llvm/lib/Transforms/Utils/LoopUnroll.cpp

//===-- UnrollLoop.cpp - Loop unrolling utilities -------------------------===//
//
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file implements some loop unrolling utilities. It does not define any
// actual pass or policy, but provides a single function to perform loop
// unrolling.
//
// The process of unrolling can produce extraneous basic blocks linked with
// unconditional branches.  This will be corrected in the future.
//
//===----------------------------------------------------------------------===//

#include "llvm/Transforms/Utils/UnrollLoop.h"
#include "llvm/ADT/SmallPtrSet.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/Analysis/AssumptionCache.h"
#include "llvm/Analysis/InstructionSimplify.h"
#include "llvm/Analysis/LoopIterator.h"
#include "llvm/Analysis/LoopPass.h"
#include "llvm/Analysis/ScalarEvolution.h"
#include "llvm/IR/BasicBlock.h"
#include "llvm/IR/DataLayout.h"
#include "llvm/IR/DiagnosticInfo.h"
#include "llvm/IR/Dominators.h"
#include "llvm/IR/LLVMContext.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Transforms/Utils/BasicBlockUtils.h"
#include "llvm/Transforms/Utils/Cloning.h"
#include "llvm/Transforms/Utils/Local.h"
#include "llvm/Transforms/Utils/LoopUtils.h"
#include "llvm/Transforms/Utils/SimplifyIndVar.h"
using namespace llvm;

#define DEBUG_TYPE "loop-unroll"

// TODO: Should these be here or in LoopUnroll?
STATISTIC(NumCompletelyUnrolled, "Number of loops completely unrolled");
STATISTIC(NumUnrolled, "Number of loops unrolled (completely or otherwise)");

/// RemapInstruction - Convert the instruction operands from referencing the
/// current values into those specified by VMap.
static inline void RemapInstruction(Instruction *I,
                                    ValueToValueMapTy &VMap) {
  for (unsigned op = 0, E = I->getNumOperands(); op != E; ++op) {
    Value *Op = I->getOperand(op);
    ValueToValueMapTy::iterator It = VMap.find(Op);
    if (It != VMap.end())
      I->setOperand(op, It->second);
  }

  if (PHINode *PN = dyn_cast<PHINode>(I)) {
    for (unsigned i = 0, e = PN->getNumIncomingValues(); i != e; ++i) {
      ValueToValueMapTy::iterator It = VMap.find(PN->getIncomingBlock(i));
      if (It != VMap.end())
        PN->setIncomingBlock(i, cast<BasicBlock>(It->second));
    }
  }
}

/// FoldBlockIntoPredecessor - Folds a basic block into its predecessor if it
/// only has one predecessor, and that predecessor only has one successor.
/// The LoopInfo Analysis that is passed will be kept consistent.  If folding is
/// successful references to the containing loop must be removed from
/// ScalarEvolution by calling ScalarEvolution::forgetLoop because SE may have
/// references to the eliminated BB.  The argument ForgottenLoops contains a set
/// of loops that have already been forgotten to prevent redundant, expensive
/// calls to ScalarEvolution::forgetLoop.  Returns the new combined block.
static BasicBlock *
FoldBlockIntoPredecessor(BasicBlock *BB, LoopInfo* LI, LPPassManager *LPM,
                         SmallPtrSetImpl<Loop *> &ForgottenLoops) {
  // Merge basic blocks into their predecessor if there is only one distinct
  // pred, and if there is only one distinct successor of the predecessor, and
  // if there are no PHI nodes.
  BasicBlock *OnlyPred = BB->getSinglePredecessor();
  if (!OnlyPred) return nullptr;

  if (OnlyPred->getTerminator()->getNumSuccessors() != 1)
    return nullptr;

  DEBUG(dbgs() << "Merging: " << *BB << "into: " << *OnlyPred);

  // Resolve any PHI nodes at the start of the block.  They are all
  // guaranteed to have exactly one entry if they exist, unless there are
  // multiple duplicate (but guaranteed to be equal) entries for the
  // incoming edges.  This occurs when there are multiple edges from
  // OnlyPred to OnlySucc.
  FoldSingleEntryPHINodes(BB);

  // Delete the unconditional branch from the predecessor...
  OnlyPred->getInstList().pop_back();

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

  // Move all definitions in the successor to the predecessor...
  OnlyPred->getInstList().splice(OnlyPred->end(), BB->getInstList());

  // OldName will be valid until erased.
  StringRef OldName = BB->getName();

  // Erase basic block from the function...

