C++ BasicBlock::eraseFromParent方法代码示例

void LowerEmAsyncify::FindContextVariables(AsyncCallEntry & Entry) {
  BasicBlock *AfterCallBlock = Entry.AfterCallBlock;

  Function & F = *AfterCallBlock->getParent();

  // Create a new entry block as if in the callback function
  // theck check variables that no longer properly dominate their uses
  BasicBlock *EntryBlock = BasicBlock::Create(TheModule->getContext(), "", &F, &F.getEntryBlock());
  BranchInst::Create(AfterCallBlock, EntryBlock);

  DominatorTreeWrapperPass DTW;
  DTW.runOnFunction(F);
  DominatorTree& DT = DTW.getDomTree();

  // These blocks may be using some values defined at or before AsyncCallBlock
  BasicBlockSet Ramifications = FindReachableBlocksFrom(AfterCallBlock); 

  SmallPtrSet<Value*, 256> ContextVariables;
  Values Pending;

  // Examine the instructions, find all variables that we need to store in the context
  for (BasicBlockSet::iterator RI = Ramifications.begin(), RE = Ramifications.end(); RI != RE; ++RI) {
    for (BasicBlock::iterator I = (*RI)->begin(), E = (*RI)->end(); I != E; ++I) {
      for (unsigned i = 0, NumOperands = I->getNumOperands(); i < NumOperands; ++i) {
        Value *O = I->getOperand(i);
        if (Instruction *Inst = dyn_cast<Instruction>(O)) {
          if (Inst == Entry.AsyncCallInst) continue; // for the original async call, we will load directly from async return value
          if (ContextVariables.count(Inst) != 0)  continue; // already examined 

          if (!DT.dominates(Inst, I->getOperandUse(i))) {
            // `I` is using `Inst`, yet `Inst` does not dominate `I` if we arrive directly at AfterCallBlock
            // so we need to save `Inst` in the context
            ContextVariables.insert(Inst);
            Pending.push_back(Inst);
          }
        } else if (Argument *Arg = dyn_cast<Argument>(O)) {
          // count() should be as fast/slow as insert, so just insert here 
          ContextVariables.insert(Arg);
        }
      }
    }
  }

  // restore F
  EntryBlock->eraseFromParent();  

  Entry.ContextVariables.clear();
  Entry.ContextVariables.reserve(ContextVariables.size());
  for (SmallPtrSet<Value*, 256>::iterator I = ContextVariables.begin(), E = ContextVariables.end(); I != E; ++I) {
    Entry.ContextVariables.push_back(*I);
  }
}

/// Fold the loop tail into the loop exit by speculating the loop tail
/// instructions. Typically, this is a single post-increment. In the case of a
/// simple 2-block loop, hoisting the increment can be much better than
/// duplicating the entire loop header. In the case of loops with early exits,
/// rotation will not work anyway, but simplifyLoopLatch will put the loop in
/// canonical form so downstream passes can handle it.
///
/// I don't believe this invalidates SCEV.
bool LoopRotate::simplifyLoopLatch(Loop *L) {
  BasicBlock *Latch = L->getLoopLatch();
  if (!Latch || Latch->hasAddressTaken())
    return false;

  BranchInst *Jmp = dyn_cast<BranchInst>(Latch->getTerminator());
  if (!Jmp || !Jmp->isUnconditional())
    return false;

  BasicBlock *LastExit = Latch->getSinglePredecessor();
  if (!LastExit || !L->isLoopExiting(LastExit))
    return false;

  BranchInst *BI = dyn_cast<BranchInst>(LastExit->getTerminator());
  if (!BI)
    return false;

  if (!shouldSpeculateInstrs(Latch->begin(), Jmp))
    return false;

  DEBUG(dbgs() << "Folding loop latch " << Latch->getName() << " into "
        << LastExit->getName() << "\n");

  // Hoist the instructions from Latch into LastExit.
  LastExit->getInstList().splice(BI, Latch->getInstList(), Latch->begin(), Jmp);

  unsigned FallThruPath = BI->getSuccessor(0) == Latch ? 0 : 1;
  BasicBlock *Header = Jmp->getSuccessor(0);
  assert(Header == L->getHeader() && "expected a backward branch");

  // Remove Latch from the CFG so that LastExit becomes the new Latch.
  BI->setSuccessor(FallThruPath, Header);
  Latch->replaceSuccessorsPhiUsesWith(LastExit);
  Jmp->eraseFromParent();

  // Nuke the Latch block.
  assert(Latch->empty() && "unable to evacuate Latch");
  LI->removeBlock(Latch);
  if (DominatorTreeWrapperPass *DTWP =
          getAnalysisIfAvailable<DominatorTreeWrapperPass>())
    DTWP->getDomTree().eraseNode(Latch);
  Latch->eraseFromParent();
  return true;
}

/// MergeBasicBlockIntoOnlyPred - DestBB is a block with one predecessor and its
/// predecessor is known to have one successor (DestBB!).  Eliminate the edge
/// between them, moving the instructions in the predecessor into DestBB and
/// deleting the predecessor block.
///
void llvm::MergeBasicBlockIntoOnlyPred(BasicBlock *DestBB) {
  // If BB has single-entry PHI nodes, fold them.
  while (PHINode *PN = dyn_cast<PHINode>(DestBB->begin())) {
    Value *NewVal = PN->getIncomingValue(0);
    // Replace self referencing PHI with undef, it must be dead.
    if (NewVal == PN) NewVal = UndefValue::get(PN->getType());
    PN->replaceAllUsesWith(NewVal);
    PN->eraseFromParent();
  }
  
  BasicBlock *PredBB = DestBB->getSinglePredecessor();
  assert(PredBB && "Block doesn't have a single predecessor!");
  
