C++ PHINode::getType方法代码示例

//.........这里部分代码省略.........
    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 (const auto &PCI : PredCount) {
          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
  // WeakTrackingVH 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).
  const DataLayout &DL = NewFunc->getParent()->getDataLayout();
  SmallSetVector<const Value *, 8> Worklist;
  for (unsigned Idx = 0, Size = PHIToResolve.size(); Idx != Size; ++Idx)
    if (isa<PHINode>(VMap[PHIToResolve[Idx]]))
      Worklist.insert(PHIToResolve[Idx]);

  // Note that we must test the size on each iteration, the worklist can grow.
  for (unsigned Idx = 0; Idx != Worklist.size(); ++Idx) {
    const Value *OrigV = Worklist[Idx];
    auto *I = dyn_cast_or_null<Instruction>(VMap.lookup(OrigV));
    if (!I)
      continue;

    // Skip over non-intrinsic callsites, we don't want to remove any nodes from
    // the CGSCC.
    CallSite CS = CallSite(I);

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

  // 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, TD);

  // 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 Begin = cast<BasicBlock>(VMap[&OldFunc->getEntryBlock()]);
  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)) {

Function *PartialInlinerImpl::unswitchFunction(Function *F) {
  // First, verify that this function is an unswitching candidate...
  BasicBlock *EntryBlock = &F->front();
  BranchInst *BR = dyn_cast<BranchInst>(EntryBlock->getTerminator());
  if (!BR || BR->isUnconditional())
    return nullptr;

  BasicBlock *ReturnBlock = nullptr;
  BasicBlock *NonReturnBlock = nullptr;
  unsigned ReturnCount = 0;
  for (BasicBlock *BB : successors(EntryBlock)) {
    if (isa<ReturnInst>(BB->getTerminator())) {
      ReturnBlock = BB;
      ReturnCount++;
    } else
      NonReturnBlock = BB;
  }

  if (ReturnCount != 1)
    return nullptr;

  // Clone the function, so that we can hack away on it.
  ValueToValueMapTy VMap;
  Function *DuplicateFunction = CloneFunction(F, VMap);
  DuplicateFunction->setLinkage(GlobalValue::InternalLinkage);
  BasicBlock *NewEntryBlock = cast<BasicBlock>(VMap[EntryBlock]);
  BasicBlock *NewReturnBlock = cast<BasicBlock>(VMap[ReturnBlock]);
  BasicBlock *NewNonReturnBlock = cast<BasicBlock>(VMap[NonReturnBlock]);

  // Go ahead and update all uses to the duplicate, so that we can just
  // use the inliner functionality when we're done hacking.
  F->replaceAllUsesWith(DuplicateFunction);

  // Special hackery is needed with PHI nodes that have inputs from more than
  // one extracted block.  For simplicity, just split the PHIs into a two-level
  // sequence of PHIs, some of which will go in the extracted region, and some
  // of which will go outside.
  BasicBlock *PreReturn = NewReturnBlock;
  NewReturnBlock = NewReturnBlock->splitBasicBlock(
      NewReturnBlock->getFirstNonPHI()->getIterator());
  BasicBlock::iterator I = PreReturn->begin();
  Instruction *Ins = &NewReturnBlock->front();
  while (I != PreReturn->end()) {
    PHINode *OldPhi = dyn_cast<PHINode>(I);
    if (!OldPhi)
      break;

    PHINode *RetPhi = PHINode::Create(OldPhi->getType(), 2, "", Ins);
    OldPhi->replaceAllUsesWith(RetPhi);
    Ins = NewReturnBlock->getFirstNonPHI();

    RetPhi->addIncoming(&*I, PreReturn);
    RetPhi->addIncoming(OldPhi->getIncomingValueForBlock(NewEntryBlock),
                        NewEntryBlock);
    OldPhi->removeIncomingValue(NewEntryBlock);

    ++I;
  }
  NewEntryBlock->getTerminator()->replaceUsesOfWith(PreReturn, NewReturnBlock);

  // Gather up the blocks that we're going to extract.
  std::vector<BasicBlock *> ToExtract;
  ToExtract.push_back(NewNonReturnBlock);
  for (BasicBlock &BB : *DuplicateFunction)
    if (&BB != NewEntryBlock && &BB != NewReturnBlock &&
        &BB != NewNonReturnBlock)
      ToExtract.push_back(&BB);

  // The CodeExtractor needs a dominator tree.
  DominatorTree DT;
  DT.recalculate(*DuplicateFunction);

  // Extract the body of the if.
  Function *ExtractedFunction =
      CodeExtractor(ToExtract, &DT).extractCodeRegion();

  // Inline the top-level if test into all callers.
  std::vector<User *> Users(DuplicateFunction->user_begin(),
                            DuplicateFunction->user_end());
  for (User *User : Users)
    if (CallInst *CI = dyn_cast<CallInst>(User))
      InlineFunction(CI, IFI);
    else if (InvokeInst *II = dyn_cast<InvokeInst>(User))
      InlineFunction(II, IFI);

  // Ditch the duplicate, since we're done with it, and rewrite all remaining
  // users (function pointers, etc.) back to the original function.
  DuplicateFunction->replaceAllUsesWith(F);
  DuplicateFunction->eraseFromParent();

