llvm/lib/Transforms/Scalar/Scalarizer.cpp

   1 //===- Scalarizer.cpp - Scalarize vector operations -----------------------===//
   2 //
   3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
   4 // See https://llvm.org/LICENSE.txt for license information.
   5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
   6 //
   7 //===----------------------------------------------------------------------===//
   8 //
   9 // This pass converts vector operations into scalar operations, in order
  10 // to expose optimization opportunities on the individual scalar operations.
  11 // It is mainly intended for targets that do not have vector units, but it
  12 // may also be useful for revectorizing code to different vector widths.
  13 //
  14 //===----------------------------------------------------------------------===//
  15
  16 #include "llvm/Transforms/Scalar/Scalarizer.h"
  17 #include "llvm/ADT/PostOrderIterator.h"
  18 #include "llvm/ADT/SmallVector.h"
  19 #include "llvm/ADT/Twine.h"
  20 #include "llvm/Analysis/VectorUtils.h"
  21 #include "llvm/IR/Argument.h"
  22 #include "llvm/IR/BasicBlock.h"
  23 #include "llvm/IR/Constants.h"
  24 #include "llvm/IR/DataLayout.h"
  25 #include "llvm/IR/DerivedTypes.h"
  26 #include "llvm/IR/Dominators.h"
  27 #include "llvm/IR/Function.h"
  28 #include "llvm/IR/IRBuilder.h"
  29 #include "llvm/IR/InstVisitor.h"
  30 #include "llvm/IR/InstrTypes.h"
  31 #include "llvm/IR/Instruction.h"
  32 #include "llvm/IR/Instructions.h"
  33 #include "llvm/IR/Intrinsics.h"
  34 #include "llvm/IR/LLVMContext.h"
  35 #include "llvm/IR/Module.h"
  36 #include "llvm/IR/Type.h"
  37 #include "llvm/IR/Value.h"
  38 #include "llvm/InitializePasses.h"
  39 #include "llvm/Pass.h"
  40 #include "llvm/Support/Casting.h"
  41 #include "llvm/Support/CommandLine.h"
  42 #include "llvm/Support/MathExtras.h"
  43 #include "llvm/Transforms/Scalar.h"
  44 #include "llvm/Transforms/Utils/Local.h"
  45 #include <cassert>
  46 #include <cstdint>
  47 #include <iterator>
  48 #include <map>
  49 #include <utility>
  50
  51 using namespace llvm;
  52
  53 #define DEBUG_TYPE "scalarizer"
  54
  55 static cl::opt<bool> ScalarizeVariableInsertExtract(
  56     "scalarize-variable-insert-extract", cl::init(true), cl::Hidden,
  57     cl::desc("Allow the scalarizer pass to scalarize "
  58              "insertelement/extractelement with variable index"));
  59
  60 // This is disabled by default because having separate loads and stores
  61 // makes it more likely that the -combiner-alias-analysis limits will be
  62 // reached.
  63 static cl::opt<bool>
  64     ScalarizeLoadStore("scalarize-load-store", cl::init(false), cl::Hidden,
  65                        cl::desc("Allow the scalarizer pass to scalarize loads and store"));
  66
  67 namespace {
  68
  69 // Used to store the scattered form of a vector.
  70 using ValueVector = SmallVector<Value *, 8>;
  71
  72 // Used to map a vector Value to its scattered form.  We use std::map
  73 // because we want iterators to persist across insertion and because the
  74 // values are relatively large.
  75 using ScatterMap = std::map<Value *, ValueVector>;
  76
  77 // Lists Instructions that have been replaced with scalar implementations,
  78 // along with a pointer to their scattered forms.
  79 using GatherList = SmallVector<std::pair<Instruction *, ValueVector *>, 16>;
  80
  81 // Provides a very limited vector-like interface for lazily accessing one
  82 // component of a scattered vector or vector pointer.
  83 class Scatterer {
  84 public:
  85   Scatterer() = default;
  86
  87   // Scatter V into Size components.  If new instructions are needed,
  88   // insert them before BBI in BB.  If Cache is nonnull, use it to cache
  89   // the results.
  90   Scatterer(BasicBlock *bb, BasicBlock::iterator bbi, Value *v,
  91             ValueVector *cachePtr = nullptr);
  92
  93   // Return component I, creating a new Value for it if necessary.
  94   Value *operator[](unsigned I);
  95
  96   // Return the number of components.
  97   unsigned size() const { return Size; }
  98
  99 private:
 100   BasicBlock *BB;
 101   BasicBlock::iterator BBI;
 102   Value *V;
 103   ValueVector *CachePtr;
 104   PointerType *PtrTy;
 105   ValueVector Tmp;
 106   unsigned Size;
 107 };
 108
 109 // FCmpSpliiter(FCI)(Builder, X, Y, Name) uses Builder to create an FCmp
 110 // called Name that compares X and Y in the same way as FCI.