  // Splice all the instructions from PredBB to DestBB.
  PredBB->getTerminator()->eraseFromParent();
  DestBB->getInstList().splice(DestBB->begin(), PredBB->getInstList());
  
  // Anything that branched to PredBB now branches to DestBB.
  PredBB->replaceAllUsesWith(DestBB);
  
  // Nuke BB.
  PredBB->eraseFromParent();
}

//.........这里部分代码省略.........
bool CodeGenPrepare::DupRetToEnableTailCallOpts(ReturnInst *RI) {
    if (!TLI)
        return false;

    Value *V = RI->getReturnValue();
    PHINode *PN = V ? dyn_cast<PHINode>(V) : NULL;
    if (V && !PN)
        return false;

    BasicBlock *BB = RI->getParent();
    if (PN && PN->getParent() != BB)
        return false;

    // It's not safe to eliminate the sign / zero extension of the return value.
    // See llvm::isInTailCallPosition().
    const Function *F = BB->getParent();
    Attributes CallerRetAttr = F->getAttributes().getRetAttributes();
    if ((CallerRetAttr & Attribute::ZExt) || (CallerRetAttr & Attribute::SExt))
        return false;

    // Make sure there are no instructions between the PHI and return, or that the
    // return is the first instruction in the block.
    if (PN) {
        BasicBlock::iterator BI = BB->begin();
        do {
            ++BI;
        }
        while (isa<DbgInfoIntrinsic>(BI));
        if (&*BI != RI)
            return false;
    } else {
        BasicBlock::iterator BI = BB->begin();
        while (isa<DbgInfoIntrinsic>(BI)) ++BI;
        if (&*BI != RI)
            return false;
    }

    /// Only dup the ReturnInst if the CallInst is likely to be emitted as a tail
    /// call.
    SmallVector<CallInst*, 4> TailCalls;
    if (PN) {
        for (unsigned I = 0, E = PN->getNumIncomingValues(); I != E; ++I) {
            CallInst *CI = dyn_cast<CallInst>(PN->getIncomingValue(I));
            // Make sure the phi value is indeed produced by the tail call.
            if (CI && CI->hasOneUse() && CI->getParent() == PN->getIncomingBlock(I) &&
                    TLI->mayBeEmittedAsTailCall(CI))
                TailCalls.push_back(CI);
        }
    } else {
        SmallPtrSet<BasicBlock*, 4> VisitedBBs;
        for (pred_iterator PI = pred_begin(BB), PE = pred_end(BB); PI != PE; ++PI) {
            if (!VisitedBBs.insert(*PI))
                continue;

            BasicBlock::InstListType &InstList = (*PI)->getInstList();
            BasicBlock::InstListType::reverse_iterator RI = InstList.rbegin();
            BasicBlock::InstListType::reverse_iterator RE = InstList.rend();
            do {
                ++RI;
            }
            while (RI != RE && isa<DbgInfoIntrinsic>(&*RI));
            if (RI == RE)
                continue;

            CallInst *CI = dyn_cast<CallInst>(&*RI);
            if (CI && CI->use_empty() && TLI->mayBeEmittedAsTailCall(CI))
                TailCalls.push_back(CI);
        }
    }

    bool Changed = false;
    for (unsigned i = 0, e = TailCalls.size(); i != e; ++i) {
        CallInst *CI = TailCalls[i];
        CallSite CS(CI);

        // Conservatively require the attributes of the call to match those of the
        // return. Ignore noalias because it doesn't affect the call sequence.
        Attributes CalleeRetAttr = CS.getAttributes().getRetAttributes();
        if ((CalleeRetAttr ^ CallerRetAttr) & ~Attribute::NoAlias)
            continue;

        // Make sure the call instruction is followed by an unconditional branch to
        // the return block.
        BasicBlock *CallBB = CI->getParent();
        BranchInst *BI = dyn_cast<BranchInst>(CallBB->getTerminator());
        if (!BI || !BI->isUnconditional() || BI->getSuccessor(0) != BB)
            continue;

        // Duplicate the return into CallBB.
        (void)FoldReturnIntoUncondBranch(RI, BB, CallBB);
        ModifiedDT = Changed = true;
        ++NumRetsDup;
    }

    // If we eliminated all predecessors of the block, delete the block now.
    if (Changed && pred_begin(BB) == pred_end(BB))
        BB->eraseFromParent();

    return Changed;
}

/// \brief Simplify one loop and queue further loops for simplification.
///
/// FIXME: Currently this accepts both lots of analyses that it uses and a raw
/// Pass pointer. The Pass pointer is used by numerous utilities to update
/// specific analyses. Rather than a pass it would be much cleaner and more
/// explicit if they accepted the analysis directly and then updated it.
static bool simplifyOneLoop(Loop *L, SmallVectorImpl<Loop *> &Worklist,
                            DominatorTree *DT, LoopInfo *LI,
                            ScalarEvolution *SE, Pass *PP,
                            AssumptionCache *AC) {
  bool Changed = false;
ReprocessLoop:

  // Check to see that no blocks (other than the header) in this loop have
  // predecessors that are not in the loop.  This is not valid for natural
  // loops, but can occur if the blocks are unreachable.  Since they are
  // unreachable we can just shamelessly delete those CFG edges!
  for (Loop::block_iterator BB = L->block_begin(), E = L->block_end();
       BB != E; ++BB) {
    if (*BB == L->getHeader()) continue;

    SmallPtrSet<BasicBlock*, 4> BadPreds;
    for (pred_iterator PI = pred_begin(*BB),
         PE = pred_end(*BB); PI != PE; ++PI) {
      BasicBlock *P = *PI;
      if (!L->contains(P))
        BadPreds.insert(P);
    }