  ++NumPartialInlined;

  return ExtractedFunction;
}

/// NormalizeLandingPads - Normalize and discover landing pads, noting them
/// in the LandingPads set.  A landing pad is normal if the only CFG edges
/// that end at it are unwind edges from invoke instructions. If we inlined
/// through an invoke we could have a normal branch from the previous
/// unwind block through to the landing pad for the original invoke.
/// Abnormal landing pads are fixed up by redirecting all unwind edges to
/// a new basic block which falls through to the original.
bool DwarfEHPrepare::NormalizeLandingPads() {
  bool Changed = false;

  const MCAsmInfo *MAI = TM->getMCAsmInfo();
  bool usingSjLjEH = MAI->getExceptionHandlingType() == ExceptionHandling::SjLj;

  for (Function::iterator I = F->begin(), E = F->end(); I != E; ++I) {
    TerminatorInst *TI = I->getTerminator();
    if (!isa<InvokeInst>(TI))
      continue;
    BasicBlock *LPad = TI->getSuccessor(1);
    // Skip landing pads that have already been normalized.
    if (LandingPads.count(LPad))
      continue;

    // Check that only invoke unwind edges end at the landing pad.
    bool OnlyUnwoundTo = true;
    bool SwitchOK = usingSjLjEH;
    for (pred_iterator PI = pred_begin(LPad), PE = pred_end(LPad);
         PI != PE; ++PI) {
      TerminatorInst *PT = (*PI)->getTerminator();
      // The SjLj dispatch block uses a switch instruction. This is effectively
      // an unwind edge, so we can disregard it here. There will only ever
      // be one dispatch, however, so if there are multiple switches, one
      // of them truly is a normal edge, not an unwind edge.
      if (SwitchOK && isa<SwitchInst>(PT)) {
        SwitchOK = false;
        continue;
      }
      if (!isa<InvokeInst>(PT) || LPad == PT->getSuccessor(0)) {
        OnlyUnwoundTo = false;
        break;
      }
    }

    if (OnlyUnwoundTo) {
      // Only unwind edges lead to the landing pad.  Remember the landing pad.
      LandingPads.insert(LPad);
      continue;
    }

    // At least one normal edge ends at the landing pad.  Redirect the unwind
    // edges to a new basic block which falls through into this one.

    // Create the new basic block.
    BasicBlock *NewBB = BasicBlock::Create(F->getContext(),
                                           LPad->getName() + "_unwind_edge");

    // Insert it into the function right before the original landing pad.
    LPad->getParent()->getBasicBlockList().insert(LPad, NewBB);

    // Redirect unwind edges from the original landing pad to NewBB.
    for (pred_iterator PI = pred_begin(LPad), PE = pred_end(LPad); PI != PE; ) {
      TerminatorInst *PT = (*PI++)->getTerminator();
      if (isa<InvokeInst>(PT) && PT->getSuccessor(1) == LPad)
        // Unwind to the new block.
        PT->setSuccessor(1, NewBB);
    }

    // If there are any PHI nodes in LPad, we need to update them so that they
    // merge incoming values from NewBB instead.
    for (BasicBlock::iterator II = LPad->begin(); isa<PHINode>(II); ++II) {
      PHINode *PN = cast<PHINode>(II);
      pred_iterator PB = pred_begin(NewBB), PE = pred_end(NewBB);

      // Check to see if all of the values coming in via unwind edges are the
      // same.  If so, we don't need to create a new PHI node.
      Value *InVal = PN->getIncomingValueForBlock(*PB);
      for (pred_iterator PI = PB; PI != PE; ++PI) {
        if (PI != PB && InVal != PN->getIncomingValueForBlock(*PI)) {
          InVal = 0;
          break;
        }
      }

      if (InVal == 0) {
        // Different unwind edges have different values.  Create a new PHI node
        // in NewBB.
        PHINode *NewPN = PHINode::Create(PN->getType(),
                                         PN->getNumIncomingValues(),
                                         PN->getName()+".unwind", NewBB);
        // Add an entry for each unwind edge, using the value from the old PHI.
        for (pred_iterator PI = PB; PI != PE; ++PI)
          NewPN->addIncoming(PN->getIncomingValueForBlock(*PI), *PI);

        // Now use this new PHI as the common incoming value for NewBB in PN.
        InVal = NewPN;
      }

      // Revector exactly one entry in the PHI node to come from NewBB
      // and delete all other entries that come from unwind edges.  If
      // there are both normal and unwind edges from the same predecessor,
      // this leaves an entry for the normal edge.
//.........这里部分代码省略.........

Function* LLVMSignatureInfo::createFunctionStub(bool special, bool virt) {
  
  std::vector<Value*> Args;
  std::vector<Value*> FunctionArgs;
  std::vector<Value*> TempArgs;
  
  J3Intrinsics& Intrinsics = *Compiler->getIntrinsics();
  Function* stub = NULL;
  FunctionType* FTy = (virt || special)? getVirtualType() : getStaticType();
  if (virt) {
    stub = Compiler->virtualStubs[FTy];
  } else if (special) {
    stub = Compiler->specialStubs[FTy];
  } else {
    stub = Compiler->staticStubs[FTy];
  }
  if (stub != NULL) {
    return stub;
  }
  if (Compiler->isStaticCompiling()) {
    vmkit::ThreadAllocator allocator;
    const char* type = virt ? "virtual_stub" : special ? "special_stub" : "static_stub";
    char* buf = (char*)allocator.Allocate(
        (signature->keyName->size << 1) + 1 + 11);
    signature->nativeName(buf, type);
    stub = Function::Create(
        FTy, GlobalValue::ExternalLinkage, buf, Compiler->getLLVMModule());
  

  } else {
    stub = Function::Create(
        FTy, GlobalValue::ExternalLinkage, "", Compiler->getLLVMModule());
  }
  LLVMContext& context = Compiler->getLLVMModule()->getContext();
  
