(dbgs() << "COMPLETELY UNROLLING loop %" << Header->getName()
          << " with trip count " << TripCount << "!\n");
    emitOptimizationRemark(Ctx, DEBUG_TYPE, *F, LoopLoc,
                           Twine("completely unrolled loop with ") +
                               Twine(TripCount) + " iterations");
  } else {
    auto EmitDiag = [&](const Twine &T) {
      emitOptimizationRemark(Ctx, DEBUG_TYPE, *F, LoopLoc,
                             "unrolled loop by a factor of " + Twine(Count) +
                                 T);
    };

    DEBUG(dbgs() << "UNROLLING loop %" << Header->getName()
          << " by " << Count);
    if (TripMultiple == 0 || BreakoutTrip != TripMultiple) {
      DEBUG(dbgs() << " with a breakout at trip " << BreakoutTrip);
      EmitDiag(" with a breakout at trip " + Twine(BreakoutTrip));
    } else if (TripMultiple != 1) {
      DEBUG(dbgs() << " with " << TripMultiple << " trips per branch");
      EmitDiag(" with " + Twine(TripMultiple) + " trips per branch");
    } else if (RuntimeTripCount) {
      DEBUG(dbgs() << " with run-time trip count");
      EmitDiag(" with run-time trip count");
    }
    DEBUG(dbgs() << "!\n");
  }

  bool ContinueOnTrue = L->contains(BI->getSuccessor(0));
  BasicBlock *LoopExit = BI->getSuccessor(ContinueOnTrue);

  // For the first iteration of the loop, we should use the precloned values for
  // PHI nodes.  Insert associations now.
  ValueToValueMapTy LastValueMap;
  std::vector<PHINode*> OrigPHINode;
  for (BasicBlock::iterator I = Header->begin(); isa<PHINode>(I); ++I) {
    OrigPHINode.push_back(cast<PHINode>(I));
  }

  std::vector<BasicBlock*> Headers;
  std::vector<BasicBlock*> Latches;
  Headers.push_back(Header);
  Latches.push_back(LatchBlock);

  // The current on-the-fly SSA update requires blocks to be processed in
  // reverse postorder so that LastValueMap contains the correct value at each
  // exit.
  LoopBlocksDFS DFS(L);
  DFS.perform(LI);

  // Stash the DFS iterators before adding blocks to the loop.
  LoopBlocksDFS::RPOIterator BlockBegin = DFS.beginRPO();
  LoopBlocksDFS::RPOIterator BlockEnd = DFS.endRPO();

  for (unsigned It = 1; It != Count; ++It) {
    std::vector<BasicBlock*> NewBlocks;
    SmallDenseMap<const Loop *, Loop *, 4> NewLoops;
    NewLoops[L] = L;

    for (LoopBlocksDFS::RPOIterator BB = BlockBegin; BB != BlockEnd; ++BB) {
      ValueToValueMapTy VMap;

//.........这里部分代码省略.........
          assert(InVal && "Unknown input value?");
          PN->setIncomingValue(pred, InVal);
          PN->setIncomingBlock(pred, MappedBlock);
        } else {
          PN->removeIncomingValue(pred, false);
          --pred, --e;  // Revisit the next entry.
        }
      } 
    }
    
    // The loop above has removed PHI entries for those blocks that are dead
    // and has updated others.  However, if a block is live (i.e. copied over)
    // but its terminator has been changed to not go to this block, then our
    // phi nodes will have invalid entries.  Update the PHI nodes in this
    // case.
    PHINode *PN = cast<PHINode>(NewBB->begin());
    NumPreds = std::distance(pred_begin(NewBB), pred_end(NewBB));
    if (NumPreds != PN->getNumIncomingValues()) {
      assert(NumPreds < PN->getNumIncomingValues());
      // Count how many times each predecessor comes to this block.
      std::map<BasicBlock*, unsigned> PredCount;
      for (pred_iterator PI = pred_begin(NewBB), E = pred_end(NewBB);
           PI != E; ++PI)
        --PredCount[*PI];
      
      // Figure out how many entries to remove from each PHI.
      for (unsigned i = 0, e = PN->getNumIncomingValues(); i != e; ++i)
        ++PredCount[PN->getIncomingBlock(i)];
      
      // At this point, the excess predecessor entries are positive in the
      // map.  Loop over all of the PHIs and remove excess predecessor
      // entries.
      BasicBlock::iterator I = NewBB->begin();
      for (; (PN = dyn_cast<PHINode>(I)); ++I) {
        for (std::map<BasicBlock*, unsigned>::iterator PCI =PredCount.begin(),
             E = PredCount.end(); PCI != E; ++PCI) {
          BasicBlock *Pred     = PCI->first;
          for (unsigned NumToRemove = PCI->second; NumToRemove; --NumToRemove)
            PN->removeIncomingValue(Pred, false);
        }
      }
    }
    