 111 struct FCmpSplitter {
 112   FCmpSplitter(FCmpInst &fci) : FCI(fci) {}
 113
 114   Value *operator()(IRBuilder<> &Builder, Value *Op0, Value *Op1,
 115                     const Twine &Name) const {
 116     return Builder.CreateFCmp(FCI.getPredicate(), Op0, Op1, Name);
 117   }
 118
 119   FCmpInst &FCI;
 120 };
 121
 122 // ICmpSpliiter(ICI)(Builder, X, Y, Name) uses Builder to create an ICmp
 123 // called Name that compares X and Y in the same way as ICI.
 124 struct ICmpSplitter {
 125   ICmpSplitter(ICmpInst &ici) : ICI(ici) {}
 126
 127   Value *operator()(IRBuilder<> &Builder, Value *Op0, Value *Op1,
 128                     const Twine &Name) const {
 129     return Builder.CreateICmp(ICI.getPredicate(), Op0, Op1, Name);
 130   }
 131
 132   ICmpInst &ICI;
 133 };
 134
 135 // UnarySpliiter(UO)(Builder, X, Name) uses Builder to create
 136 // a unary operator like UO called Name with operand X.
 137 struct UnarySplitter {
 138   UnarySplitter(UnaryOperator &uo) : UO(uo) {}
 139
 140   Value *operator()(IRBuilder<> &Builder, Value *Op, const Twine &Name) const {
 141     return Builder.CreateUnOp(UO.getOpcode(), Op, Name);
 142   }
 143
 144   UnaryOperator &UO;
 145 };
 146
 147 // BinarySpliiter(BO)(Builder, X, Y, Name) uses Builder to create
 148 // a binary operator like BO called Name with operands X and Y.
 149 struct BinarySplitter {
 150   BinarySplitter(BinaryOperator &bo) : BO(bo) {}
 151
 152   Value *operator()(IRBuilder<> &Builder, Value *Op0, Value *Op1,
 153                     const Twine &Name) const {
 154     return Builder.CreateBinOp(BO.getOpcode(), Op0, Op1, Name);
 155   }
 156
 157   BinaryOperator &BO;
 158 };
 159
 160 // Information about a load or store that we're scalarizing.
 161 struct VectorLayout {
 162   VectorLayout() = default;
 163
 164   // Return the alignment of element I.
 165   Align getElemAlign(unsigned I) {
 166     return commonAlignment(VecAlign, I * ElemSize);
 167   }
 168
 169   // The type of the vector.
 170   VectorType *VecTy = nullptr;
 171
 172   // The type of each element.
 173   Type *ElemTy = nullptr;
 174
 175   // The alignment of the vector.
 176   Align VecAlign;
 177
 178   // The size of each element.
 179   uint64_t ElemSize = 0;
 180 };
 181
 182 class ScalarizerVisitor : public InstVisitor<ScalarizerVisitor, bool> {
 183 public:
 184   ScalarizerVisitor(unsigned ParallelLoopAccessMDKind, DominatorTree *DT)
 185     : ParallelLoopAccessMDKind(ParallelLoopAccessMDKind), DT(DT) {
 186   }
 187
 188   bool visit(Function &F);
 189
 190   // InstVisitor methods.  They return true if the instruction was scalarized,
 191   // false if nothing changed.
 192   bool visitInstruction(Instruction &I) { return false; }
 193   bool visitSelectInst(SelectInst &SI);
 194   bool visitICmpInst(ICmpInst &ICI);
 195   bool visitFCmpInst(FCmpInst &FCI);
 196   bool visitUnaryOperator(UnaryOperator &UO);
 197   bool visitBinaryOperator(BinaryOperator &BO);
 198   bool visitGetElementPtrInst(GetElementPtrInst &GEPI);
 199   bool visitCastInst(CastInst &CI);
 200   bool visitBitCastInst(BitCastInst &BCI);
 201   bool visitInsertElementInst(InsertElementInst &IEI);
 202   bool visitExtractElementInst(ExtractElementInst &EEI);
 203   bool visitShuffleVectorInst(ShuffleVectorInst &SVI);
 204   bool visitPHINode(PHINode &PHI);
 205   bool visitLoadInst(LoadInst &LI);
 206   bool visitStoreInst(StoreInst &SI);
 207   bool visitCallInst(CallInst &ICI);
 208
 209 private:
 210   Scatterer scatter(Instruction *Point, Value *V);
 211   void gather(Instruction *Op, const ValueVector &CV);
 212   bool canTransferMetadata(unsigned Kind);
 213   void transferMetadataAndIRFlags(Instruction *Op, const ValueVector &CV);
 214   Optional<VectorLayout> getVectorLayout(Type *Ty, Align Alignment,
 215                                          const DataLayout &DL);
 216   bool finish();
 217
 218   template<typename T> bool splitUnary(Instruction &, const T &);
 219   template<typename T> bool splitBinary(Instruction &, const T &);
 220
 221   bool splitCall(CallInst &CI);
 222
 223   ScatterMap Scattered;
 224   GatherList Gathered;
 225
 226   SmallVector<WeakTrackingVH, 32> PotentiallyDeadInstrs;
 227
 228   unsigned ParallelLoopAccessMDKind;