  BasicBlock* currentBlock = BasicBlock::Create(context, "enter", stub);
  BasicBlock* endBlock = BasicBlock::Create(context, "end", stub);
  BasicBlock* callBlock = BasicBlock::Create(context, "call", stub);
  PHINode* node = NULL;
  if (!signature->getReturnType()->isVoid()) {
    node = PHINode::Create(stub->getReturnType(), 2, "", endBlock);
  }
    

  for (Function::arg_iterator arg = stub->arg_begin();
       arg != stub->arg_end(); ++arg) {
    Value* temp = arg;
    if (Compiler->useCooperativeGC() &&
        arg->getType() == Intrinsics.JavaObjectType) {
      temp = new AllocaInst(Intrinsics.JavaObjectType, "", currentBlock);
      new StoreInst(arg, temp, "", currentBlock);
      Value* GCArgs[2] = {
        new BitCastInst(temp, Intrinsics.ptrPtrType, "", currentBlock),
        Intrinsics.constantPtrNull
      };
        
      CallInst::Create(Intrinsics.llvm_gc_gcroot, GCArgs, "", currentBlock);
    }
    
    TempArgs.push_back(temp);
  }

  if (virt) {
    if (Compiler->useCooperativeGC()) {
      Args.push_back(new LoadInst(TempArgs[0], "", false, currentBlock));
    } else {
      Args.push_back(TempArgs[0]);
    }
  }

  Value* val = CallInst::Create(virt ? Intrinsics.ResolveVirtualStubFunction :
                                special ? Intrinsics.ResolveSpecialStubFunction:
                                          Intrinsics.ResolveStaticStubFunction,
                                Args, "", currentBlock);
  
  Constant* nullValue = Constant::getNullValue(val->getType());
  Value* cmp = new ICmpInst(*currentBlock, ICmpInst::ICMP_EQ,
                            nullValue, val, "");
  BranchInst::Create(endBlock, callBlock, cmp, currentBlock);
  if (node) node->addIncoming(Constant::getNullValue(node->getType()),
                              currentBlock);

  currentBlock = callBlock;
  Value* Func = new BitCastInst(val, stub->getType(), "", currentBlock);
  
  int i = 0;
  for (Function::arg_iterator arg = stub->arg_begin();
       arg != stub->arg_end(); ++arg, ++i) {
    Value* temp = arg;
    if (Compiler->useCooperativeGC() &&
        arg->getType() == Intrinsics.JavaObjectType) {
      temp = new LoadInst(TempArgs[i], "", false, currentBlock);
    }
    FunctionArgs.push_back(temp);
  }
  Value* res = CallInst::Create(Func, FunctionArgs, "", currentBlock);
  if (node) node->addIncoming(res, currentBlock);
  BranchInst::Create(endBlock, currentBlock);

  currentBlock = endBlock;
//.........这里部分代码省略.........

/// SplitBlockPredecessors - This method transforms BB by introducing a new
/// basic block into the function, and moving some of the predecessors of BB to
/// be predecessors of the new block.  The new predecessors are indicated by the
/// Preds array, which has NumPreds elements in it.  The new block is given a
/// suffix of 'Suffix'.
///
/// This currently updates the LLVM IR, AliasAnalysis, DominatorTree,
/// LoopInfo, and LCCSA but no other analyses. In particular, it does not
/// preserve LoopSimplify (because it's complicated to handle the case where one
/// of the edges being split is an exit of a loop with other exits).
///
BasicBlock *llvm::SplitBlockPredecessors(BasicBlock *BB, 
                                         BasicBlock *const *Preds,
                                         unsigned NumPreds, const char *Suffix,
                                         Pass *P) {
  // Create new basic block, insert right before the original block.
  BasicBlock *NewBB = BasicBlock::Create(BB->getContext(), BB->getName()+Suffix,
                                         BB->getParent(), BB);
  
  // The new block unconditionally branches to the old block.
  BranchInst *BI = BranchInst::Create(BB, NewBB);
  
  LoopInfo *LI = P ? P->getAnalysisIfAvailable<LoopInfo>() : 0;
  Loop *L = LI ? LI->getLoopFor(BB) : 0;
  bool PreserveLCSSA = P->mustPreserveAnalysisID(LCSSAID);

  // Move the edges from Preds to point to NewBB instead of BB.
  // While here, if we need to preserve loop analyses, collect
  // some information about how this split will affect loops.
  bool HasLoopExit = false;
  bool IsLoopEntry = !!L;
  bool SplitMakesNewLoopHeader = false;
  for (unsigned i = 0; i != NumPreds; ++i) {
    // This is slightly more strict than necessary; the minimum requirement
    // is that there be no more than one indirectbr branching to BB. And
    // all BlockAddress uses would need to be updated.
    assert(!isa<IndirectBrInst>(Preds[i]->getTerminator()) &&
           "Cannot split an edge from an IndirectBrInst");

    Preds[i]->getTerminator()->replaceUsesOfWith(BB, NewBB);

    if (LI) {
      // If we need to preserve LCSSA, determine if any of
      // the preds is a loop exit.
      if (PreserveLCSSA)
        if (Loop *PL = LI->getLoopFor(Preds[i]))
          if (!PL->contains(BB))
            HasLoopExit = true;
      // If we need to preserve LoopInfo, note whether any of the
      // preds crosses an interesting loop boundary.
      if (L) {
        if (L->contains(Preds[i]))
          IsLoopEntry = false;
        else
          SplitMakesNewLoopHeader = true;
      }
    }
  }

  // Update dominator tree if available.
  DominatorTree *DT = P ? P->getAnalysisIfAvailable<DominatorTree>() : 0;
  if (DT)
    DT->splitBlock(NewBB);

  // Insert a new PHI node into NewBB for every PHI node in BB and that new PHI
  // node becomes an incoming value for BB's phi node.  However, if the Preds
  // list is empty, we need to insert dummy entries into the PHI nodes in BB to
  // account for the newly created predecessor.
  if (NumPreds == 0) {
    // Insert dummy values as the incoming value.
    for (BasicBlock::iterator I = BB->begin(); isa<PHINode>(I); ++I)
      cast<PHINode>(I)->addIncoming(UndefValue::get(I->getType()), NewBB);
    return NewBB;
  }