  // ScalarEvolution holds references to loop exit blocks.
  if (LPM) {
    if (auto *SEWP =
            LPM->getAnalysisIfAvailable<ScalarEvolutionWrapperPass>()) {
      if (Loop *L = LI->getLoopFor(BB)) {
        if (ForgottenLoops.insert(L).second)
          SEWP->getSE().forgetLoop(L);
      }
    }
  }
  LI->removeBlock(BB);

  // Inherit predecessor's name if it exists...
  if (!OldName.empty() && !OnlyPred->hasName())
    OnlyPred->setName(OldName);

  BB->eraseFromParent();

  return OnlyPred;
}

/// Unroll the given loop by Count. The loop must be in LCSSA form. Returns true
/// if unrolling was successful, or false if the loop was unmodified. Unrolling
/// can only fail when the loop's latch block is not terminated by a conditional
/// branch instruction. However, if the trip count (and multiple) are not known,
/// loop unrolling will mostly produce more code that is no faster.
///
/// TripCount is generally defined as the number of times the loop header
/// executes. UnrollLoop relaxes the definition to permit early exits: here
/// TripCount is the iteration on which control exits LatchBlock if no early
/// exits were taken. Note that UnrollLoop assumes that the loop counter test
/// terminates LatchBlock in order to remove unnecesssary instances of the
/// test. In other words, control may exit the loop prior to TripCount
/// iterations via an early branch, but control may not exit the loop from the
/// LatchBlock's terminator prior to TripCount iterations.
///
/// Similarly, TripMultiple divides the number of times that the LatchBlock may
/// execute without exiting the loop.
///
/// If AllowRuntime is true then UnrollLoop will consider unrolling loops that
/// have a runtime (i.e. not compile time constant) trip count.  Unrolling these
/// loops require a unroll "prologue" that runs "RuntimeTripCount % Count"
/// iterations before branching into the unrolled loop.  UnrollLoop will not
/// runtime-unroll the loop if computing RuntimeTripCount will be expensive and
/// AllowExpensiveTripCount is false.
///
/// The LoopInfo Analysis that is passed will be kept consistent.
///
/// If a LoopPassManager is passed in, and the loop is fully removed, it will be
/// removed from the LoopPassManager as well. LPM can also be NULL.
///
/// This utility preserves LoopInfo. If DominatorTree or ScalarEvolution are
/// available from the Pass it must also preserve those analyses.
bool llvm::UnrollLoop(Loop *L, unsigned Count, unsigned TripCount,
                      bool AllowRuntime, bool AllowExpensiveTripCount,
                      unsigned TripMultiple, LoopInfo *LI, Pass *PP,
                      LPPassManager *LPM, AssumptionCache *AC) {
  BasicBlock *Preheader = L->getLoopPreheader();
  if (!Preheader) {
    DEBUG(dbgs() << "  Can't unroll; loop preheader-insertion failed.\n");
    return false;
  }

  BasicBlock *LatchBlock = L->getLoopLatch();
  if (!LatchBlock) {
    DEBUG(dbgs() << "  Can't unroll; loop exit-block-insertion failed.\n");
    return false;
  }

  // Loops with indirectbr cannot be cloned.
  if (!L->isSafeToClone()) {
    DEBUG(dbgs() << "  Can't unroll; Loop body cannot be cloned.\n");
    return false;
  }

  BasicBlock *Header = L->getHeader();
  BranchInst *BI = dyn_cast<BranchInst>(LatchBlock->getTerminator());

  if (!BI || BI->isUnconditional()) {
    // The loop-rotate pass can be helpful to avoid this in many cases.
    DEBUG(dbgs() <<
             "  Can't unroll; loop not terminated by a conditional branch.\n");
    return false;
  }

  if (Header->hasAddressTaken()) {
    // The loop-rotate pass can be helpful to avoid this in many cases.
    DEBUG(dbgs() <<
          "  Won't unroll loop: address of header block is taken.\n");
    return false;
  }

  if (TripCount != 0)
    DEBUG(dbgs() << "  Trip Count = " << TripCount << "\n");
  if (TripMultiple != 1)
    DEBUG(dbgs() << "  Trip Multiple = " << TripMultiple << "\n");