 229
 230   DominatorTree *DT;
 231 };
 232
 233 class ScalarizerLegacyPass : public FunctionPass {
 234 public:
 235   static char ID;
 236
 237   ScalarizerLegacyPass() : FunctionPass(ID) {
 238     initializeScalarizerLegacyPassPass(*PassRegistry::getPassRegistry());
 239   }
 240
 241   bool runOnFunction(Function &F) override;
 242
 243   void getAnalysisUsage(AnalysisUsage& AU) const override {
 244     AU.addRequired<DominatorTreeWrapperPass>();
 245     AU.addPreserved<DominatorTreeWrapperPass>();
 246   }
 247 };
 248
 249 } // end anonymous namespace
 250
 251 char ScalarizerLegacyPass::ID = 0;
 252 INITIALIZE_PASS_BEGIN(ScalarizerLegacyPass, "scalarizer",
 253                       "Scalarize vector operations", false, false)
 254 INITIALIZE_PASS_DEPENDENCY(DominatorTreeWrapperPass)
 255 INITIALIZE_PASS_END(ScalarizerLegacyPass, "scalarizer",
 256                     "Scalarize vector operations", false, false)
 257
 258 Scatterer::Scatterer(BasicBlock *bb, BasicBlock::iterator bbi, Value *v,
 259                      ValueVector *cachePtr)
 260   : BB(bb), BBI(bbi), V(v), CachePtr(cachePtr) {
 261   Type *Ty = V->getType();
 262   PtrTy = dyn_cast<PointerType>(Ty);
 263   if (PtrTy)
 264     Ty = PtrTy->getElementType();
 265   Size = cast<FixedVectorType>(Ty)->getNumElements();
 266   if (!CachePtr)
 267     Tmp.resize(Size, nullptr);
 268   else if (CachePtr->empty())
 269     CachePtr->resize(Size, nullptr);
 270   else
 271     assert(Size == CachePtr->size() && "Inconsistent vector sizes");
 272 }
 273
 274 // Return component I, creating a new Value for it if necessary.
 275 Value *Scatterer::operator[](unsigned I) {
 276   ValueVector &CV = (CachePtr ? *CachePtr : Tmp);
 277   // Try to reuse a previous value.
 278   if (CV[I])
 279     return CV[I];
 280   IRBuilder<> Builder(BB, BBI);
 281   if (PtrTy) {
 282     Type *ElTy = cast<VectorType>(PtrTy->getElementType())->getElementType();
 283     if (!CV[0]) {
 284       Type *NewPtrTy = PointerType::get(ElTy, PtrTy->getAddressSpace());
 285       CV[0] = Builder.CreateBitCast(V, NewPtrTy, V->getName() + ".i0");
 286     }
 287     if (I != 0)
 288       CV[I] = Builder.CreateConstGEP1_32(ElTy, CV[0], I,
 289                                          V->getName() + ".i" + Twine(I));
 290   } else {
 291     // Search through a chain of InsertElementInsts looking for element I.
 292     // Record other elements in the cache.  The new V is still suitable
 293     // for all uncached indices.
 294     while (true) {
 295       InsertElementInst *Insert = dyn_cast<InsertElementInst>(V);
 296       if (!Insert)
 297         break;
 298       ConstantInt *Idx = dyn_cast<ConstantInt>(Insert->getOperand(2));
 299       if (!Idx)
 300         break;
 301       unsigned J = Idx->getZExtValue();
 302       V = Insert->getOperand(0);
 303       if (I == J) {
 304         CV[J] = Insert->getOperand(1);
 305         return CV[J];
 306       } else if (!CV[J]) {
 307         // Only cache the first entry we find for each index we're not actively
 308         // searching for. This prevents us from going too far up the chain and
 309         // caching incorrect entries.
 310         CV[J] = Insert->getOperand(1);
 311       }
 312     }
 313     CV[I] = Builder.CreateExtractElement(V, Builder.getInt32(I),
 314                                          V->getName() + ".i" + Twine(I));
 315   }
 316   return CV[I];
 317 }
 318
 319 bool ScalarizerLegacyPass::runOnFunction(Function &F) {
 320   if (skipFunction(F))
 321     return false;
 322
 323   Module &M = *F.getParent();
 324   unsigned ParallelLoopAccessMDKind =
 325       M.getContext().getMDKindID("llvm.mem.parallel_loop_access");
 326   DominatorTree *DT = &getAnalysis<DominatorTreeWrapperPass>().getDomTree();
 327   ScalarizerVisitor Impl(ParallelLoopAccessMDKind, DT);
 328   return Impl.visit(F);
 329 }
 330
 331 FunctionPass *llvm::createScalarizerPass() {
 332   return new ScalarizerLegacyPass();
 333 }
 334
 335 bool ScalarizerVisitor::visit(Function &F) {
 336   assert(Gathered.empty() && Scattered.empty());
 337
 338   // To ensure we replace gathered components correctly we need to do an ordered
 339   // traversal of the basic blocks in the function.