    // If the loops above have made these phi nodes have 0 or 1 operand,
    // replace them with undef or the input value.  We must do this for
    // correctness, because 0-operand phis are not valid.
    PN = cast<PHINode>(NewBB->begin());
    if (PN->getNumIncomingValues() == 0) {
      BasicBlock::iterator I = NewBB->begin();
      BasicBlock::const_iterator OldI = OldBB->begin();
      while ((PN = dyn_cast<PHINode>(I++))) {
        Value *NV = UndefValue::get(PN->getType());
        PN->replaceAllUsesWith(NV);
        assert(VMap[OldI] == PN && "VMap mismatch");
        VMap[OldI] = NV;
        PN->eraseFromParent();
        ++OldI;
      }
    }
    // NOTE: We cannot eliminate single entry phi nodes here, because of
    // VMap.  Single entry phi nodes can have multiple VMap entries
    // pointing at them.  Thus, deleting one would require scanning the VMap
    // to update any entries in it that would require that.  This would be
    // really slow.
  }
  
  // Now that the inlined function body has been fully constructed, go through
  // and zap unconditional fall-through branches.  This happen all the time when
  // specializing code: code specialization turns conditional branches into
  // uncond branches, and this code folds them.
  Function::iterator I = cast<BasicBlock>(VMap[&OldFunc->getEntryBlock()]);
  while (I != NewFunc->end()) {
    BranchInst *BI = dyn_cast<BranchInst>(I->getTerminator());
    if (!BI || BI->isConditional()) { ++I; continue; }
    
    // Note that we can't eliminate uncond branches if the destination has
    // single-entry PHI nodes.  Eliminating the single-entry phi nodes would
    // require scanning the VMap to update any entries that point to the phi
    // node.
    BasicBlock *Dest = BI->getSuccessor(0);
    if (!Dest->getSinglePredecessor() || isa<PHINode>(Dest->begin())) {
      ++I; continue;
    }
    
    // We know all single-entry PHI nodes in the inlined function have been
    // removed, so we just need to splice the blocks.
    BI->eraseFromParent();
    
    // Move all the instructions in the succ to the pred.
    I->getInstList().splice(I->end(), Dest->getInstList());
    
    // Make all PHI nodes that referred to Dest now refer to I as their source.
    Dest->replaceAllUsesWith(I);

    // Remove the dest block.
    Dest->eraseFromParent();
    
    // Do not increment I, iteratively merge all things this block branches to.
  }
}

//.........这里部分代码省略.........
    // We now have a loop-invariant count of loop iterations (which is not the
    // constant zero) for which we know that this loop will not exit via this
    // exisiting block.

    // We need to make sure that this block will run on every loop iteration.
    // For this to be true, we must dominate all blocks with backedges. Such
    // blocks are in-loop predecessors to the header block.
    bool NotAlways = false;
    for (pred_iterator PI = pred_begin(L->getHeader()),
         PIE = pred_end(L->getHeader()); PI != PIE; ++PI) {
      if (!L->contains(*PI))
        continue;

      if (!DT->dominates(*I, *PI)) {
        NotAlways = true;
        break;
      }
    }

    if (NotAlways)
      continue;

    // Make sure this blocks ends with a conditional branch.
    Instruction *TI = (*I)->getTerminator();
    if (!TI)
      continue;

    if (BranchInst *BI = dyn_cast<BranchInst>(TI)) {
      if (!BI->isConditional())
        continue;

      CountedExitBranch = BI;
    } else
      continue;

    // Note that this block may not be the loop latch block, even if the loop
    // has a latch block.
    CountedExitBlock = *I;
    ExitCount = EC;
    break;
  }

  if (!CountedExitBlock)
    return MadeChange;

  BasicBlock *Preheader = L->getLoopPreheader();

  // If we don't have a preheader, then insert one. If we already have a
  // preheader, then we can use it (except if the preheader contains a use of
  // the CTR register because some such uses might be reordered by the
  // selection DAG after the mtctr instruction).
  if (!Preheader || mightUseCTR(Preheader))
    Preheader = InsertPreheaderForLoop(L, DT, LI, PreserveLCSSA);
  if (!Preheader)
    return MadeChange;

  DEBUG(dbgs() << "Preheader for exit count: " << Preheader->getName() << "\n");

  // Insert the count into the preheader and replace the condition used by the
  // selected branch.
  MadeChange = true;

  SCEVExpander SCEVE(*SE, *DL, "loopcnt");
  LLVMContext &C = SE->getContext();
  Type *CountType = TM->isPPC64() ? Type::getInt64Ty(C) : Type::getInt32Ty(C);
  if (!ExitCount->getType()->isPointerTy() &&
      ExitCount->getType() != CountType)
    ExitCount = SE->getZeroExtendExpr(ExitCount, CountType);
  ExitCount = SE->getAddExpr(ExitCount, SE->getOne(CountType));
  Value *ECValue =
      SCEVE.expandCodeFor(ExitCount, CountType, Preheader->getTerminator());