  // Effectively "DCE" unrolled iterations that are beyond the tripcount
  // and will never be executed.
  if (TripCount != 0 && Count > TripCount)
    Count = TripCount;

  // Don't enter the unroll code if there is nothing to do. This way we don't
  // need to support "partial unrolling by 1".
  if (TripCount == 0 && Count < 2)
    return false;

  assert(Count > 0);
  assert(TripMultiple > 0);
  assert(TripCount == 0 || TripCount % TripMultiple == 0);

  // Are we eliminating the loop control altogether?
  bool CompletelyUnroll = Count == TripCount;

  // We assume a run-time trip count if the compiler cannot
  // figure out the loop trip count and the unroll-runtime
  // flag is specified.
  bool RuntimeTripCount = (TripCount == 0 && Count > 0 && AllowRuntime);

  if (RuntimeTripCount &&
      !UnrollRuntimeLoopProlog(L, Count, AllowExpensiveTripCount, LI, LPM))
    return false;

  // Notify ScalarEvolution that the loop will be substantially changed,
  // if not outright eliminated.
  auto *SEWP =
      PP ? PP->getAnalysisIfAvailable<ScalarEvolutionWrapperPass>() : nullptr;
  ScalarEvolution *SE = SEWP ? &SEWP->getSE() : nullptr;
  if (SE)
    SE->forgetLoop(L);

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

  // Report the unrolling decision.
  DebugLoc LoopLoc = L->getStartLoc();
  Function *F = Header->getParent();
  LLVMContext &Ctx = F->getContext();

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

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

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

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

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

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

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

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

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

      // Tell LI about New.
      if (*BB == Header) {
        assert(LI->getLoopFor(*BB) == L && "Header should not be in a sub-loop");
        L->addBasicBlockToLoop(New, *LI);
      } else {
        // Figure out which loop New is in.
        const Loop *OldLoop = LI->getLoopFor(*BB);
        assert(OldLoop && "Should (at least) be in the loop being unrolled!");

        Loop *&NewLoop = NewLoops[OldLoop];
        if (!NewLoop) {
          // Found a new sub-loop.
          assert(*BB == OldLoop->getHeader() &&
                 "Header should be first in RPO");

          Loop *NewLoopParent = NewLoops.lookup(OldLoop->getParentLoop());
          assert(NewLoopParent &&
                 "Expected parent loop before sub-loop in RPO");
          NewLoop = new Loop;
          NewLoopParent->addChildLoop(NewLoop);

          // Forget the old loop, since its inputs may have changed.
          if (SE)
            SE->forgetLoop(OldLoop);
        }
        NewLoop->addBasicBlockToLoop(New, *LI);
      }

      if (*BB == Header)
        // Loop over all of the PHI nodes in the block, changing them to use
        // the incoming values from the previous block.
        for (unsigned i = 0, e = OrigPHINode.size(); i != e; ++i) {
          PHINode *NewPHI = cast<PHINode>(VMap[OrigPHINode[i]]);
          Value *InVal = NewPHI->getIncomingValueForBlock(LatchBlock);
          if (Instruction *InValI = dyn_cast<Instruction>(InVal))
            if (It > 1 && L->contains(InValI))
              InVal = LastValueMap[InValI];
          VMap[OrigPHINode[i]] = InVal;
          New->getInstList().erase(NewPHI);
        }

      // Update our running map of newest clones
      LastValueMap[*BB] = New;
      for (ValueToValueMapTy::iterator VI = VMap.begin(), VE = VMap.end();
           VI != VE; ++VI)
        LastValueMap[VI->first] = VI->second;

      // Add phi entries for newly created values to all exit blocks.
      for (succ_iterator SI = succ_begin(*BB), SE = succ_end(*BB);
           SI != SE; ++SI) {
        if (L->contains(*SI))
          continue;
        for (BasicBlock::iterator BBI = (*SI)->begin();
             PHINode *phi = dyn_cast<PHINode>(BBI); ++BBI) {
          Value *Incoming = phi->getIncomingValueForBlock(*BB);
          ValueToValueMapTy::iterator It = LastValueMap.find(Incoming);
          if (It != LastValueMap.end())
            Incoming = It->second;
          phi->addIncoming(Incoming, New);
        }
      }
      // Keep track of new headers and latches as we create them, so that
      // we can insert the proper branches later.
      if (*BB == Header)
        Headers.push_back(New);
      if (*BB == LatchBlock)
        Latches.push_back(New);