 340   ReversePostOrderTraversal<BasicBlock *> RPOT(&F.getEntryBlock());
 341   for (BasicBlock *BB : RPOT) {
 342     for (BasicBlock::iterator II = BB->begin(), IE = BB->end(); II != IE;) {
 343       Instruction *I = &*II;
 344       bool Done = InstVisitor::visit(I);
 345       ++II;
 346       if (Done && I->getType()->isVoidTy())
 347         I->eraseFromParent();
 348     }
 349   }
 350   return finish();
 351 }
 352
 353 // Return a scattered form of V that can be accessed by Point.  V must be a
 354 // vector or a pointer to a vector.
 355 Scatterer ScalarizerVisitor::scatter(Instruction *Point, Value *V) {
 356   if (Argument *VArg = dyn_cast<Argument>(V)) {
 357     // Put the scattered form of arguments in the entry block,
 358     // so that it can be used everywhere.
 359     Function *F = VArg->getParent();
 360     BasicBlock *BB = &F->getEntryBlock();
 361     return Scatterer(BB, BB->begin(), V, &Scattered[V]);
 362   }
 363   if (Instruction *VOp = dyn_cast<Instruction>(V)) {
 364     // When scalarizing PHI nodes we might try to examine/rewrite InsertElement
 365     // nodes in predecessors. If those predecessors are unreachable from entry,
 366     // then the IR in those blocks could have unexpected properties resulting in
 367     // infinite loops in Scatterer::operator[]. By simply treating values
 368     // originating from instructions in unreachable blocks as undef we do not
 369     // need to analyse them further.
 370     if (!DT->isReachableFromEntry(VOp->getParent()))
 371       return Scatterer(Point->getParent(), Point->getIterator(),
 372                        UndefValue::get(V->getType()));
 373     // Put the scattered form of an instruction directly after the
 374     // instruction.
 375     BasicBlock *BB = VOp->getParent();
 376     return Scatterer(BB, std::next(BasicBlock::iterator(VOp)),
 377                      V, &Scattered[V]);
 378   }
 379   // In the fallback case, just put the scattered before Point and
 380   // keep the result local to Point.
 381   return Scatterer(Point->getParent(), Point->getIterator(), V);
 382 }
 383
 384 // Replace Op with the gathered form of the components in CV.  Defer the
 385 // deletion of Op and creation of the gathered form to the end of the pass,
 386 // so that we can avoid creating the gathered form if all uses of Op are
 387 // replaced with uses of CV.
 388 void ScalarizerVisitor::gather(Instruction *Op, const ValueVector &CV) {
 389   transferMetadataAndIRFlags(Op, CV);
 390
 391   // If we already have a scattered form of Op (created from ExtractElements
 392   // of Op itself), replace them with the new form.
 393   ValueVector &SV = Scattered[Op];
 394   if (!SV.empty()) {
 395     for (unsigned I = 0, E = SV.size(); I != E; ++I) {
 396       Value *V = SV[I];
 397       if (V == nullptr || SV[I] == CV[I])
 398         continue;
 399
 400       Instruction *Old = cast<Instruction>(V);
 401       if (isa<Instruction>(CV[I]))
 402         CV[I]->takeName(Old);
 403       Old->replaceAllUsesWith(CV[I]);
 404       PotentiallyDeadInstrs.emplace_back(Old);
 405     }
 406   }
 407   SV = CV;
 408   Gathered.push_back(GatherList::value_type(Op, &SV));
 409 }
 410
 411 // Return true if it is safe to transfer the given metadata tag from
 412 // vector to scalar instructions.
 413 bool ScalarizerVisitor::canTransferMetadata(unsigned Tag) {
 414   return (Tag == LLVMContext::MD_tbaa
 415           || Tag == LLVMContext::MD_fpmath
 416           || Tag == LLVMContext::MD_tbaa_struct
 417           || Tag == LLVMContext::MD_invariant_load
 418           || Tag == LLVMContext::MD_alias_scope
 419           || Tag == LLVMContext::MD_noalias
 420           || Tag == ParallelLoopAccessMDKind
 421           || Tag == LLVMContext::MD_access_group);
 422 }
 423
 424 // Transfer metadata from Op to the instructions in CV if it is known
 425 // to be safe to do so.
 426 void ScalarizerVisitor::transferMetadataAndIRFlags(Instruction *Op,
 427                                                    const ValueVector &CV) {
 428   SmallVector<std::pair<unsigned, MDNode *>, 4> MDs;
 429   Op->getAllMetadataOtherThanDebugLoc(MDs);
 430   for (unsigned I = 0, E = CV.size(); I != E; ++I) {
 431     if (Instruction *New = dyn_cast<Instruction>(CV[I])) {
 432       for (const auto &MD : MDs)
 433         if (canTransferMetadata(MD.first))
 434           New->setMetadata(MD.first, MD.second);
 435       New->copyIRFlags(Op);
 436       if (Op->getDebugLoc() && !New->getDebugLoc())
 437         New->setDebugLoc(Op->getDebugLoc());
 438     }
 439   }
 440 }
 441
 442 // Try to fill in Layout from Ty, returning true on success.  Alignment is
 443 // the alignment of the vector, or None if the ABI default should be used.