  IRBuilder<> CountBuilder(Preheader->getTerminator());
  Module *M = Preheader->getParent()->getParent();
  Value *MTCTRFunc = Intrinsic::getDeclaration(M, Intrinsic::ppc_mtctr,
                                               CountType);
  CountBuilder.CreateCall(MTCTRFunc, ECValue);

  IRBuilder<> CondBuilder(CountedExitBranch);
  Value *DecFunc =
    Intrinsic::getDeclaration(M, Intrinsic::ppc_is_decremented_ctr_nonzero);
  Value *NewCond = CondBuilder.CreateCall(DecFunc, {});
  Value *OldCond = CountedExitBranch->getCondition();
  CountedExitBranch->setCondition(NewCond);

  // The false branch must exit the loop.
  if (!L->contains(CountedExitBranch->getSuccessor(0)))
    CountedExitBranch->swapSuccessors();

  // The old condition may be dead now, and may have even created a dead PHI
  // (the original induction variable).
  RecursivelyDeleteTriviallyDeadInstructions(OldCond);
  // Run through the basic blocks of the loop and see if any of them have dead
  // PHIs that can be removed.
  for (auto I : L->blocks())
    DeleteDeadPHIs(I);

  ++NumCTRLoops;
  return MadeChange;
}

bool LoopBogusCF::runOnLoop(Loop *loop, LPPassManager &LPM) {
  if (disableLoopBcf)
    return false;

  ++NumLoops;
  DEBUG(errs() << "LoopBogusCF: Dumping loop info\n");
  // DEBUG(loop->dump());

  BasicBlock *header = loop->getHeader();

  BranchInst *branch = dyn_cast<BranchInst>(header->getTerminator());
  if (!branch || !branch->isConditional()) {
    DEBUG(errs() << "\t Not trivial loop -- skipping\n");
    return false;
  }

  if (!loop->isLoopSimplifyForm()) {
    DEBUG(errs() << "\t Not simplified loop -- skipping\n");
    return false;
  }

  BasicBlock *exitBlock = loop->getUniqueExitBlock();
  if (!exitBlock) {
    DEBUG(errs() << "\t No unique exit block -- skipping\n");
    return false;
  }

  if (branch->getSuccessor(0) != exitBlock &&
      branch->getSuccessor(1) != exitBlock) {
    DEBUG(errs() << "\t Not trivial loop -- skipping\n");
    return false;
  }

  ++NumLoopsObf;
  // DEBUG(header->getParent()->viewCFG());

  DEBUG(errs() << "\tCreating dummy block\n");
  LoopInfo &info = getAnalysis<LoopInfo>();
  // Split header block
  BasicBlock *dummy = header->splitBasicBlock(header->getTerminator());
  loop->addBasicBlockToLoop(dummy, info.getBase());

  BasicBlock *trueBlock, *falseBlock = exitBlock;

  if (branch->getSuccessor(0) == exitBlock) {
    trueBlock = branch->getSuccessor(1);
    branch->setSuccessor(1, dummy);
  } else {
    trueBlock = branch->getSuccessor(0);
    branch->setSuccessor(0, dummy);
  }

  branch->moveBefore(header->getTerminator());
  header->getTerminator()->eraseFromParent();

  OpaquePredicate::createStub(dummy, trueBlock, falseBlock,
                              OpaquePredicate::PredicateTrue, false);

  // DEBUG(header->getParent()->viewCFG());

  return true;
}

//.........这里部分代码省略.........

    // We now have a loop-invariant count of loop iterations (which is not the
    // constant zero) for which we know that this loop will not exit via this
    // exisiting block.

    // We need to make sure that this block will run on every loop iteration.
    // For this to be true, we must dominate all blocks with backedges. Such
    // blocks are in-loop predecessors to the header block.
    bool NotAlways = false;
    for (pred_iterator PI = pred_begin(L->getHeader()),
         PIE = pred_end(L->getHeader()); PI != PIE; ++PI) {
      if (!L->contains(*PI))
        continue;

      if (!DT->dominates(*I, *PI)) {
        NotAlways = true;
        break;
      }
    }

    if (NotAlways)
      continue;

    // Make sure this blocks ends with a conditional branch.
    Instruction *TI = (*I)->getTerminator();
    if (!TI)
      continue;

    if (BranchInst *BI = dyn_cast<BranchInst>(TI)) {
      if (!BI->isConditional())
        continue;

      CountedExitBranch = BI;
    } else
      continue;

    // Note that this block may not be the loop latch block, even if the loop
    // has a latch block.
    CountedExitBlock = *I;
    ExitCount = EC;
    break;
  }

  if (!CountedExitBlock)
    return MadeChange;

  BasicBlock *Preheader = L->getLoopPreheader();