    // Delete each unique out-of-loop (and thus dead) predecessor.
    for (BasicBlock *P : BadPreds) {

      DEBUG(dbgs() << "LoopSimplify: Deleting edge from dead predecessor "
                   << P->getName() << "\n");

      // Inform each successor of each dead pred.
      for (succ_iterator SI = succ_begin(P), SE = succ_end(P); SI != SE; ++SI)
        (*SI)->removePredecessor(P);
      // Zap the dead pred's terminator and replace it with unreachable.
      TerminatorInst *TI = P->getTerminator();
       TI->replaceAllUsesWith(UndefValue::get(TI->getType()));
      P->getTerminator()->eraseFromParent();
      new UnreachableInst(P->getContext(), P);
      Changed = true;
    }
  }

  // If there are exiting blocks with branches on undef, resolve the undef in
  // the direction which will exit the loop. This will help simplify loop
  // trip count computations.
  SmallVector<BasicBlock*, 8> ExitingBlocks;
  L->getExitingBlocks(ExitingBlocks);
  for (SmallVectorImpl<BasicBlock *>::iterator I = ExitingBlocks.begin(),
       E = ExitingBlocks.end(); I != E; ++I)
    if (BranchInst *BI = dyn_cast<BranchInst>((*I)->getTerminator()))
      if (BI->isConditional()) {
        if (UndefValue *Cond = dyn_cast<UndefValue>(BI->getCondition())) {

          DEBUG(dbgs() << "LoopSimplify: Resolving \"br i1 undef\" to exit in "
                       << (*I)->getName() << "\n");

          BI->setCondition(ConstantInt::get(Cond->getType(),
                                            !L->contains(BI->getSuccessor(0))));

          // This may make the loop analyzable, force SCEV recomputation.
          if (SE)
            SE->forgetLoop(L);

          Changed = true;
        }
      }

  // Does the loop already have a preheader?  If so, don't insert one.
  BasicBlock *Preheader = L->getLoopPreheader();
  if (!Preheader) {
    Preheader = InsertPreheaderForLoop(L, PP);
    if (Preheader) {
      ++NumInserted;
      Changed = true;
    }
  }

  // Next, check to make sure that all exit nodes of the loop only have
  // predecessors that are inside of the loop.  This check guarantees that the
  // loop preheader/header will dominate the exit blocks.  If the exit block has
  // predecessors from outside of the loop, split the edge now.
  SmallVector<BasicBlock*, 8> ExitBlocks;
  L->getExitBlocks(ExitBlocks);

  SmallSetVector<BasicBlock *, 8> ExitBlockSet(ExitBlocks.begin(),
                                               ExitBlocks.end());
  for (SmallSetVector<BasicBlock *, 8>::iterator I = ExitBlockSet.begin(),
         E = ExitBlockSet.end(); I != E; ++I) {
    BasicBlock *ExitBlock = *I;
    for (pred_iterator PI = pred_begin(ExitBlock), PE = pred_end(ExitBlock);
         PI != PE; ++PI)
      // Must be exactly this loop: no subloops, parent loops, or non-loop preds
      // allowed.
      if (!L->contains(*PI)) {
        if (rewriteLoopExitBlock(L, ExitBlock, DT, LI, PP)) {
//.........这里部分代码省略.........

//.........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();
    
    // Make all PHI nodes that referred to Dest now refer to I as their source.
    Dest->replaceAllUsesWith(I);

    // Move all the instructions in the succ to the pred.
    I->getInstList().splice(I->end(), Dest->getInstList());
    
    // Remove the dest block.
    Dest->eraseFromParent();
    
    // Do not increment I, iteratively merge all things this block branches to.
  }
}

/// removeBlocks - Remove basic block DeadBB and all blocks dominated by DeadBB.
/// This routine is used to remove split condition's dead branch, dominated by
/// DeadBB. LiveBB dominates split conidition's other branch.
void LoopIndexSplit::removeBlocks(BasicBlock *DeadBB, Loop *LP, 
                                  BasicBlock *LiveBB) {

  // First update DeadBB's dominance frontier. 
  SmallVector<BasicBlock *, 8> FrontierBBs;
  DominanceFrontier::iterator DeadBBDF = DF->find(DeadBB);
  if (DeadBBDF != DF->end()) {
    SmallVector<BasicBlock *, 8> PredBlocks;
    
    DominanceFrontier::DomSetType DeadBBSet = DeadBBDF->second;
    for (DominanceFrontier::DomSetType::iterator DeadBBSetI = DeadBBSet.begin(),
           DeadBBSetE = DeadBBSet.end(); DeadBBSetI != DeadBBSetE; ++DeadBBSetI) 
      {
      BasicBlock *FrontierBB = *DeadBBSetI;
      FrontierBBs.push_back(FrontierBB);

      // Rremove any PHI incoming edge from blocks dominated by DeadBB.
      PredBlocks.clear();
      for(pred_iterator PI = pred_begin(FrontierBB), PE = pred_end(FrontierBB);
          PI != PE; ++PI) {
        BasicBlock *P = *PI;
        if (P == DeadBB || DT->dominates(DeadBB, P))
          PredBlocks.push_back(P);
      }

      for(BasicBlock::iterator FBI = FrontierBB->begin(), FBE = FrontierBB->end();
          FBI != FBE; ++FBI) {
        if (PHINode *PN = dyn_cast<PHINode>(FBI)) {
          for(SmallVector<BasicBlock *, 8>::iterator PI = PredBlocks.begin(),
                PE = PredBlocks.end(); PI != PE; ++PI) {
            BasicBlock *P = *PI;
            PN->removeIncomingValue(P);
          }
        }
        else
          break;
      }      
    }
  }
  