  AliasAnalysis *AA = P ? P->getAnalysisIfAvailable<AliasAnalysis>() : 0;

  if (L) {
    if (IsLoopEntry) {
      // Add the new block to the nearest enclosing loop (and not an
      // adjacent loop). To find this, examine each of the predecessors and
      // determine which loops enclose them, and select the most-nested loop
      // which contains the loop containing the block being split.
      Loop *InnermostPredLoop = 0;
      for (unsigned i = 0; i != NumPreds; ++i)
        if (Loop *PredLoop = LI->getLoopFor(Preds[i])) {
          // Seek a loop which actually contains the block being split (to
          // avoid adjacent loops).
          while (PredLoop && !PredLoop->contains(BB))
            PredLoop = PredLoop->getParentLoop();
          // Select the most-nested of these loops which contains the block.
          if (PredLoop &&
              PredLoop->contains(BB) &&
              (!InnermostPredLoop ||
               InnermostPredLoop->getLoopDepth() < PredLoop->getLoopDepth()))
            InnermostPredLoop = PredLoop;
        }
      if (InnermostPredLoop)
        InnermostPredLoop->addBasicBlockToLoop(NewBB, LI->getBase());
    } else {
//.........这里部分代码省略.........

/// Create a clone of the blocks in a loop and connect them together.
/// If CreateRemainderLoop is false, loop structure will not be cloned,
/// otherwise a new loop will be created including all cloned blocks, and the
/// iterator of it switches to count NewIter down to 0.
/// The cloned blocks should be inserted between InsertTop and InsertBot.
/// If loop structure is cloned InsertTop should be new preheader, InsertBot
/// new loop exit.
/// Return the new cloned loop that is created when CreateRemainderLoop is true.
static Loop *
CloneLoopBlocks(Loop *L, Value *NewIter, const bool CreateRemainderLoop,
                const bool UseEpilogRemainder, const bool UnrollRemainder,
                BasicBlock *InsertTop,
                BasicBlock *InsertBot, BasicBlock *Preheader,
                std::vector<BasicBlock *> &NewBlocks, LoopBlocksDFS &LoopBlocks,
                ValueToValueMapTy &VMap, DominatorTree *DT, LoopInfo *LI) {
  StringRef suffix = UseEpilogRemainder ? "epil" : "prol";
  BasicBlock *Header = L->getHeader();
  BasicBlock *Latch = L->getLoopLatch();
  Function *F = Header->getParent();
  LoopBlocksDFS::RPOIterator BlockBegin = LoopBlocks.beginRPO();
  LoopBlocksDFS::RPOIterator BlockEnd = LoopBlocks.endRPO();
  Loop *ParentLoop = L->getParentLoop();
  NewLoopsMap NewLoops;
  NewLoops[ParentLoop] = ParentLoop;
  if (!CreateRemainderLoop)
    NewLoops[L] = ParentLoop;

  // For each block in the original loop, create a new copy,
  // and update the value map with the newly created values.
  for (LoopBlocksDFS::RPOIterator BB = BlockBegin; BB != BlockEnd; ++BB) {
    BasicBlock *NewBB = CloneBasicBlock(*BB, VMap, "." + suffix, F);
    NewBlocks.push_back(NewBB);

    // If we're unrolling the outermost loop, there's no remainder loop,
    // and this block isn't in a nested loop, then the new block is not
    // in any loop. Otherwise, add it to loopinfo.
    if (CreateRemainderLoop || LI->getLoopFor(*BB) != L || ParentLoop)
      addClonedBlockToLoopInfo(*BB, NewBB, LI, NewLoops);

    VMap[*BB] = NewBB;
    if (Header == *BB) {
      // For the first block, add a CFG connection to this newly
      // created block.
      InsertTop->getTerminator()->setSuccessor(0, NewBB);
    }

    if (DT) {
      if (Header == *BB) {
        // The header is dominated by the preheader.
        DT->addNewBlock(NewBB, InsertTop);
      } else {
        // Copy information from original loop to unrolled loop.
        BasicBlock *IDomBB = DT->getNode(*BB)->getIDom()->getBlock();
        DT->addNewBlock(NewBB, cast<BasicBlock>(VMap[IDomBB]));
      }
    }

    if (Latch == *BB) {
      // For the last block, if CreateRemainderLoop is false, create a direct
      // jump to InsertBot. If not, create a loop back to cloned head.
      VMap.erase((*BB)->getTerminator());
      BasicBlock *FirstLoopBB = cast<BasicBlock>(VMap[Header]);
      BranchInst *LatchBR = cast<BranchInst>(NewBB->getTerminator());
      IRBuilder<> Builder(LatchBR);
      if (!CreateRemainderLoop) {
        Builder.CreateBr(InsertBot);
      } else {
        PHINode *NewIdx = PHINode::Create(NewIter->getType(), 2,
                                          suffix + ".iter",
                                          FirstLoopBB->getFirstNonPHI());
        Value *IdxSub =
            Builder.CreateSub(NewIdx, ConstantInt::get(NewIdx->getType(), 1),
                              NewIdx->getName() + ".sub");
        Value *IdxCmp =
            Builder.CreateIsNotNull(IdxSub, NewIdx->getName() + ".cmp");
        Builder.CreateCondBr(IdxCmp, FirstLoopBB, InsertBot);
        NewIdx->addIncoming(NewIter, InsertTop);
        NewIdx->addIncoming(IdxSub, NewBB);
      }
      LatchBR->eraseFromParent();
    }
  }

  // Change the incoming values to the ones defined in the preheader or
  // cloned loop.
  for (BasicBlock::iterator I = Header->begin(); isa<PHINode>(I); ++I) {
    PHINode *NewPHI = cast<PHINode>(VMap[&*I]);
    if (!CreateRemainderLoop) {
      if (UseEpilogRemainder) {
        unsigned idx = NewPHI->getBasicBlockIndex(Preheader);
        NewPHI->setIncomingBlock(idx, InsertTop);
        NewPHI->removeIncomingValue(Latch, false);
      } else {
        VMap[&*I] = NewPHI->getIncomingValueForBlock(Preheader);
        cast<BasicBlock>(VMap[Header])->getInstList().erase(NewPHI);
      }
    } else {
      unsigned idx = NewPHI->getBasicBlockIndex(Preheader);
      NewPHI->setIncomingBlock(idx, InsertTop);
      BasicBlock *NewLatch = cast<BasicBlock>(VMap[Latch]);
//.........这里部分代码省略.........