      NewBlocks.push_back(New);
    }

    // Remap all instructions in the most recent iteration
    for (unsigned i = 0; i < NewBlocks.size(); ++i)
      for (BasicBlock::iterator I = NewBlocks[i]->begin(),
           E = NewBlocks[i]->end(); I != E; ++I)
        ::RemapInstruction(&*I, LastValueMap);
  }

  // Loop over the PHI nodes in the original block, setting incoming values.
  for (unsigned i = 0, e = OrigPHINode.size(); i != e; ++i) {
    PHINode *PN = OrigPHINode[i];
    if (CompletelyUnroll) {
      PN->replaceAllUsesWith(PN->getIncomingValueForBlock(Preheader));
      Header->getInstList().erase(PN);
    }
    else if (Count > 1) {
      Value *InVal = PN->removeIncomingValue(LatchBlock, false);
      // If this value was defined in the loop, take the value defined by the
      // last iteration of the loop.
      if (Instruction *InValI = dyn_cast<Instruction>(InVal)) {
        if (L->contains(InValI))
          InVal = LastValueMap[InVal];
      }
      assert(Latches.back() == LastValueMap[LatchBlock] && "bad last latch");
      PN->addIncoming(InVal, Latches.back());
    }
  }

  // Now that all the basic blocks for the unrolled iterations are in place,
  // set up the branches to connect them.
  for (unsigned i = 0, e = Latches.size(); i != e; ++i) {
    // The original branch was replicated in each unrolled iteration.
    BranchInst *Term = cast<BranchInst>(Latches[i]->getTerminator());

    // The branch destination.
    unsigned j = (i + 1) % e;
    BasicBlock *Dest = Headers[j];
    bool NeedConditional = true;

    if (RuntimeTripCount && j != 0) {
      NeedConditional = false;
    }

    // For a complete unroll, make the last iteration end with a branch
    // to the exit block.
    if (CompletelyUnroll) {
      if (j == 0)
        Dest = LoopExit;
      NeedConditional = false;
    }

    // If we know the trip count or a multiple of it, we can safely use an
    // unconditional branch for some iterations.
    if (j != BreakoutTrip && (TripMultiple == 0 || j % TripMultiple != 0)) {
      NeedConditional = false;
    }

    if (NeedConditional) {
      // Update the conditional branch's successor for the following
      // iteration.
      Term->setSuccessor(!ContinueOnTrue, Dest);
    } else {
      // Remove phi operands at this loop exit
      if (Dest != LoopExit) {
        BasicBlock *BB = Latches[i];
        for (succ_iterator SI = succ_begin(BB), SE = succ_end(BB);
             SI != SE; ++SI) {
          if (*SI == Headers[i])
            continue;
          for (BasicBlock::iterator BBI = (*SI)->begin();
               PHINode *Phi = dyn_cast<PHINode>(BBI); ++BBI) {
            Phi->removeIncomingValue(BB, false);
          }
        }
      }
      // Replace the conditional branch with an unconditional one.
      BranchInst::Create(Dest, Term);
      Term->eraseFromParent();
    }
  }

  // Merge adjacent basic blocks, if possible.
  SmallPtrSet<Loop *, 4> ForgottenLoops;
  for (unsigned i = 0, e = Latches.size(); i != e; ++i) {
    BranchInst *Term = cast<BranchInst>(Latches[i]->getTerminator());
    if (Term->isUnconditional()) {
      BasicBlock *Dest = Term->getSuccessor(0);
      if (BasicBlock *Fold = FoldBlockIntoPredecessor(Dest, LI, LPM,
                                                      ForgottenLoops))
        std::replace(Latches.begin(), Latches.end(), Dest, Fold);
    }
  }

  // FIXME: We could register any cloned assumptions instead of clearing the
  // whole function's cache.
  AC->clear();