  // Now remove DeadBB and all nodes dominated by DeadBB in df order.
  SmallVector<BasicBlock *, 32> WorkList;
  DomTreeNode *DN = DT->getNode(DeadBB);
  for (df_iterator<DomTreeNode*> DI = df_begin(DN),
         E = df_end(DN); DI != E; ++DI) {
    BasicBlock *BB = DI->getBlock();
    WorkList.push_back(BB);
    BB->replaceAllUsesWith(UndefValue::get(
                                       Type::getLabelTy(DeadBB->getContext())));
  }

  while (!WorkList.empty()) {
    BasicBlock *BB = WorkList.back(); WorkList.pop_back();
    LPM->deleteSimpleAnalysisValue(BB, LP);
    for(BasicBlock::iterator BBI = BB->begin(), BBE = BB->end(); 
        BBI != BBE; ) {
      Instruction *I = BBI;
      ++BBI;
      I->replaceAllUsesWith(UndefValue::get(I->getType()));
      LPM->deleteSimpleAnalysisValue(I, LP);
      I->eraseFromParent();
    }
    DT->eraseNode(BB);
    DF->removeBlock(BB);
    LI->removeBlock(BB);
    BB->eraseFromParent();
  }

  // Update Frontier BBs' dominator info.
  while (!FrontierBBs.empty()) {
    BasicBlock *FBB = FrontierBBs.back(); FrontierBBs.pop_back();
    BasicBlock *NewDominator = FBB->getSinglePredecessor();
    if (!NewDominator) {
      pred_iterator PI = pred_begin(FBB), PE = pred_end(FBB);
      NewDominator = *PI;
      ++PI;
      if (NewDominator != LiveBB) {
        for(; PI != PE; ++PI) {
          BasicBlock *P = *PI;
          if (P == LiveBB) {
            NewDominator = LiveBB;
            break;
          }
          NewDominator = DT->findNearestCommonDominator(NewDominator, P);
        }
      }
    }
    assert (NewDominator && "Unable to fix dominator info.");
    DT->changeImmediateDominator(FBB, NewDominator);
    DF->changeImmediateDominator(FBB, NewDominator, DT);
  }

}

/// SimplifyCode - Okay, now that we have simplified some instructions in the
/// loop, walk over it and constant prop, dce, and fold control flow where
/// possible.  Note that this is effectively a very simple loop-structure-aware
/// optimizer.  During processing of this loop, L could very well be deleted, so
/// it must not be used.
///
/// FIXME: When the loop optimizer is more mature, separate this out to a new
/// pass.
///
void LoopUnswitch::SimplifyCode(std::vector<Instruction*> &Worklist, Loop *L) {
  while (!Worklist.empty()) {
    Instruction *I = Worklist.back();
    Worklist.pop_back();

    // Simple DCE.
    if (isInstructionTriviallyDead(I)) {
      DEBUG(dbgs() << "Remove dead instruction '" << *I);
      
      // Add uses to the worklist, which may be dead now.
      for (unsigned i = 0, e = I->getNumOperands(); i != e; ++i)
        if (Instruction *Use = dyn_cast<Instruction>(I->getOperand(i)))
          Worklist.push_back(Use);
      LPM->deleteSimpleAnalysisValue(I, L);
      RemoveFromWorklist(I, Worklist);
      I->eraseFromParent();
      ++NumSimplify;
      continue;
    }

    // See if instruction simplification can hack this up.  This is common for
    // things like "select false, X, Y" after unswitching made the condition be
    // 'false'.
    if (Value *V = SimplifyInstruction(I, 0, DT))
      if (LI->replacementPreservesLCSSAForm(I, V)) {
        ReplaceUsesOfWith(I, V, Worklist, L, LPM);
        continue;
      }

    // Special case hacks that appear commonly in unswitched code.
    if (BranchInst *BI = dyn_cast<BranchInst>(I)) {
      if (BI->isUnconditional()) {
        // If BI's parent is the only pred of the successor, fold the two blocks
        // together.
        BasicBlock *Pred = BI->getParent();
        BasicBlock *Succ = BI->getSuccessor(0);
        BasicBlock *SinglePred = Succ->getSinglePredecessor();
        if (!SinglePred) continue;  // Nothing to do.
        assert(SinglePred == Pred && "CFG broken");

        DEBUG(dbgs() << "Merging blocks: " << Pred->getName() << " <- " 
              << Succ->getName() << "\n");
        
        // Resolve any single entry PHI nodes in Succ.
        while (PHINode *PN = dyn_cast<PHINode>(Succ->begin()))
          ReplaceUsesOfWith(PN, PN->getIncomingValue(0), Worklist, L, LPM);
        
        // Move all of the successor contents from Succ to Pred.
        Pred->getInstList().splice(BI, Succ->getInstList(), Succ->begin(),
                                   Succ->end());
        LPM->deleteSimpleAnalysisValue(BI, L);
        BI->eraseFromParent();
        RemoveFromWorklist(BI, Worklist);
        
        // If Succ has any successors with PHI nodes, update them to have
        // entries coming from Pred instead of Succ.
        Succ->replaceAllUsesWith(Pred);
        
        // Remove Succ from the loop tree.
        LI->removeBlock(Succ);
        LPM->deleteSimpleAnalysisValue(Succ, L);
        Succ->eraseFromParent();
        ++NumSimplify;
        continue;
      }
      
      if (ConstantInt *CB = dyn_cast<ConstantInt>(BI->getCondition())){
        // Conditional branch.  Turn it into an unconditional branch, then
        // remove dead blocks.
        continue;  // FIXME: Enable.