  // If we don't have a preheader, then insert one. If we already have a
  // preheader, then we can use it (except if the preheader contains a use of
  // the CTR register because some such uses might be reordered by the
  // selection DAG after the mtctr instruction).
  if (!Preheader || mightUseCTR(TT, Preheader))
    Preheader = InsertPreheaderForLoop(L, this);
  if (!Preheader)
    return MadeChange;

  DEBUG(dbgs() << "Preheader for exit count: " << Preheader->getName() << "\n");

  // Insert the count into the preheader and replace the condition used by the
  // selected branch.
  MadeChange = true;

  SCEVExpander SCEVE(*SE, "loopcnt");
  LLVMContext &C = SE->getContext();
  Type *CountType = TT.isArch64Bit() ? Type::getInt64Ty(C) :
                                       Type::getInt32Ty(C);
  if (!ExitCount->getType()->isPointerTy() &&
      ExitCount->getType() != CountType)
    ExitCount = SE->getZeroExtendExpr(ExitCount, CountType);
  ExitCount = SE->getAddExpr(ExitCount,
                             SE->getConstant(CountType, 1)); 
  Value *ECValue = SCEVE.expandCodeFor(ExitCount, CountType,
                                       Preheader->getTerminator());

  IRBuilder<> CountBuilder(Preheader->getTerminator());
  Module *M = Preheader->getParent()->getParent();
  Value *MTCTRFunc = Intrinsic::getDeclaration(M, Intrinsic::ppc_mtctr,
                                               CountType);
  CountBuilder.CreateCall(MTCTRFunc, ECValue);

  IRBuilder<> CondBuilder(CountedExitBranch);
  Value *DecFunc =
    Intrinsic::getDeclaration(M, Intrinsic::ppc_is_decremented_ctr_nonzero);
  Value *NewCond = CondBuilder.CreateCall(DecFunc);
  Value *OldCond = CountedExitBranch->getCondition();
  CountedExitBranch->setCondition(NewCond);

  // The false branch must exit the loop.
  if (!L->contains(CountedExitBranch->getSuccessor(0)))
    CountedExitBranch->swapSuccessors();

  // The old condition may be dead now, and may have even created a dead PHI
  // (the original induction variable).
  RecursivelyDeleteTriviallyDeadInstructions(OldCond);
  DeleteDeadPHIs(CountedExitBlock);

  ++NumCTRLoops;
  return MadeChange;
}

//.........这里部分代码省略.........
    // Only conditional branches are allowed beyond this point.
    assert(PBI->isConditional());

    // Condition's unique use should be the branch instruction.
    Value *PC = PBI->getCondition();
    if (!PC || !PC->hasOneUse())
      return false;

    if (PP && Preds.count(PP)) {
      // These are internal condition blocks to be merged from, e.g.,
      // BB2 in both cases.
      // Should not be address-taken.
      if (Pred->hasAddressTaken())
        return false;

      // Instructions in the internal condition blocks should be safe
      // to hoist up.
      for (BasicBlock::iterator BI = Pred->begin(), BE = PBI->getIterator();
           BI != BE;) {
        Instruction *CI = &*BI++;
        if (isa<PHINode>(CI) || !isSafeToSpeculativelyExecute(CI))
          return false;
      }
    } else {
      // This is the condition block to be merged into, e.g. BB1 in
      // both cases.
      if (FirstCondBlock)
        return false;
      FirstCondBlock = Pred;
    }

    // Find whether BB is uniformly on the true (or false) path
    // for all of its predecessors.
    BasicBlock *PS1 = PBI->getSuccessor(0);
    BasicBlock *PS2 = PBI->getSuccessor(1);
    BasicBlock *PS = (PS1 == BB) ? PS2 : PS1;
    int CIdx = (PS1 == BB) ? 0 : 1;

    if (Idx == -1)
      Idx = CIdx;
    else if (CIdx != Idx)
      return false;

    // PS is the successor which is not BB. Check successors to identify
    // the last conditional branch.
    if (Preds.count(PS) == 0) {
      // Case 2.
      LastCondBlock = Pred;
    } else {
      // Case 1
      BranchInst *BPS = dyn_cast<BranchInst>(PS->getTerminator());
      if (BPS && BPS->isUnconditional()) {
        // Case 1: PS(BB3) should be an unconditional branch.
        LastCondBlock = Pred;
      }
    }
  }

  if (!FirstCondBlock || !LastCondBlock || (FirstCondBlock == LastCondBlock))
    return false;

  TerminatorInst *TBB = LastCondBlock->getTerminator();
  BasicBlock *PS1 = TBB->getSuccessor(0);
  BasicBlock *PS2 = TBB->getSuccessor(1);
  BranchInst *PBI1 = dyn_cast<BranchInst>(PS1->getTerminator());
  BranchInst *PBI2 = dyn_cast<BranchInst>(PS2->getTerminator());