/// RewriteLoopExitValues - Check to see if this loop has a computable
/// loop-invariant execution count.  If so, this means that we can compute the
/// final value of any expressions that are recurrent in the loop, and
/// substitute the exit values from the loop into any instructions outside of
/// the loop that use the final values of the current expressions.
///
/// This is mostly redundant with the regular IndVarSimplify activities that
/// happen later, except that it's more powerful in some cases, because it's
/// able to brute-force evaluate arbitrary instructions as long as they have
/// constant operands at the beginning of the loop.
void IndVarSimplify::RewriteLoopExitValues(Loop *L,
                                           SCEVExpander &Rewriter) {
  // Verify the input to the pass in already in LCSSA form.
  assert(L->isLCSSAForm());

  SmallVector<BasicBlock*, 8> ExitBlocks;
  L->getUniqueExitBlocks(ExitBlocks);

  // Find all values that are computed inside the loop, but used outside of it.
  // Because of LCSSA, these values will only occur in LCSSA PHI Nodes.  Scan
  // the exit blocks of the loop to find them.
  for (unsigned i = 0, e = ExitBlocks.size(); i != e; ++i) {
    BasicBlock *ExitBB = ExitBlocks[i];

    // If there are no PHI nodes in this exit block, then no values defined
    // inside the loop are used on this path, skip it.
    PHINode *PN = dyn_cast<PHINode>(ExitBB->begin());
    if (!PN) continue;

    unsigned NumPreds = PN->getNumIncomingValues();

    // Iterate over all of the PHI nodes.
    BasicBlock::iterator BBI = ExitBB->begin();
    while ((PN = dyn_cast<PHINode>(BBI++))) {
      if (PN->use_empty())
        continue; // dead use, don't replace it

      // SCEV only supports integer expressions for now.
      if (!PN->getType()->isIntegerTy() && !PN->getType()->isPointerTy())
        continue;

      // It's necessary to tell ScalarEvolution about this explicitly so that
      // it can walk the def-use list and forget all SCEVs, as it may not be
      // watching the PHI itself. Once the new exit value is in place, there
      // may not be a def-use connection between the loop and every instruction
      // which got a SCEVAddRecExpr for that loop.
      SE->forgetValue(PN);

      // Iterate over all of the values in all the PHI nodes.
      for (unsigned i = 0; i != NumPreds; ++i) {
        // If the value being merged in is not integer or is not defined
        // in the loop, skip it.
        Value *InVal = PN->getIncomingValue(i);
        if (!isa<Instruction>(InVal))
          continue;

        // If this pred is for a subloop, not L itself, skip it.
        if (LI->getLoopFor(PN->getIncomingBlock(i)) != L)
          continue; // The Block is in a subloop, skip it.

        // Check that InVal is defined in the loop.
        Instruction *Inst = cast<Instruction>(InVal);
        if (!L->contains(Inst))
          continue;

        // Okay, this instruction has a user outside of the current loop
        // and varies predictably *inside* the loop.  Evaluate the value it
        // contains when the loop exits, if possible.
        const SCEV *ExitValue = SE->getSCEVAtScope(Inst, L->getParentLoop());
        if (!ExitValue->isLoopInvariant(L))
          continue;

        Changed = true;
        ++NumReplaced;

        Value *ExitVal = Rewriter.expandCodeFor(ExitValue, PN->getType(), Inst);

        DEBUG(dbgs() << "INDVARS: RLEV: AfterLoopVal = " << *ExitVal << '\n'
                     << "  LoopVal = " << *Inst << "\n");

        PN->setIncomingValue(i, ExitVal);

        // If this instruction is dead now, delete it.
        RecursivelyDeleteTriviallyDeadInstructions(Inst);

        if (NumPreds == 1) {
          // Completely replace a single-pred PHI. This is safe, because the
          // NewVal won't be variant in the loop, so we don't need an LCSSA phi
          // node anymore.
          PN->replaceAllUsesWith(ExitVal);
          RecursivelyDeleteTriviallyDeadInstructions(PN);
        }
      }
      if (NumPreds != 1) {
        // Clone the PHI and delete the original one. This lets IVUsers and
        // any other maps purge the original user from their records.
        PHINode *NewPN = cast<PHINode>(PN->clone());
        NewPN->takeName(PN);
        NewPN->insertBefore(PN);
        PN->replaceAllUsesWith(NewPN);
//.........这里部分代码省略.........