 444 Optional<VectorLayout>
 445 ScalarizerVisitor::getVectorLayout(Type *Ty, Align Alignment,
 446                                    const DataLayout &DL) {
 447   VectorLayout Layout;
 448   // Make sure we're dealing with a vector.
 449   Layout.VecTy = dyn_cast<VectorType>(Ty);
 450   if (!Layout.VecTy)
 451     return None;
 452   // Check that we're dealing with full-byte elements.
 453   Layout.ElemTy = Layout.VecTy->getElementType();
 454   if (!DL.typeSizeEqualsStoreSize(Layout.ElemTy))
 455     return None;
 456   Layout.VecAlign = Alignment;
 457   Layout.ElemSize = DL.getTypeStoreSize(Layout.ElemTy);
 458   return Layout;
 459 }
 460
 461 // Scalarize one-operand instruction I, using Split(Builder, X, Name)
 462 // to create an instruction like I with operand X and name Name.
 463 template<typename Splitter>
 464 bool ScalarizerVisitor::splitUnary(Instruction &I, const Splitter &Split) {
 465   VectorType *VT = dyn_cast<VectorType>(I.getType());
 466   if (!VT)
 467     return false;
 468
 469   unsigned NumElems = cast<FixedVectorType>(VT)->getNumElements();
 470   IRBuilder<> Builder(&I);
 471   Scatterer Op = scatter(&I, I.getOperand(0));
 472   assert(Op.size() == NumElems && "Mismatched unary operation");
 473   ValueVector Res;
 474   Res.resize(NumElems);
 475   for (unsigned Elem = 0; Elem < NumElems; ++Elem)
 476     Res[Elem] = Split(Builder, Op[Elem], I.getName() + ".i" + Twine(Elem));
 477   gather(&I, Res);
 478   return true;
 479 }
 480
 481 // Scalarize two-operand instruction I, using Split(Builder, X, Y, Name)
 482 // to create an instruction like I with operands X and Y and name Name.
 483 template<typename Splitter>
 484 bool ScalarizerVisitor::splitBinary(Instruction &I, const Splitter &Split) {
 485   VectorType *VT = dyn_cast<VectorType>(I.getType());
 486   if (!VT)
 487     return false;
 488
 489   unsigned NumElems = cast<FixedVectorType>(VT)->getNumElements();
 490   IRBuilder<> Builder(&I);
 491   Scatterer VOp0 = scatter(&I, I.getOperand(0));
 492   Scatterer VOp1 = scatter(&I, I.getOperand(1));
 493   assert(VOp0.size() == NumElems && "Mismatched binary operation");
 494   assert(VOp1.size() == NumElems && "Mismatched binary operation");
 495   ValueVector Res;
 496   Res.resize(NumElems);
 497   for (unsigned Elem = 0; Elem < NumElems; ++Elem) {
 498     Value *Op0 = VOp0[Elem];
 499     Value *Op1 = VOp1[Elem];
 500     Res[Elem] = Split(Builder, Op0, Op1, I.getName() + ".i" + Twine(Elem));
 501   }
 502   gather(&I, Res);
 503   return true;
 504 }
 505
 506 static bool isTriviallyScalariable(Intrinsic::ID ID) {
 507   return isTriviallyVectorizable(ID);
 508 }
 509
 510 // All of the current scalarizable intrinsics only have one mangled type.
 511 static Function *getScalarIntrinsicDeclaration(Module *M,
 512                                                Intrinsic::ID ID,
 513                                                ArrayRef<Type*> Tys) {
 514   return Intrinsic::getDeclaration(M, ID, Tys);
 515 }
 516
 517 /// If a call to a vector typed intrinsic function, split into a scalar call per
 518 /// element if possible for the intrinsic.
 519 bool ScalarizerVisitor::splitCall(CallInst &CI) {
 520   VectorType *VT = dyn_cast<VectorType>(CI.getType());
 521   if (!VT)
 522     return false;
 523
 524   Function *F = CI.getCalledFunction();
 525   if (!F)
 526     return false;
 527
 528   Intrinsic::ID ID = F->getIntrinsicID();
 529   if (ID == Intrinsic::not_intrinsic || !isTriviallyScalariable(ID))
 530     return false;
 531
 532   unsigned NumElems = cast<FixedVectorType>(VT)->getNumElements();
 533   unsigned NumArgs = CI.getNumArgOperands();
 534
 535   ValueVector ScalarOperands(NumArgs);
 536   SmallVector<Scatterer, 8> Scattered(NumArgs);
 537
 538   Scattered.resize(NumArgs);
 539
 540   SmallVector<llvm::Type *, 3> Tys;
 541   Tys.push_back(VT->getScalarType());
 542
 543   // Assumes that any vector type has the same number of elements as the return
 544   // vector type, which is true for all current intrinsics.