        DEBUG(dbgs() << "Folded branch: " << *BI);
        BasicBlock *DeadSucc = BI->getSuccessor(CB->getZExtValue());
        BasicBlock *LiveSucc = BI->getSuccessor(!CB->getZExtValue());
        DeadSucc->removePredecessor(BI->getParent(), true);
        Worklist.push_back(BranchInst::Create(LiveSucc, BI));
        LPM->deleteSimpleAnalysisValue(BI, L);
        BI->eraseFromParent();
        RemoveFromWorklist(BI, Worklist);
        ++NumSimplify;

        RemoveBlockIfDead(DeadSucc, Worklist, L);
      }
      continue;
    }
  }
}

//.........这里部分代码省略.........
      }

      ValueMap[I] = ActualArg;
    }

    // We want the inliner to prune the code as it copies.  We would LOVE to
    // have no dead or constant instructions leftover after inlining occurs
    // (which can happen, e.g., because an argument was constant), but we'll be
    // happy with whatever the cloner can do.
    CloneAndPruneFunctionInto(Caller, CalledFunc, ValueMap, Returns, ".i",
                              &InlinedFunctionInfo, TD);

    // Remember the first block that is newly cloned over.
    FirstNewBlock = LastBlock; ++FirstNewBlock;

    // Update the callgraph if requested.
    if (CG)
      UpdateCallGraphAfterInlining(CS, FirstNewBlock, ValueMap, *CG);
  }

  // If there are any alloca instructions in the block that used to be the entry
  // block for the callee, move them to the entry block of the caller.  First
  // calculate which instruction they should be inserted before.  We insert the
  // instructions at the end of the current alloca list.
  //
  {
    BasicBlock::iterator InsertPoint = Caller->begin()->begin();
    for (BasicBlock::iterator I = FirstNewBlock->begin(),
           E = FirstNewBlock->end(); I != E; )
      if (AllocaInst *AI = dyn_cast<AllocaInst>(I++)) {
        // If the alloca is now dead, remove it.  This often occurs due to code
        // specialization.
        if (AI->use_empty()) {
          AI->eraseFromParent();
          continue;
        }

        if (isa<Constant>(AI->getArraySize())) {
          // Scan for the block of allocas that we can move over, and move them
          // all at once.
          while (isa<AllocaInst>(I) &&
                 isa<Constant>(cast<AllocaInst>(I)->getArraySize()))
            ++I;

          // Transfer all of the allocas over in a block.  Using splice means
          // that the instructions aren't removed from the symbol table, then
          // reinserted.
          Caller->getEntryBlock().getInstList().splice(
              InsertPoint,
              FirstNewBlock->getInstList(),
              AI, I);
        }
      }
  }

  // If the inlined code contained dynamic alloca instructions, wrap the inlined
  // code with llvm.stacksave/llvm.stackrestore intrinsics.
  if (InlinedFunctionInfo.ContainsDynamicAllocas) {
    Module *M = Caller->getParent();
    // Get the two intrinsics we care about.
    Constant *StackSave, *StackRestore;
    StackSave    = Intrinsic::getDeclaration(M, Intrinsic::stacksave);
    StackRestore = Intrinsic::getDeclaration(M, Intrinsic::stackrestore);

    // If we are preserving the callgraph, add edges to the stacksave/restore
    // functions for the calls we insert.

//.........这里部分代码省略.........
  bool IsMustTailCall = CS.isMustTailCall();

  PHINode *CallPN = nullptr;

  // `musttail` calls must be followed by optional `bitcast`, and `ret`. The
  // split blocks will be terminated right after that so there're no users for
  // this phi in a `TailBB`.
  if (!IsMustTailCall && Instr->getNumUses())
    CallPN = PHINode::Create(Instr->getType(), Preds.size(), "phi.call");

  DEBUG(dbgs() << "split call-site : " << *Instr << " into \n");

  assert(Preds.size() == 2 && "The ValueToValueMaps array has size 2.");
  // ValueToValueMapTy is neither copy nor moveable, so we use a simple array
  // here.
  ValueToValueMapTy ValueToValueMaps[2];
  for (unsigned i = 0; i < Preds.size(); i++) {
    BasicBlock *PredBB = Preds[i].first;
    BasicBlock *SplitBlock = DuplicateInstructionsInSplitBetween(
        TailBB, PredBB, &*std::next(Instr->getIterator()), ValueToValueMaps[i]);
    assert(SplitBlock && "Unexpected new basic block split.");

    Instruction *NewCI =
        &*std::prev(SplitBlock->getTerminator()->getIterator());
    CallSite NewCS(NewCI);
    addConditions(NewCS, Preds[i].second);

    // Handle PHIs used as arguments in the call-site.
    for (PHINode &PN : TailBB->phis()) {
      unsigned ArgNo = 0;
      for (auto &CI : CS.args()) {
        if (&*CI == &PN) {
          NewCS.setArgument(ArgNo, PN.getIncomingValueForBlock(SplitBlock));
        }
        ++ArgNo;
      }
    }
    DEBUG(dbgs() << "    " << *NewCI << " in " << SplitBlock->getName()
                 << "\n");
    if (CallPN)
      CallPN->addIncoming(NewCI, SplitBlock);

    // Clone and place bitcast and return instructions before `TI`
    if (IsMustTailCall)
      copyMustTailReturn(SplitBlock, Instr, NewCI);
  }

  NumCallSiteSplit++;

  // FIXME: remove TI in `copyMustTailReturn`
  if (IsMustTailCall) {
    // Remove superfluous `br` terminators from the end of the Split blocks
    // NOTE: Removing terminator removes the SplitBlock from the TailBB's
    // predecessors. Therefore we must get complete list of Splits before
    // attempting removal.
    SmallVector<BasicBlock *, 2> Splits(predecessors((TailBB)));
    assert(Splits.size() == 2 && "Expected exactly 2 splits!");
    for (unsigned i = 0; i < Splits.size(); i++)
      Splits[i]->getTerminator()->eraseFromParent();