/// InsertUniqueBackedgeBlock - This method is called when the specified loop
/// has more than one backedge in it.  If this occurs, revector all of these
/// backedges to target a new basic block and have that block branch to the loop
/// header.  This ensures that loops have exactly one backedge.
///
BasicBlock *
LoopSimplify::InsertUniqueBackedgeBlock(Loop *L, BasicBlock *Preheader) {
  assert(L->getNumBackEdges() > 1 && "Must have > 1 backedge!");

  // Get information about the loop
  BasicBlock *Header = L->getHeader();
  Function *F = Header->getParent();

  // Unique backedge insertion currently depends on having a preheader.
  if (!Preheader)
    return 0;

  // Figure out which basic blocks contain back-edges to the loop header.
  std::vector<BasicBlock*> BackedgeBlocks;
  for (pred_iterator I = pred_begin(Header), E = pred_end(Header); I != E; ++I){
    BasicBlock *P = *I;

    // Indirectbr edges cannot be split, so we must fail if we find one.
    if (isa<IndirectBrInst>(P->getTerminator()))
      return 0;

    if (P != Preheader) BackedgeBlocks.push_back(P);
  }

  // Create and insert the new backedge block...
  BasicBlock *BEBlock = BasicBlock::Create(Header->getContext(),
                                           Header->getName()+".backedge", F);
  BranchInst *BETerminator = BranchInst::Create(Header, BEBlock);

  DEBUG(dbgs() << "LoopSimplify: Inserting unique backedge block "
               << BEBlock->getName() << "\n");

  // Move the new backedge block to right after the last backedge block.
  Function::iterator InsertPos = BackedgeBlocks.back(); ++InsertPos;
  F->getBasicBlockList().splice(InsertPos, F->getBasicBlockList(), BEBlock);

  // Now that the block has been inserted into the function, create PHI nodes in
  // the backedge block which correspond to any PHI nodes in the header block.
  for (BasicBlock::iterator I = Header->begin(); isa<PHINode>(I); ++I) {
    PHINode *PN = cast<PHINode>(I);
    PHINode *NewPN = PHINode::Create(PN->getType(), PN->getName()+".be",
                                     BETerminator);
    NewPN->reserveOperandSpace(BackedgeBlocks.size());
    if (AA) AA->copyValue(PN, NewPN);

    // Loop over the PHI node, moving all entries except the one for the
    // preheader over to the new PHI node.
    unsigned PreheaderIdx = ~0U;
    bool HasUniqueIncomingValue = true;
    Value *UniqueValue = 0;
    for (unsigned i = 0, e = PN->getNumIncomingValues(); i != e; ++i) {
      BasicBlock *IBB = PN->getIncomingBlock(i);
      Value *IV = PN->getIncomingValue(i);
      if (IBB == Preheader) {
        PreheaderIdx = i;
      } else {
        NewPN->addIncoming(IV, IBB);
        if (HasUniqueIncomingValue) {
          if (UniqueValue == 0)
            UniqueValue = IV;
          else if (UniqueValue != IV)
            HasUniqueIncomingValue = false;
        }
      }
    }

    // Delete all of the incoming values from the old PN except the preheader's
    assert(PreheaderIdx != ~0U && "PHI has no preheader entry??");
    if (PreheaderIdx != 0) {
      PN->setIncomingValue(0, PN->getIncomingValue(PreheaderIdx));
      PN->setIncomingBlock(0, PN->getIncomingBlock(PreheaderIdx));
    }
    // Nuke all entries except the zero'th.
    for (unsigned i = 0, e = PN->getNumIncomingValues()-1; i != e; ++i)
      PN->removeIncomingValue(e-i, false);

    // Finally, add the newly constructed PHI node as the entry for the BEBlock.
    PN->addIncoming(NewPN, BEBlock);

    // As an optimization, if all incoming values in the new PhiNode (which is a
    // subset of the incoming values of the old PHI node) have the same value,
    // eliminate the PHI Node.
    if (HasUniqueIncomingValue) {
      NewPN->replaceAllUsesWith(UniqueValue);
      if (AA) AA->deleteValue(NewPN);
      BEBlock->getInstList().erase(NewPN);
    }
  }

  // Now that all of the PHI nodes have been inserted and adjusted, modify the
  // backedge blocks to just to the BEBlock instead of the header.
  for (unsigned i = 0, e = BackedgeBlocks.size(); i != e; ++i) {
    TerminatorInst *TI = BackedgeBlocks[i]->getTerminator();
    for (unsigned Op = 0, e = TI->getNumSuccessors(); Op != e; ++Op)
      if (TI->getSuccessor(Op) == Header)
//.........这里部分代码省略.........

void RegionExtractor::findInputsOutputs(ValueSet &Inputs,
                                      ValueSet &Outputs) const {
  for (SetVector<BasicBlock *>::const_iterator I = Blocks.begin(),
                                               E = Blocks.end();
       I != E; ++I) {
    BasicBlock *BB = *I;

    // If a used value is defined outside the region, it's an input.  If an
    // instruction is used outside the region, it's an output.
    for (BasicBlock::iterator II = BB->begin(), IE = BB->end();
         II != IE; ++II) {
      for (User::op_iterator OI = II->op_begin(), OE = II->op_end();
           OI != OE; ++OI)
        if (definedInCaller(Blocks, *OI))
          Inputs.insert(*OI);
#if LLVM_VERSION_MINOR == 5
      for (User *U : II->users())
        if (!definedInRegion(Blocks, U)) {
#else
      for (Value::use_iterator UI = II->use_begin(), UE = II->use_end();
           UI != UE; ++UI)
        if (!definedInRegion(Blocks, *UI)) {
#endif
          Outputs.insert(II);
          break;
        }
    }
  }
}

/// severSplitPHINodes - If a PHI node has multiple inputs from outside of the
/// region, we need to split the entry block of the region so that the PHI node
/// is easier to deal with.
void RegionExtractor::severSplitPHINodes(BasicBlock *&Header) {
  unsigned NumPredsFromRegion = 0;
  unsigned NumPredsOutsideRegion = 0;

  if (Header != &Header->getParent()->getEntryBlock()) {
    PHINode *PN = dyn_cast<PHINode>(Header->begin());
    if (!PN) return; // No PHI nodes.