 545   for (unsigned I = 0; I != NumArgs; ++I) {
 546     Value *OpI = CI.getOperand(I);
 547     if (OpI->getType()->isVectorTy()) {
 548       Scattered[I] = scatter(&CI, OpI);
 549       assert(Scattered[I].size() == NumElems && "mismatched call operands");
 550     } else {
 551       ScalarOperands[I] = OpI;
 552       if (hasVectorInstrinsicOverloadedScalarOpd(ID, I))
 553         Tys.push_back(OpI->getType());
 554     }
 555   }
 556
 557   ValueVector Res(NumElems);
 558   ValueVector ScalarCallOps(NumArgs);
 559
 560   Function *NewIntrin = getScalarIntrinsicDeclaration(F->getParent(), ID, Tys);
 561   IRBuilder<> Builder(&CI);
 562
 563   // Perform actual scalarization, taking care to preserve any scalar operands.
 564   for (unsigned Elem = 0; Elem < NumElems; ++Elem) {
 565     ScalarCallOps.clear();
 566
 567     for (unsigned J = 0; J != NumArgs; ++J) {
 568       if (hasVectorInstrinsicScalarOpd(ID, J))
 569         ScalarCallOps.push_back(ScalarOperands[J]);
 570       else
 571         ScalarCallOps.push_back(Scattered[J][Elem]);
 572     }
 573
 574     Res[Elem] = Builder.CreateCall(NewIntrin, ScalarCallOps,
 575                                    CI.getName() + ".i" + Twine(Elem));
 576   }
 577
 578   gather(&CI, Res);
 579   return true;
 580 }
 581
 582 bool ScalarizerVisitor::visitSelectInst(SelectInst &SI) {
 583   VectorType *VT = dyn_cast<VectorType>(SI.getType());
 584   if (!VT)
 585     return false;
 586
 587   unsigned NumElems = cast<FixedVectorType>(VT)->getNumElements();
 588   IRBuilder<> Builder(&SI);
 589   Scatterer VOp1 = scatter(&SI, SI.getOperand(1));
 590   Scatterer VOp2 = scatter(&SI, SI.getOperand(2));
 591   assert(VOp1.size() == NumElems && "Mismatched select");
 592   assert(VOp2.size() == NumElems && "Mismatched select");
 593   ValueVector Res;
 594   Res.resize(NumElems);
 595
 596   if (SI.getOperand(0)->getType()->isVectorTy()) {
 597     Scatterer VOp0 = scatter(&SI, SI.getOperand(0));
 598     assert(VOp0.size() == NumElems && "Mismatched select");
 599     for (unsigned I = 0; I < NumElems; ++I) {
 600       Value *Op0 = VOp0[I];
 601       Value *Op1 = VOp1[I];
 602       Value *Op2 = VOp2[I];
 603       Res[I] = Builder.CreateSelect(Op0, Op1, Op2,
 604                                     SI.getName() + ".i" + Twine(I));
 605     }
 606   } else {
 607     Value *Op0 = SI.getOperand(0);
 608     for (unsigned I = 0; I < NumElems; ++I) {
 609       Value *Op1 = VOp1[I];
 610       Value *Op2 = VOp2[I];
 611       Res[I] = Builder.CreateSelect(Op0, Op1, Op2,
 612                                     SI.getName() + ".i" + Twine(I));
 613     }
 614   }
 615   gather(&SI, Res);
 616   return true;
 617 }
 618
 619 bool ScalarizerVisitor::visitICmpInst(ICmpInst &ICI) {
 620   return splitBinary(ICI, ICmpSplitter(ICI));
 621 }
 622
 623 bool ScalarizerVisitor::visitFCmpInst(FCmpInst &FCI) {
 624   return splitBinary(FCI, FCmpSplitter(FCI));
 625 }
 626
 627 bool ScalarizerVisitor::visitUnaryOperator(UnaryOperator &UO) {
 628   return splitUnary(UO, UnarySplitter(UO));
 629 }
 630
 631 bool ScalarizerVisitor::visitBinaryOperator(BinaryOperator &BO) {
 632   return splitBinary(BO, BinarySplitter(BO));
 633 }
 634
 635 bool ScalarizerVisitor::visitGetElementPtrInst(GetElementPtrInst &GEPI) {
 636   VectorType *VT = dyn_cast<VectorType>(GEPI.getType());
 637   if (!VT)
 638     return false;
 639
 640   IRBuilder<> Builder(&GEPI);
 641   unsigned NumElems = cast<FixedVectorType>(VT)->getNumElements();
 642   unsigned NumIndices = GEPI.getNumIndices();
 643
 644   // The base pointer might be scalar even if it's a vector GEP. In those cases,
 645   // splat the pointer into a vector value, and scatter that vector.
 646   Value *Op0 = GEPI.getOperand(0);
 647   if (!Op0->getType()->isVectorTy())
 648     Op0 = Builder.CreateVectorSplat(NumElems, Op0);
 649   Scatterer Base = scatter(&GEPI, Op0);
 650
 651   SmallVector<Scatterer, 8> Ops;
 652   Ops.resize(NumIndices);
 653   for (unsigned I = 0; I < NumIndices; ++I) {
 654     Value *Op = GEPI.getOperand(I + 1);
 655
 656     // The indices might be scalars even if it's a vector GEP. In those cases,
 657     // splat the scalar into a vector value, and scatter that vector.