    // Erase the tail block once done with musttail patching
    TailBB->eraseFromParent();
    return;
  }

  auto *OriginalBegin = &*TailBB->begin();
  // Replace users of the original call with a PHI mering call-sites split.
  if (CallPN) {
    CallPN->insertBefore(OriginalBegin);
    Instr->replaceAllUsesWith(CallPN);
  }

  // Remove instructions moved to split blocks from TailBB, from the duplicated
  // call instruction to the beginning of the basic block. If an instruction
  // has any uses, add a new PHI node to combine the values coming from the
  // split blocks. The new PHI nodes are placed before the first original
  // instruction, so we do not end up deleting them. By using reverse-order, we
  // do not introduce unnecessary PHI nodes for def-use chains from the call
  // instruction to the beginning of the block.
  auto I = Instr->getReverseIterator();
  while (I != TailBB->rend()) {
    Instruction *CurrentI = &*I++;
    if (!CurrentI->use_empty()) {
      // If an existing PHI has users after the call, there is no need to create
      // a new one.
      if (isa<PHINode>(CurrentI))
        continue;
      PHINode *NewPN = PHINode::Create(CurrentI->getType(), Preds.size());
      for (auto &Mapping : ValueToValueMaps)
        NewPN->addIncoming(Mapping[CurrentI],
                           cast<Instruction>(Mapping[CurrentI])->getParent());
      NewPN->insertBefore(&*TailBB->begin());
      CurrentI->replaceAllUsesWith(NewPN);
    }
    CurrentI->eraseFromParent();
    // We are done once we handled the first original instruction in TailBB.
    if (CurrentI == OriginalBegin)
      break;
  }
}

//.........这里部分代码省略.........
           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;
      }
    }
  }

  // Make a second pass over the PHINodes now that all of them have been
  // remapped into the new function, simplifying the PHINode and performing any
  // recursive simplifications exposed. This will transparently update the
  // WeakVH in the VMap. Notably, we rely on that so that if we coalesce
  // two PHINodes, the iteration over the old PHIs remains valid, and the
  // mapping will just map us to the new node (which may not even be a PHI
  // node).
  for (unsigned Idx = 0, Size = PHIToResolve.size(); Idx != Size; ++Idx)
    if (PHINode *PN = dyn_cast<PHINode>(VMap[PHIToResolve[Idx]]))
      recursivelySimplifyInstruction(PN);

  // Now that the inlined function body has been fully constructed, go through
  // and zap unconditional fall-through branches. This happens all the time when
  // specializing code: code specialization turns conditional branches into
  // uncond branches, and this code folds them.
  Function::iterator Begin = cast<BasicBlock>(VMap[StartingBB])->getIterator();
  Function::iterator I = Begin;
  while (I != NewFunc->end()) {
    // Check if this block has become dead during inlining or other
    // simplifications. Note that the first block will appear dead, as it has
    // not yet been wired up properly.
    if (I != Begin && (pred_begin(&*I) == pred_end(&*I) ||
                       I->getSinglePredecessor() == &*I)) {
      BasicBlock *DeadBB = &*I++;
      DeleteDeadBlock(DeadBB);
      continue;
    }

/// Check whether \param BB is the merge block of a if-region.  If yes, check
/// whether there exists an adjacent if-region upstream, the two if-regions
/// contain identical instuctions and can be legally merged.  \returns true if
/// the two if-regions are merged.
///
/// From:
/// if (a)
///   statement;
/// if (b)
///   statement;
///
/// To:
/// if (a || b)
///   statement;
///
bool FlattenCFGOpt::MergeIfRegion(BasicBlock *BB, IRBuilder<> &Builder,
                                  Pass *P) {
  BasicBlock *IfTrue2, *IfFalse2;
  Value *IfCond2 = GetIfCondition(BB, IfTrue2, IfFalse2);
  Instruction *CInst2 = dyn_cast_or_null<Instruction>(IfCond2);
  if (!CInst2)
    return false;

  BasicBlock *SecondEntryBlock = CInst2->getParent();
  if (SecondEntryBlock->hasAddressTaken())
    return false;

  BasicBlock *IfTrue1, *IfFalse1;
  Value *IfCond1 = GetIfCondition(SecondEntryBlock, IfTrue1, IfFalse1);
  Instruction *CInst1 = dyn_cast_or_null<Instruction>(IfCond1);
  if (!CInst1)
    return false;

  BasicBlock *FirstEntryBlock = CInst1->getParent();

  // Either then-path or else-path should be empty.
  if ((IfTrue1 != FirstEntryBlock) && (IfFalse1 != FirstEntryBlock))
    return false;
  if ((IfTrue2 != SecondEntryBlock) && (IfFalse2 != SecondEntryBlock))
    return false;

  TerminatorInst *PTI2 = SecondEntryBlock->getTerminator();
  Instruction *PBI2 = SecondEntryBlock->begin();

  if (!CompareIfRegionBlock(FirstEntryBlock, SecondEntryBlock, IfTrue1,
                            IfTrue2))
    return false;

  if (!CompareIfRegionBlock(FirstEntryBlock, SecondEntryBlock, IfFalse1,
                            IfFalse2))
    return false;

  // Check whether \param SecondEntryBlock has side-effect and is safe to
  // speculate.
  for (BasicBlock::iterator BI = PBI2, BE = PTI2; BI != BE; ++BI) {
    Instruction *CI = BI;
    if (isa<PHINode>(CI) || CI->mayHaveSideEffects() ||
        !isSafeToSpeculativelyExecute(CI))
      return false;
  }