//.........这里部分代码省略.........
                      << "!\n");
    ORE->emit(OptimizationRemark(DEBUG_TYPE, "FullyUnrolled", L->getStartLoc(),
                                 L->getHeader())
              << "completely unroll and jammed loop with "
              << NV("UnrollCount", TripCount) << " iterations");
  } else {
    auto DiagBuilder = [&]() {
      OptimizationRemark Diag(DEBUG_TYPE, "PartialUnrolled", L->getStartLoc(),
                              L->getHeader());
      return Diag << "unroll and jammed loop by a factor of "
                  << NV("UnrollCount", Count);
    };

    LLVM_DEBUG(dbgs() << "UNROLL AND JAMMING loop %" << Header->getName()
                      << " by " << Count);
    if (TripMultiple != 1) {
      LLVM_DEBUG(dbgs() << " with " << TripMultiple << " trips per branch");
      ORE->emit([&]() {
        return DiagBuilder() << " with " << NV("TripMultiple", TripMultiple)
                             << " trips per branch";
      });
    } else {
      LLVM_DEBUG(dbgs() << " with run-time trip count");
      ORE->emit([&]() { return DiagBuilder() << " with run-time trip count"; });
    }
    LLVM_DEBUG(dbgs() << "!\n");
  }

  BasicBlock *Preheader = L->getLoopPreheader();
  BasicBlock *LatchBlock = L->getLoopLatch();
  BranchInst *BI = dyn_cast<BranchInst>(LatchBlock->getTerminator());
  assert(Preheader && LatchBlock && Header);
  assert(BI && !BI->isUnconditional());
  bool ContinueOnTrue = L->contains(BI->getSuccessor(0));
  BasicBlock *LoopExit = BI->getSuccessor(ContinueOnTrue);
  bool SubLoopContinueOnTrue = SubLoop->contains(
      SubLoop->getLoopLatch()->getTerminator()->getSuccessor(0));

  // Partition blocks in an outer/inner loop pair into blocks before and after
  // the loop
  BasicBlockSet SubLoopBlocks;
  BasicBlockSet ForeBlocks;
  BasicBlockSet AftBlocks;
  partitionOuterLoopBlocks(L, SubLoop, ForeBlocks, SubLoopBlocks, AftBlocks,
                           DT);

  // We keep track of the entering/first and exiting/last block of each of
  // Fore/SubLoop/Aft in each iteration. This helps make the stapling up of
  // blocks easier.
  std::vector<BasicBlock *> ForeBlocksFirst;
  std::vector<BasicBlock *> ForeBlocksLast;
  std::vector<BasicBlock *> SubLoopBlocksFirst;
  std::vector<BasicBlock *> SubLoopBlocksLast;
  std::vector<BasicBlock *> AftBlocksFirst;
  std::vector<BasicBlock *> AftBlocksLast;
  ForeBlocksFirst.push_back(Header);
  ForeBlocksLast.push_back(SubLoop->getLoopPreheader());
  SubLoopBlocksFirst.push_back(SubLoop->getHeader());
  SubLoopBlocksLast.push_back(SubLoop->getExitingBlock());
  AftBlocksFirst.push_back(SubLoop->getExitBlock());
  AftBlocksLast.push_back(L->getExitingBlock());
  // Maps Blocks[0] -> Blocks[It]
  ValueToValueMapTy LastValueMap;

  // Move any instructions from fore phi operands from AftBlocks into Fore.
  moveHeaderPhiOperandsToForeBlocks(

bool llvm::UnrollRuntimeLoopRemainder(Loop *L, unsigned Count,
                                      bool AllowExpensiveTripCount,
                                      bool UseEpilogRemainder,
                                      bool UnrollRemainder, LoopInfo *LI,
                                      ScalarEvolution *SE, DominatorTree *DT,
                                      AssumptionCache *AC, bool PreserveLCSSA,
                                      Loop **ResultLoop) {
  LLVM_DEBUG(dbgs() << "Trying runtime unrolling on Loop: \n");
  LLVM_DEBUG(L->dump());
  LLVM_DEBUG(UseEpilogRemainder ? dbgs() << "Using epilog remainder.\n"
                                : dbgs() << "Using prolog remainder.\n");

  // Make sure the loop is in canonical form.
  if (!L->isLoopSimplifyForm()) {
    LLVM_DEBUG(dbgs() << "Not in simplify form!\n");
    return false;
  }

  // Guaranteed by LoopSimplifyForm.
  BasicBlock *Latch = L->getLoopLatch();
  BasicBlock *Header = L->getHeader();

  BranchInst *LatchBR = cast<BranchInst>(Latch->getTerminator());

  if (!LatchBR || LatchBR->isUnconditional()) {
    // The loop-rotate pass can be helpful to avoid this in many cases.
    LLVM_DEBUG(
        dbgs()
        << "Loop latch not terminated by a conditional branch.\n");
    return false;
  }

  unsigned ExitIndex = LatchBR->getSuccessor(0) == Header ? 1 : 0;
  BasicBlock *LatchExit = LatchBR->getSuccessor(ExitIndex);

  if (L->contains(LatchExit)) {
    // Cloning the loop basic blocks (`CloneLoopBlocks`) requires that one of the
    // targets of the Latch be an exit block out of the loop.
    LLVM_DEBUG(
        dbgs()
        << "One of the loop latch successors must be the exit block.\n");
    return false;
  }