  DominatorTree *DT = nullptr;
  if (PP) {
    // FIXME: Reconstruct dom info, because it is not preserved properly.
    // Incrementally updating domtree after loop unrolling would be easy.
    if (DominatorTreeWrapperPass *DTWP =
            PP->getAnalysisIfAvailable<DominatorTreeWrapperPass>()) {
      DT = &DTWP->getDomTree();
      DT->recalculate(*L->getHeader()->getParent());
    }

    // Simplify any new induction variables in the partially unrolled loop.
    if (SE && !CompletelyUnroll) {
      SmallVector<WeakVH, 16> DeadInsts;
      simplifyLoopIVs(L, SE, DT, LPM, DeadInsts);

      // Aggressively clean up dead instructions that simplifyLoopIVs already
      // identified. Any remaining should be cleaned up below.
      while (!DeadInsts.empty())
        if (Instruction *Inst =
            dyn_cast_or_null<Instruction>(&*DeadInsts.pop_back_val()))
          RecursivelyDeleteTriviallyDeadInstructions(Inst);
    }
  }
  // At this point, the code is well formed.  We now do a quick sweep over the
  // inserted code, doing constant propagation and dead code elimination as we
  // go.
  const DataLayout &DL = Header->getModule()->getDataLayout();
  const std::vector<BasicBlock*> &NewLoopBlocks = L->getBlocks();
  for (std::vector<BasicBlock*>::const_iterator BB = NewLoopBlocks.begin(),
       BBE = NewLoopBlocks.end(); BB != BBE; ++BB)
    for (BasicBlock::iterator I = (*BB)->begin(), E = (*BB)->end(); I != E; ) {
      Instruction *Inst = &*I++;

      if (isInstructionTriviallyDead(Inst))
        (*BB)->getInstList().erase(Inst);
      else if (Value *V = SimplifyInstruction(Inst, DL))
        if (LI->replacementPreservesLCSSAForm(Inst, V)) {
          Inst->replaceAllUsesWith(V);
          (*BB)->getInstList().erase(Inst);
        }
    }

  NumCompletelyUnrolled += CompletelyUnroll;
  ++NumUnrolled;

  Loop *OuterL = L->getParentLoop();
  // Remove the loop from the LoopPassManager if it's completely removed.
  if (CompletelyUnroll && LPM != nullptr)
    LPM->deleteLoopFromQueue(L);

  // If we have a pass and a DominatorTree we should re-simplify impacted loops
  // to ensure subsequent analyses can rely on this form. We want to simplify
  // at least one layer outside of the loop that was unrolled so that any
  // changes to the parent loop exposed by the unrolling are considered.
  if (PP && DT) {
    if (!OuterL && !CompletelyUnroll)
      OuterL = L;
    if (OuterL) {
      simplifyLoop(OuterL, DT, LI, PP, SE, AC);

      // LCSSA must be performed on the outermost affected loop. The unrolled
      // loop's last loop latch is guaranteed to be in the outermost loop after
      // deleteLoopFromQueue updates LoopInfo.
      Loop *LatchLoop = LI->getLoopFor(Latches.back());
      if (!OuterL->contains(LatchLoop))
        while (OuterL->getParentLoop() != LatchLoop)
          OuterL = OuterL->getParentLoop();

      formLCSSARecursively(*OuterL, *DT, LI, SE);
    }
  }

  return true;
}

/// Given an llvm.loop loop id metadata node, returns the loop hint metadata
/// node with the given name (for example, "llvm.loop.unroll.count"). If no
/// such metadata node exists, then nullptr is returned.
MDNode *llvm::GetUnrollMetadata(MDNode *LoopID, StringRef Name) {
  // First operand should refer to the loop id itself.
  assert(LoopID->getNumOperands() > 0 && "requires at least one operand");
  assert(LoopID->getOperand(0) == LoopID && "invalid loop id");

  for (unsigned i = 1, e = LoopID->getNumOperands(); i < e; ++i) {
    MDNode *MD = dyn_cast<MDNode>(LoopID->getOperand(i));
    if (!MD)
      continue;

    MDString *S = dyn_cast<MDString>(MD->getOperand(0));
    if (!S)
      continue;

    if (Name.equals(S->getString()))
      return MD;
  }
  return nullptr;
}