  // Merge \param SecondEntryBlock into \param FirstEntryBlock.
  FirstEntryBlock->getInstList().pop_back();
  FirstEntryBlock->getInstList()
      .splice(FirstEntryBlock->end(), SecondEntryBlock->getInstList());
  BranchInst *PBI = dyn_cast<BranchInst>(FirstEntryBlock->getTerminator());
  Value *CC = PBI->getCondition();
  BasicBlock *SaveInsertBB = Builder.GetInsertBlock();
  BasicBlock::iterator SaveInsertPt = Builder.GetInsertPoint();
  Builder.SetInsertPoint(PBI);
  Value *NC = Builder.CreateOr(CInst1, CC);
  PBI->replaceUsesOfWith(CC, NC);
  Builder.SetInsertPoint(SaveInsertBB, SaveInsertPt);

  // Remove IfTrue1
  if (IfTrue1 != FirstEntryBlock) {
    IfTrue1->dropAllReferences();
    IfTrue1->eraseFromParent();
  }

  // Remove IfFalse1
  if (IfFalse1 != FirstEntryBlock) {
    IfFalse1->dropAllReferences();
    IfFalse1->eraseFromParent();
  }

  // Remove \param SecondEntryBlock
  SecondEntryBlock->dropAllReferences();
  SecondEntryBlock->eraseFromParent();
  DEBUG(dbgs() << "If conditions merged into:\n" << *FirstEntryBlock);
  return true;
}

//.........这里部分代码省略.........
    // We want the inliner to prune the code as it copies.  We would LOVE to
    // have no dead or constant instructions leftover after inlining occurs
    // (which can happen, e.g., because an argument was constant), but we'll be
    // happy with whatever the cloner can do.
    CloneAndPruneFunctionInto(Caller, CalledFunc, VMap, 
                              /*ModuleLevelChanges=*/false, Returns, ".i",
                              &InlinedFunctionInfo, IFI.TD, TheCall);

    // Remember the first block that is newly cloned over.
    FirstNewBlock = LastBlock; ++FirstNewBlock;

    // Update the callgraph if requested.
    if (IFI.CG)
      UpdateCallGraphAfterInlining(CS, FirstNewBlock, VMap, IFI);

    // Update inlined instructions' line number information.
    fixupLineNumbers(Caller, FirstNewBlock, TheCall);
  }

  // If there are any alloca instructions in the block that used to be the entry
  // block for the callee, move them to the entry block of the caller.  First
  // calculate which instruction they should be inserted before.  We insert the
  // instructions at the end of the current alloca list.
  {
    BasicBlock::iterator InsertPoint = Caller->begin()->begin();
    for (BasicBlock::iterator I = FirstNewBlock->begin(),
         E = FirstNewBlock->end(); I != E; ) {
      AllocaInst *AI = dyn_cast<AllocaInst>(I++);
      if (AI == 0) continue;
      
      // If the alloca is now dead, remove it.  This often occurs due to code
      // specialization.
      if (AI->use_empty()) {
        AI->eraseFromParent();
        continue;
      }

      if (!isa<Constant>(AI->getArraySize()))
        continue;
      
      // Keep track of the static allocas that we inline into the caller.
      IFI.StaticAllocas.push_back(AI);
      
      // Scan for the block of allocas that we can move over, and move them
      // all at once.
      while (isa<AllocaInst>(I) &&
             isa<Constant>(cast<AllocaInst>(I)->getArraySize())) {
        IFI.StaticAllocas.push_back(cast<AllocaInst>(I));
        ++I;
      }

      // Transfer all of the allocas over in a block.  Using splice means
      // that the instructions aren't removed from the symbol table, then
      // reinserted.
      Caller->getEntryBlock().getInstList().splice(InsertPoint,
                                                   FirstNewBlock->getInstList(),
                                                   AI, I);
    }
  }

  // Leave lifetime markers for the static alloca's, scoping them to the
  // function we just inlined.
  if (InsertLifetime && !IFI.StaticAllocas.empty()) {
    IRBuilder<> builder(FirstNewBlock->begin());
    for (unsigned ai = 0, ae = IFI.StaticAllocas.size(); ai != ae; ++ai) {
      AllocaInst *AI = IFI.StaticAllocas[ai];

// InlineFunction - This function inlines the called function into the basic
// block of the caller.  This returns false if it is not possible to inline this
// call.  The program is still in a well defined state if this occurs though.
//
// Note that this only does one level of inlining.  For example, if the
// instruction 'call B' is inlined, and 'B' calls 'C', then the call to 'C' now
// exists in the instruction stream.  Similiarly this will inline a recursive
// function by one level.
//
bool llvm::InlineFunction(CallSite CS, CallGraph *CG, const TargetData *TD) {
  Instruction *TheCall = CS.getInstruction();
  assert(TheCall->getParent() && TheCall->getParent()->getParent() &&
         "Instruction not in function!");

  const Function *CalledFunc = CS.getCalledFunction();
  if (CalledFunc == 0 ||          // Can't inline external function or indirect
      CalledFunc->isDeclaration() || // call, or call to a vararg function!
      CalledFunc->getFunctionType()->isVarArg()) return false;


  // If the call to the callee is a non-tail call, we must clear the 'tail'
  // flags on any calls that we inline.
  bool MustClearTailCallFlags =
    isa<CallInst>(TheCall) && !cast<CallInst>(TheCall)->isTailCall();

  BasicBlock *OrigBB = TheCall->getParent();
  Function *Caller = OrigBB->getParent();

  // Get an iterator to the last basic block in the function, which will have
  // the new function inlined after it.
  //
  Function::iterator LastBlock = &Caller->back();