  // These are exit blocks other than the target of the latch exiting block.
  SmallVector<BasicBlock *, 4> OtherExits;
  bool isMultiExitUnrollingEnabled =
      canSafelyUnrollMultiExitLoop(L, OtherExits, LatchExit, PreserveLCSSA,
                                   UseEpilogRemainder) &&
      canProfitablyUnrollMultiExitLoop(L, OtherExits, LatchExit, PreserveLCSSA,
                                       UseEpilogRemainder);
  // Support only single exit and exiting block unless multi-exit loop unrolling is enabled.
  if (!isMultiExitUnrollingEnabled &&
      (!L->getExitingBlock() || OtherExits.size())) {
    LLVM_DEBUG(
        dbgs()
        << "Multiple exit/exiting blocks in loop and multi-exit unrolling not "
           "enabled!\n");
    return false;
  }
  // Use Scalar Evolution to compute the trip count. This allows more loops to
  // be unrolled than relying on induction var simplification.
  if (!SE)
    return false;

  // Only unroll loops with a computable trip count, and the trip count needs
  // to be an int value (allowing a pointer type is a TODO item).
  // We calculate the backedge count by using getExitCount on the Latch block,
  // which is proven to be the only exiting block in this loop. This is same as
  // calculating getBackedgeTakenCount on the loop (which computes SCEV for all
  // exiting blocks).
  const SCEV *BECountSC = SE->getExitCount(L, Latch);
  if (isa<SCEVCouldNotCompute>(BECountSC) ||
      !BECountSC->getType()->isIntegerTy()) {
    LLVM_DEBUG(dbgs() << "Could not compute exit block SCEV\n");
    return false;
  }

  unsigned BEWidth = cast<IntegerType>(BECountSC->getType())->getBitWidth();

  // Add 1 since the backedge count doesn't include the first loop iteration.
  const SCEV *TripCountSC =
      SE->getAddExpr(BECountSC, SE->getConstant(BECountSC->getType(), 1));
  if (isa<SCEVCouldNotCompute>(TripCountSC)) {
    LLVM_DEBUG(dbgs() << "Could not compute trip count SCEV.\n");
    return false;
  }

  BasicBlock *PreHeader = L->getLoopPreheader();
  BranchInst *PreHeaderBR = cast<BranchInst>(PreHeader->getTerminator());
  const DataLayout &DL = Header->getModule()->getDataLayout();
  SCEVExpander Expander(*SE, DL, "loop-unroll");
  if (!AllowExpensiveTripCount &&
      Expander.isHighCostExpansion(TripCountSC, L, PreHeaderBR)) {
    LLVM_DEBUG(dbgs() << "High cost for expanding trip count scev!\n");
    return false;
  }

  // This constraint lets us deal with an overflowing trip count easily; see the
  // comment on ModVal below.
//.........这里部分代码省略.........

bool FunctionStaticSlicer::computeBC() {
  bool changed = false;
#ifdef DEBUG_BC
  errs() << " ====== BEG BC Computation ======\n";
#endif
  if (!forward) {
    PostDominanceFrontier &PDF = MP->getAnalysis<PostDominanceFrontier>(fun);
    for (inst_iterator I = inst_begin(fun), E = inst_end(fun); I != E; I++) {
      Instruction *i = &*I;
      const InsInfo *ii = getInsInfo(i);
      if (ii->isSliced())
        continue;
      BasicBlock *BB = i->getParent();
#ifdef DEBUG_BC
      errs() << "  ";
      i->print(errs());
      errs() << " -> bb=" << BB->getName() << '\n';
#endif
      PostDominanceFrontier::const_iterator frontier = PDF.find(BB);
      if (frontier == PDF.end())
        continue;
      changed |= updateRCSC(frontier->second.begin(), frontier->second.end());
    }
  }
  else {
    for (inst_iterator I = inst_begin(fun), E = inst_end(fun); I != E; I++) {
      Instruction *i = &*I;
      if (!isa<BranchInst>(i))
        continue;
      BranchInst * bi = dyn_cast<BranchInst>(i);
      if (bi->isUnconditional()) // skip unconditional inst
        continue;
      const InsInfo *ii = getInsInfo(i);
      if (ii->isSliced())
        continue;
      unsigned succs = bi->getNumSuccessors(); 
      for (unsigned si = 0; si < succs; si++) {
        BasicBlock * BB = bi->getSuccessor(si);
        BasicBlock * Pred = BB->getUniquePredecessor();
        if (Pred == NULL)
          continue;
        for (BasicBlock::iterator BI = BB->begin(), BE = BB->end(); BI != BE; BI++) {
          InsInfo *bii = getInsInfo(&*BI);
          bii->deslice();
          for (ValSet::const_iterator VI = bii->DEF_begin(), VE = bii->DEF_end();
            VI != VE; VI++) {
            if (bii->addRC(*VI)) {
              changed = true;
#ifdef DEBUG_RC
              errs() << "  added ";
              printVal(*VI); 
              errs() << "\n";
#endif
            }
          }
        }
      }
    }
  }
#ifdef DEBUG_BC
  errs() << " ====== END BC Computation ======\n";
#endif
  return changed;
}