 658     if (!Op->getType()->isVectorTy())
 659       Op = Builder.CreateVectorSplat(NumElems, Op);
 660
 661     Ops[I] = scatter(&GEPI, Op);
 662   }
 663
 664   ValueVector Res;
 665   Res.resize(NumElems);
 666   for (unsigned I = 0; I < NumElems; ++I) {
 667     SmallVector<Value *, 8> Indices;
 668     Indices.resize(NumIndices);
 669     for (unsigned J = 0; J < NumIndices; ++J)
 670       Indices[J] = Ops[J][I];
 671     Res[I] = Builder.CreateGEP(GEPI.getSourceElementType(), Base[I], Indices,
 672                                GEPI.getName() + ".i" + Twine(I));
 673     if (GEPI.isInBounds())
 674       if (GetElementPtrInst *NewGEPI = dyn_cast<GetElementPtrInst>(Res[I]))
 675         NewGEPI->setIsInBounds();
 676   }
 677   gather(&GEPI, Res);
 678   return true;
 679 }
 680
 681 bool ScalarizerVisitor::visitCastInst(CastInst &CI) {
 682   VectorType *VT = dyn_cast<VectorType>(CI.getDestTy());
 683   if (!VT)
 684     return false;
 685
 686   unsigned NumElems = cast<FixedVectorType>(VT)->getNumElements();
 687   IRBuilder<> Builder(&CI);
 688   Scatterer Op0 = scatter(&CI, CI.getOperand(0));
 689   assert(Op0.size() == NumElems && "Mismatched cast");
 690   ValueVector Res;
 691   Res.resize(NumElems);
 692   for (unsigned I = 0; I < NumElems; ++I)
 693     Res[I] = Builder.CreateCast(CI.getOpcode(), Op0[I], VT->getElementType(),
 694                                 CI.getName() + ".i" + Twine(I));
 695   gather(&CI, Res);
 696   return true;
 697 }
 698
 699 bool ScalarizerVisitor::visitBitCastInst(BitCastInst &BCI) {
 700   VectorType *DstVT = dyn_cast<VectorType>(BCI.getDestTy());
 701   VectorType *SrcVT = dyn_cast<VectorType>(BCI.getSrcTy());
 702   if (!DstVT || !SrcVT)
 703     return false;
 704
 705   unsigned DstNumElems = cast<FixedVectorType>(DstVT)->getNumElements();
 706   unsigned SrcNumElems = cast<FixedVectorType>(SrcVT)->getNumElements();
 707   IRBuilder<> Builder(&BCI);
 708   Scatterer Op0 = scatter(&BCI, BCI.getOperand(0));
 709   ValueVector Res;
 710   Res.resize(DstNumElems);
 711
 712   if (DstNumElems == SrcNumElems) {
 713     for (unsigned I = 0; I < DstNumElems; ++I)
 714       Res[I] = Builder.CreateBitCast(Op0[I], DstVT->getElementType(),
 715                                      BCI.getName() + ".i" + Twine(I));
 716   } else if (DstNumElems > SrcNumElems) {
 717     // <M x t1> -> <N*M x t2>.  Convert each t1 to <N x t2> and copy the
 718     // individual elements to the destination.
 719     unsigned FanOut = DstNumElems / SrcNumElems;
 720     auto *MidTy = FixedVectorType::get(DstVT->getElementType(), FanOut);
 721     unsigned ResI = 0;
 722     for (unsigned Op0I = 0; Op0I < SrcNumElems; ++Op0I) {
 723       Value *V = Op0[Op0I];
 724       Instruction *VI;
 725       // Look through any existing bitcasts before converting to <N x t2>.
 726       // In the best case, the resulting conversion might be a no-op.
 727       while ((VI = dyn_cast<Instruction>(V)) &&
 728              VI->getOpcode() == Instruction::BitCast)
 729         V = VI->getOperand(0);
 730       V = Builder.CreateBitCast(V, MidTy, V->getName() + ".cast");
 731       Scatterer Mid = scatter(&BCI, V);
 732       for (unsigned MidI = 0; MidI < FanOut; ++MidI)
 733         Res[ResI++] = Mid[MidI];
 734     }
 735   } else {
 736     // <N*M x t1> -> <M x t2>.  Convert each group of <N x t1> into a t2.