  // Make sure to capture all of the return instructions from the cloned
  // function.
  std::vector<ReturnInst*> Returns;
  ClonedCodeInfo InlinedFunctionInfo;
  Function::iterator FirstNewBlock;
  
  { // Scope to destroy ValueMap after cloning.
    DenseMap<const Value*, Value*> ValueMap;

    // Calculate the vector of arguments to pass into the function cloner, which
    // matches up the formal to the actual argument values.
    assert(std::distance(CalledFunc->arg_begin(), CalledFunc->arg_end()) ==
           std::distance(CS.arg_begin(), CS.arg_end()) &&
           "No varargs calls can be inlined!");
    CallSite::arg_iterator AI = CS.arg_begin();
    for (Function::const_arg_iterator I = CalledFunc->arg_begin(),
           E = CalledFunc->arg_end(); I != E; ++I, ++AI)
      ValueMap[I] = *AI;

    // We want the inliner to prune the code as it copies.  We would LOVE to
    // have no dead or constant instructions leftover after inlining occurs
    // (which can happen, e.g., because an argument was constant), but we'll be
    // happy with whatever the cloner can do.
    CloneAndPruneFunctionInto(Caller, CalledFunc, ValueMap, Returns, ".i",
                              &InlinedFunctionInfo, TD);
    
    // Remember the first block that is newly cloned over.
    FirstNewBlock = LastBlock; ++FirstNewBlock;
    
    // Update the callgraph if requested.
    if (CG)
      UpdateCallGraphAfterInlining(Caller, CalledFunc, FirstNewBlock, ValueMap,
                                   *CG);
  }
 
  // If there are any alloca instructions in the block that used to be the entry
  // block for the callee, move them to the entry block of the caller.  First
  // calculate which instruction they should be inserted before.  We insert the
  // instructions at the end of the current alloca list.
  //
  {
    BasicBlock::iterator InsertPoint = Caller->begin()->begin();
    for (BasicBlock::iterator I = FirstNewBlock->begin(),
           E = FirstNewBlock->end(); I != E; )
      if (AllocaInst *AI = dyn_cast<AllocaInst>(I++)) {
        // If the alloca is now dead, remove it.  This often occurs due to code
        // specialization.
        if (AI->use_empty()) {
          AI->eraseFromParent();
          continue;
        }
        
        if (isa<Constant>(AI->getArraySize())) {
          // Scan for the block of allocas that we can move over, and move them
          // all at once.
          while (isa<AllocaInst>(I) &&
                 isa<Constant>(cast<AllocaInst>(I)->getArraySize()))
            ++I;

          // Transfer all of the allocas over in a block.  Using splice means
          // that the instructions aren't removed from the symbol table, then
          // reinserted.
          Caller->getEntryBlock().getInstList().splice(
              InsertPoint,
              FirstNewBlock->getInstList(),
              AI, I);
        }
//.........这里部分代码省略.........

void DSWP::cleanup(Loop *L, LPPassManager &LPM) {
  // Move some instructions that may not have been inserted in the right
  // place, delete the old loop, and clean up our aux data structures for this
  // loop.

  /*
   * move the produce instructions, which have been inserted after the branch,
   * in front of it
   */
  for (int i = 0; i < MAX_THREAD; i++) {
      for (Function::iterator bi = allFunc[i]->begin(), be = allFunc[i]->end(); bi != be; ++bi) {
          BasicBlock *bb = bi;
          TerminatorInst *term = NULL;
          for (BasicBlock::iterator ii = bb->begin(), ie = bb->end(); ii != ie; ++ii) {
              Instruction *inst = ii;
              if (isa<TerminatorInst>(inst)) {
                  term = dyn_cast<TerminatorInst>(inst);
                  break;
              }
          }

          if (term == NULL) {
              error("term cannot be null");
          }

          while (true) {
              Instruction *last = &bb->getInstList().back();
              if (isa<TerminatorInst>(last))
                  break;
              last->moveBefore(term);
          }
      }
  }

  /*
   * move the phi nodes to the top of the block
   */
  for (int i = 0; i < MAX_THREAD; i++) {
      for (Function::iterator bi = allFunc[i]->begin(), be = allFunc[i]->end(); bi != be; ++bi) {
          BasicBlock *bb = bi;
          Instruction *first_nonphi = bb->getFirstNonPHI();
          BasicBlock::iterator ii = bb->begin(), ie = bb->end();

          // advance the iterator up to one past first_nonphi
          while (&(*ii) != first_nonphi) {
              ++ii;
          }
          ++ii;

          // move any phi nodes after the first nonphi to before it
          for (BasicBlock::iterator i_next; ii != ie; ii = i_next) {
              i_next = ii;
              ++i_next;
              Instruction *inst = ii;
              if (isa<PHINode>(inst)) {
                  inst->moveBefore(first_nonphi);
              }
          }
      }
  }

  cout << "begin to delete loop" << endl;
  for (Loop::block_iterator bi = L->block_begin(), be = L->block_end(); bi != be; ++bi) {
      BasicBlock *BB = *bi;
      for (BasicBlock::iterator ii = BB->begin(), i_next, ie = BB->end(); ii != ie; ii = i_next) {
          i_next = ii;
          ++i_next;
          Instruction &inst = *ii;
          inst.replaceAllUsesWith(UndefValue::get(inst.getType()));
          inst.eraseFromParent();
      }
  }

  // Delete the basic blocks only afterwards
  // so that backwards branch instructions don't break
  for (Loop::block_iterator bi = L->block_begin(), be = L->block_end(); bi != be; ++bi) {
      BasicBlock *BB = *bi;
      BB->eraseFromParent();
  }

  LPM.deleteLoopFromQueue(L);

}