void Executor::executeBranch(Instruction *i, bool cond) {
    
    assert(i && "Expecting an instruction!");
    LastExecutedBB = i->getParent(); 
    if (CollectBlockTraces) {
        ExecutedBlocks.push_back(LastExecutedBB);
    }
    BranchInst *br = dyn_cast<BranchInst>(i);
    if (br->isConditional()) {        
        // Loops: check if the control-flow
        //        jump back to a loop header.
        std::string bbname = PathLoopInfo->getLoopLatchName(br);  
        BasicBlock *nextBB1 = br->getSuccessor(0);
        BasicBlock *nextBB2 = br->getSuccessor(1);
        if ((cond && bbname==nextBB1->getName().str()) 
            || (!cond && bbname==nextBB2->getName().str())) {
            unsigned line = 0;
            if (MDNode *N = br->getMetadata("dbg")) { 
                DILocation Loc(N); 
                line = Loc.getLineNumber();
            }
            //std::cout << "Current: " << LastExecutedBB->getName().str() << std::endl;
            //std::cout << "LoopLatch: " << bbname << std::endl;
            std::cout << "Violated property: unwinding assertion";
            std::cout << " (line " << line << ")\n";
            //exit(1);
            DisabledSymbolicExeCurRun = true;
            return;
        }
        // Next block to be executed
        if (cond) {
            NextBB = nextBB1;
        } else {
            NextBB = nextBB2;
        }
    } else {
        // Next block to be executed
        NextBB = br->getSuccessor(0);
    }
    if (!DisabledSymbolicExeCurRun) {
        PathContext->propagatePointers(LastExecutedBB, NextBB);
    }
    // For program coverage analysis
    //Profile->covered(br, cond);
    //Profile->covered(br->getParent());
    if(br->isConditional()) {
        // Update the number of time we jump in the IR
        GotoCount++;
        if (GotoCount>MaxDepth) {
            std::cout << "warning: maximum goto-count (";
            std::cout << GotoCount  << ") reached.\n";
            DisabledSymbolicExeCurRun = true;
            return;
        }
        if (DisabledSymbolicExeCurRun) {
            return;
        }
        Value *vcond = br->getCondition();
        // vcond in domain(S) 
        if (SMAP->contains(vcond)) {
            SymbolPtr Svcond = SMAP->get(vcond);
            // Report(S(vcond), branchTaken)
            Path->addBranchCell(Svcond, cond, i);
        } 
    }
}

/// [LIBUNWIND] Find the (possibly absent) call to @llvm.eh.selector
/// in the given landing pad.  In principle, llvm.eh.exception is
/// required to be in the landing pad; in practice, SplitCriticalEdge
/// can break that invariant, and then inlining can break it further.
/// There's a real need for a reliable solution here, but until that
/// happens, we have some fragile workarounds here.
static EHSelectorInst *findSelectorForLandingPad(BasicBlock *lpad) {
  // Look for an exception call in the actual landing pad.
  EHExceptionInst *exn = findExceptionInBlock(lpad);
  if (exn) return findSelectorForException(exn);

  // Okay, if that failed, look for one in an obvious successor.  If
  // we find one, we'll fix the IR by moving things back to the
  // landing pad.

  bool dominates = true; // does the lpad dominate the exn call
  BasicBlock *nonDominated = 0; // if not, the first non-dominated block
  BasicBlock *lastDominated = 0; // and the block which branched to it

  BasicBlock *exnBlock = lpad;

  // We need to protect against lpads that lead into infinite loops.
  SmallPtrSet<BasicBlock*,4> visited;
  visited.insert(exnBlock);

  do {
    // We're not going to apply this hack to anything more complicated
    // than a series of unconditional branches, so if the block
    // doesn't terminate in an unconditional branch, just fail.  More
    // complicated cases can arise when, say, sinking a call into a
    // split unwind edge and then inlining it; but that can do almost
    // *anything* to the CFG, including leaving the selector
    // completely unreachable.  The only way to fix that properly is
    // to (1) prohibit transforms which move the exception or selector
    // values away from the landing pad, e.g. by producing them with
    // instructions that are pinned to an edge like a phi, or
    // producing them with not-really-instructions, and (2) making
    // transforms which split edges deal with that.
    BranchInst *branch = dyn_cast<BranchInst>(&exnBlock->back());
    if (!branch || branch->isConditional()) return 0;

    BasicBlock *successor = branch->getSuccessor(0);