    // If the header node contains any PHI nodes, check to see if there is more
    // than one entry from outside the region.  If so, we need to sever the
    // header block into two.
    for (unsigned i = 0, e = PN->getNumIncomingValues(); i != e; ++i)
      if (Blocks.count(PN->getIncomingBlock(i)))
        ++NumPredsFromRegion;
      else
        ++NumPredsOutsideRegion;

    // If there is one (or fewer) predecessor from outside the region, we don't
    // need to do anything special.
    if (NumPredsOutsideRegion <= 1) return;
  }

  // Otherwise, we need to split the header block into two pieces: one
  // containing PHI nodes merging values from outside of the region, and a
  // second that contains all of the code for the block and merges back any
  // incoming values from inside of the region.
  BasicBlock::iterator AfterPHIs = Header->getFirstNonPHI();
  BasicBlock *NewBB = Header->splitBasicBlock(AfterPHIs,
                                              Header->getName()+".ce");

  // We only want to code extract the second block now, and it becomes the new
  // header of the region.
  BasicBlock *OldPred = Header;
  Blocks.remove(OldPred);
  Blocks.insert(NewBB);
  Header = NewBB;

  // Okay, update dominator sets. The blocks that dominate the new one are the
  // blocks that dominate TIBB plus the new block itself.
  if (DT)
    DT->splitBlock(NewBB);

  // Okay, now we need to adjust the PHI nodes and any branches from within the
  // region to go to the new header block instead of the old header block.
  if (NumPredsFromRegion) {
    PHINode *PN = cast<PHINode>(OldPred->begin());
    // Loop over all of the predecessors of OldPred that are in the region,
    // changing them to branch to NewBB instead.
    for (unsigned i = 0, e = PN->getNumIncomingValues(); i != e; ++i)
      if (Blocks.count(PN->getIncomingBlock(i))) {
        TerminatorInst *TI = PN->getIncomingBlock(i)->getTerminator();
        TI->replaceUsesOfWith(OldPred, NewBB);
      }

    // Okay, everything within the region is now branching to the right block, we
    // just have to update the PHI nodes now, inserting PHI nodes into NewBB.
    for (AfterPHIs = OldPred->begin(); isa<PHINode>(AfterPHIs); ++AfterPHIs) {
      PHINode *PN = cast<PHINode>(AfterPHIs);
      // Create a new PHI node in the new region, which has an incoming value
      // from OldPred of PN.
      PHINode *NewPN = PHINode::Create(PN->getType(), 1 + NumPredsFromRegion,
                                       PN->getName()+".ce", NewBB->begin());
      NewPN->addIncoming(PN, OldPred);

      // Loop over all of the incoming value in PN, moving them to NewPN if they
      // are from the extracted region.
      for (unsigned i = 0; i != PN->getNumIncomingValues(); ++i) {
//.........这里部分代码省略.........

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

// 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, InlineFunctionInfo &IFI) {
  Instruction *TheCall = CS.getInstruction();
  LLVMContext &Context = TheCall->getContext();
  assert(TheCall->getParent() && TheCall->getParent()->getParent() &&
         "Instruction not in function!");

  // If IFI has any state in it, zap it before we fill it in.
  IFI.reset();
  
  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 not a tail call, we must clear the 'tail'
  // flags on any calls that we inline.
  bool MustClearTailCallFlags =
    !(isa<CallInst>(TheCall) && cast<CallInst>(TheCall)->isTailCall());

  // If the call to the callee cannot throw, set the 'nounwind' flag on any
  // calls that we inline.
  bool MarkNoUnwind = CS.doesNotThrow();

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

  // GC poses two hazards to inlining, which only occur when the callee has GC:
  //  1. If the caller has no GC, then the callee's GC must be propagated to the
  //     caller.
  //  2. If the caller has a differing GC, it is invalid to inline.
  if (CalledFunc->hasGC()) {
    if (!Caller->hasGC())
      Caller->setGC(CalledFunc->getGC());
    else if (CalledFunc->getGC() != Caller->getGC())
      return false;
  }

  // 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.
  SmallVector<ReturnInst*, 8> Returns;
  ClonedCodeInfo InlinedFunctionInfo;
  Function::iterator FirstNewBlock;

  { // Scope to destroy VMap after cloning.
    ValueMap<const Value*, Value*> VMap;

    assert(CalledFunc->arg_size() == CS.arg_size() &&
           "No varargs calls can be inlined!");

    // Calculate the vector of arguments to pass into the function cloner, which
    // matches up the formal to the actual argument values.
    CallSite::arg_iterator AI = CS.arg_begin();
    unsigned ArgNo = 0;
    for (Function::const_arg_iterator I = CalledFunc->arg_begin(),
         E = CalledFunc->arg_end(); I != E; ++I, ++AI, ++ArgNo) {
      Value *ActualArg = *AI;

      // When byval arguments actually inlined, we need to make the copy implied
      // by them explicit.  However, we don't do this if the callee is readonly
      // or readnone, because the copy would be unneeded: the callee doesn't
      // modify the struct.
      if (CalledFunc->paramHasAttr(ArgNo+1, Attribute::ByVal) &&
          !CalledFunc->onlyReadsMemory()) {
        const Type *AggTy = cast<PointerType>(I->getType())->getElementType();
        const Type *VoidPtrTy = 
            Type::getInt8PtrTy(Context);

        // Create the alloca.  If we have TargetData, use nice alignment.
        unsigned Align = 1;
        if (IFI.TD) Align = IFI.TD->getPrefTypeAlignment(AggTy);
        Value *NewAlloca = new AllocaInst(AggTy, 0, Align, 
                                          I->getName(), 
                                          &*Caller->begin()->begin());
        // Emit a memcpy.
        const Type *Tys[3] = {VoidPtrTy, VoidPtrTy, Type::getInt64Ty(Context)};
        Function *MemCpyFn = Intrinsic::getDeclaration(Caller->getParent(),
                                                       Intrinsic::memcpy, 
                                                       Tys, 3);
        Value *DestCast = new BitCastInst(NewAlloca, VoidPtrTy, "tmp", TheCall);
        Value *SrcCast = new BitCastInst(*AI, VoidPtrTy, "tmp", TheCall);

        Value *Size;
        if (IFI.TD == 0)
          Size = ConstantExpr::getSizeOf(AggTy);
        else
//.........这里部分代码省略.........