    // Fail if we found an infinite loop.
    if (!visited.insert(successor)) return 0;

    // If the successor isn't dominated by exnBlock:
    if (!successor->getSinglePredecessor()) {
      // We don't want to have to deal with threading the exception
      // through multiple levels of phi, so give up if we've already
      // followed a non-dominating edge.
      if (!dominates) return 0;

      // Otherwise, remember this as a non-dominating edge.
      dominates = false;
      nonDominated = successor;
      lastDominated = exnBlock;
    }

    exnBlock = successor;

    // Can we stop here?
    exn = findExceptionInBlock(exnBlock);
  } while (!exn);

  // Look for a selector call for the exception we found.
  EHSelectorInst *selector = findSelectorForException(exn);
  if (!selector) return 0;

  // The easy case is when the landing pad still dominates the
  // exception call, in which case we can just move both calls back to
  // the landing pad.
  if (dominates) {
    selector->moveBefore(lpad->getFirstNonPHI());
    exn->moveBefore(selector);
    return selector;
  }

  // Otherwise, we have to split at the first non-dominating block.
  // The CFG looks basically like this:
  //    lpad:
  //      phis_0
  //      insnsAndBranches_1
  //      br label %nonDominated
  //    nonDominated:
  //      phis_2
  //      insns_3
  //      %exn = call i8* @llvm.eh.exception()
  //      insnsAndBranches_4
  //      %selector = call @llvm.eh.selector(i8* %exn, ...
  // We need to turn this into:
  //    lpad:
  //      phis_0
  //      %exn0 = call i8* @llvm.eh.exception()
  //      %selector0 = call @llvm.eh.selector(i8* %exn0, ...
  //      insnsAndBranches_1
  //      br label %split // from lastDominated
  //    nonDominated:
  //      phis_2 (without edge from lastDominated)
  //      %exn1 = call i8* @llvm.eh.exception()
//.........这里部分代码省略.........

//.........这里部分代码省略.........
  ScalarEvolution *SE = LPM->getAnalysisIfAvailable<ScalarEvolution>();
  if (SE)
    SE->forgetLoop(L);

  // If we know the trip count, we know the multiple...
  unsigned BreakoutTrip = 0;
  if (TripCount != 0) {
    BreakoutTrip = TripCount % Count;
    TripMultiple = 0;
  } else {
    // Figure out what multiple to use.
    BreakoutTrip = TripMultiple =
      (unsigned)GreatestCommonDivisor64(Count, TripMultiple);
  }

  if (CompletelyUnroll) {
    DEBUG(dbgs() << "COMPLETELY UNROLLING loop %" << Header->getName()
          << " with trip count " << TripCount << "!\n");
  } else {
    DEBUG(dbgs() << "UNROLLING loop %" << Header->getName()
          << " by " << Count);
    if (TripMultiple == 0 || BreakoutTrip != TripMultiple) {
      DEBUG(dbgs() << " with a breakout at trip " << BreakoutTrip);
    } else if (TripMultiple != 1) {
      DEBUG(dbgs() << " with " << TripMultiple << " trips per branch");
    } else if (RuntimeTripCount) {
      DEBUG(dbgs() << " with run-time trip count");
    }
    DEBUG(dbgs() << "!\n");
  }

  std::vector<BasicBlock*> LoopBlocks = L->getBlocks();

  bool ContinueOnTrue = L->contains(BI->getSuccessor(0));
  BasicBlock *LoopExit = BI->getSuccessor(ContinueOnTrue);

  // For the first iteration of the loop, we should use the precloned values for
  // PHI nodes.  Insert associations now.
  ValueToValueMapTy LastValueMap;
  std::vector<PHINode*> OrigPHINode;
  for (BasicBlock::iterator I = Header->begin(); isa<PHINode>(I); ++I) {
    OrigPHINode.push_back(cast<PHINode>(I));
  }

  std::vector<BasicBlock*> Headers;
  std::vector<BasicBlock*> Latches;
  Headers.push_back(Header);
  Latches.push_back(LatchBlock);

  // The current on-the-fly SSA update requires blocks to be processed in
  // reverse postorder so that LastValueMap contains the correct value at each
  // exit.
  LoopBlocksDFS DFS(L);
  DFS.perform(LI);

  // Stash the DFS iterators before adding blocks to the loop.
  LoopBlocksDFS::RPOIterator BlockBegin = DFS.beginRPO();
  LoopBlocksDFS::RPOIterator BlockEnd = DFS.endRPO();

  for (unsigned It = 1; It != Count; ++It) {
    std::vector<BasicBlock*> NewBlocks;

    for (LoopBlocksDFS::RPOIterator BB = BlockBegin; BB != BlockEnd; ++BB) {
      ValueToValueMapTy VMap;
      BasicBlock *New = CloneBasicBlock(*BB, VMap, "." + Twine(It));
      Header->getParent()->getBasicBlockList().push_back(New);