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.........

//.........这里部分代码省略.........
      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];

      // If the alloca is already scoped to something smaller than the whole
      // function then there's no need to add redundant, less accurate markers.
      if (hasLifetimeMarkers(AI))
        continue;

      // Try to determine the size of the allocation.
      ConstantInt *AllocaSize = 0;
      if (ConstantInt *AIArraySize =
          dyn_cast<ConstantInt>(AI->getArraySize())) {
        if (IFI.TD) {
          Type *AllocaType = AI->getAllocatedType();
          uint64_t AllocaTypeSize = IFI.TD->getTypeAllocSize(AllocaType);
          uint64_t AllocaArraySize = AIArraySize->getLimitedValue();
          assert(AllocaArraySize > 0 && "array size of AllocaInst is zero");
          // Check that array size doesn't saturate uint64_t and doesn't
          // overflow when it's multiplied by type size.
          if (AllocaArraySize != ~0ULL &&
              UINT64_MAX / AllocaArraySize >= AllocaTypeSize) {
            AllocaSize = ConstantInt::get(Type::getInt64Ty(AI->getContext()),
                                          AllocaArraySize * AllocaTypeSize);
          }
        }
      }

      builder.CreateLifetimeStart(AI, AllocaSize);
      for (unsigned ri = 0, re = Returns.size(); ri != re; ++ri) {
        IRBuilder<> builder(Returns[ri]);
        builder.CreateLifetimeEnd(AI, AllocaSize);
      }
    }
  }

  // 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.
    Function *StackSave = Intrinsic::getDeclaration(M, Intrinsic::stacksave);
    Function *StackRestore=Intrinsic::getDeclaration(M,Intrinsic::stackrestore);

    // Insert the llvm.stacksave.
    CallInst *SavedPtr = IRBuilder<>(FirstNewBlock, FirstNewBlock->begin())
      .CreateCall(StackSave, "savedstack");

/// severSplitPHINodes - If a PHI node has multiple inputs from outside of the
/// region, we need to split the entry block of the region so that the PHI node
/// is easier to deal with.
void CodeExtractor::severSplitPHINodes(BasicBlock *&Header) {
  unsigned NumPredsFromRegion = 0;
  unsigned NumPredsOutsideRegion = 0;

  if (Header != &Header->getParent()->getEntryBlock()) {
    PHINode *PN = dyn_cast<PHINode>(Header->begin());
    if (!PN) return;  // No PHI nodes.

    // If the header node contains any PHI nodes, check to see if there is more
    // than one entry from outside the region.  If so, we need to sever the
    // header block into two.
    for (unsigned i = 0, e = PN->getNumIncomingValues(); i != e; ++i)
      if (Blocks.count(PN->getIncomingBlock(i)))
        ++NumPredsFromRegion;
      else
        ++NumPredsOutsideRegion;

    // If there is one (or fewer) predecessor from outside the region, we don't
    // need to do anything special.
    if (NumPredsOutsideRegion <= 1) return;
  }

  // Otherwise, we need to split the header block into two pieces: one
  // containing PHI nodes merging values from outside of the region, and a
  // second that contains all of the code for the block and merges back any
  // incoming values from inside of the region.
  BasicBlock *NewBB = SplitBlock(Header, Header->getFirstNonPHI(), DT);

  // We only want to code extract the second block now, and it becomes the new
  // header of the region.
  BasicBlock *OldPred = Header;
  Blocks.remove(OldPred);
  Blocks.insert(NewBB);
  Header = NewBB;

  // Okay, now we need to adjust the PHI nodes and any branches from within the
  // region to go to the new header block instead of the old header block.
  if (NumPredsFromRegion) {
    PHINode *PN = cast<PHINode>(OldPred->begin());
    // Loop over all of the predecessors of OldPred that are in the region,
    // changing them to branch to NewBB instead.
    for (unsigned i = 0, e = PN->getNumIncomingValues(); i != e; ++i)
      if (Blocks.count(PN->getIncomingBlock(i))) {
        TerminatorInst *TI = PN->getIncomingBlock(i)->getTerminator();
        TI->replaceUsesOfWith(OldPred, NewBB);
      }

    // Okay, everything within the region is now branching to the right block, we
    // just have to update the PHI nodes now, inserting PHI nodes into NewBB.
    BasicBlock::iterator AfterPHIs;
    for (AfterPHIs = OldPred->begin(); isa<PHINode>(AfterPHIs); ++AfterPHIs) {
      PHINode *PN = cast<PHINode>(AfterPHIs);
      // Create a new PHI node in the new region, which has an incoming value
      // from OldPred of PN.
      PHINode *NewPN = PHINode::Create(PN->getType(), 1 + NumPredsFromRegion,
                                       PN->getName() + ".ce", &NewBB->front());
      PN->replaceAllUsesWith(NewPN);
      NewPN->addIncoming(PN, OldPred);

      // Loop over all of the incoming value in PN, moving them to NewPN if they
      // are from the extracted region.
      for (unsigned i = 0; i != PN->getNumIncomingValues(); ++i) {
        if (Blocks.count(PN->getIncomingBlock(i))) {
          NewPN->addIncoming(PN->getIncomingValue(i), PN->getIncomingBlock(i));
          PN->removeIncomingValue(i);
          --i;
        }
      }
    }
  }
}