 737     unsigned FanIn = SrcNumElems / DstNumElems;
 738     auto *MidTy = FixedVectorType::get(SrcVT->getElementType(), FanIn);
 739     unsigned Op0I = 0;
 740     for (unsigned ResI = 0; ResI < DstNumElems; ++ResI) {
 741       Value *V = PoisonValue::get(MidTy);
 742       for (unsigned MidI = 0; MidI < FanIn; ++MidI)
 743         V = Builder.CreateInsertElement(V, Op0[Op0I++], Builder.getInt32(MidI),
 744                                         BCI.getName() + ".i" + Twine(ResI)
 745                                         + ".upto" + Twine(MidI));
 746       Res[ResI] = Builder.CreateBitCast(V, DstVT->getElementType(),
 747                                         BCI.getName() + ".i" + Twine(ResI));
 748     }
 749   }
 750   gather(&BCI, Res);
 751   return true;
 752 }
 753
 754 bool ScalarizerVisitor::visitInsertElementInst(InsertElementInst &IEI) {
 755   VectorType *VT = dyn_cast<VectorType>(IEI.getType());
 756   if (!VT)
 757     return false;
 758
 759   unsigned NumElems = cast<FixedVectorType>(VT)->getNumElements();
 760   IRBuilder<> Builder(&IEI);
 761   Scatterer Op0 = scatter(&IEI, IEI.getOperand(0));
 762   Value *NewElt = IEI.getOperand(1);
 763   Value *InsIdx = IEI.getOperand(2);
 764
 765   ValueVector Res;
 766   Res.resize(NumElems);
 767
 768   if (auto *CI = dyn_cast<ConstantInt>(InsIdx)) {
 769     for (unsigned I = 0; I < NumElems; ++I)
 770       Res[I] = CI->getValue().getZExtValue() == I ? NewElt : Op0[I];
 771   } else {
 772     if (!ScalarizeVariableInsertExtract)
 773       return false;
 774
 775     for (unsigned I = 0; I < NumElems; ++I) {
 776       Value *ShouldReplace =
 777           Builder.CreateICmpEQ(InsIdx, ConstantInt::get(InsIdx->getType(), I),
 778                                InsIdx->getName() + ".is." + Twine(I));
 779       Value *OldElt = Op0[I];
 780       Res[I] = Builder.CreateSelect(ShouldReplace, NewElt, OldElt,
 781                                     IEI.getName() + ".i" + Twine(I));
 782     }
 783   }
 784
 785   gather(&IEI, Res);
 786   return true;
 787 }
 788
 789 bool ScalarizerVisitor::visitExtractElementInst(ExtractElementInst &EEI) {
 790   VectorType *VT = dyn_cast<VectorType>(EEI.getOperand(0)->getType());
 791   if (!VT)
 792     return false;
 793
 794   unsigned NumSrcElems = cast<FixedVectorType>(VT)->getNumElements();
 795   IRBuilder<> Builder(&EEI);
 796   Scatterer Op0 = scatter(&EEI, EEI.getOperand(0));
 797   Value *ExtIdx = EEI.getOperand(1);
 798
 799   if (auto *CI = dyn_cast<ConstantInt>(ExtIdx)) {
 800     Value *Res = Op0[CI->getValue().getZExtValue()];
 801     gather(&EEI, {Res});
 802     return true;
 803   }
 804
 805   if (!ScalarizeVariableInsertExtract)
 806     return false;
 807
 808   Value *Res = UndefValue::get(VT->getElementType());
 809   for (unsigned I = 0; I < NumSrcElems; ++I) {
 810     Value *ShouldExtract =
 811         Builder.CreateICmpEQ(ExtIdx, ConstantInt::get(ExtIdx->getType(), I),
 812                              ExtIdx->getName() + ".is." + Twine(I));
 813     Value *Elt = Op0[I];
 814     Res = Builder.CreateSelect(ShouldExtract, Elt, Res,
 815                                EEI.getName() + ".upto" + Twine(I));
 816   }
 817   gather(&EEI, {Res});
 818   return true;
 819 }
 820
 821 bool ScalarizerVisitor::visitShuffleVectorInst(ShuffleVectorInst &SVI) {
 822   VectorType *VT = dyn_cast<VectorType>(SVI.getType());
 823   if (!VT)
 824     return false;
 825
 826   unsigned NumElems = cast<FixedVectorType>(VT)->getNumElements();
 827   Scatterer Op0 = scatter(&SVI, SVI.getOperand(0));
 828   Scatterer Op1 = scatter(&SVI, SVI.getOperand(1));
 829   ValueVector Res;
 830   Res.resize(NumElems);
 831
 832   for (unsigned I = 0; I < NumElems; ++I) {
 833     int Selector = SVI.getMaskValue(I);
 834     if (Selector < 0)
 835       Res[I] = UndefValue::get(VT->getElementType());
 836     else if (unsigned(Selector) < Op0.size())
 837       Res[I] = Op0[Selector];
 838     else
 839       Res[I] = Op1[Selector - Op0.size()];
 840   }
 841   gather(&SVI, Res);
 842   return true;
 843 }
 844
 845 bool ScalarizerVisitor::visitPHINode(PHINode &PHI) {
 846   VectorType *VT = dyn_cast<VectorType>(PHI.getType());
 847   if (!VT)
 848     return false;
 849
 850   unsigned NumElems = cast<FixedVectorType>(VT)->getNumElements();
 851   IRBuilder<> Builder(&PHI);
 852   ValueVector Res;
 853   Res.resize(NumElems);
 854
 855   unsigned NumOps = PHI.getNumOperands();
 856   for (unsigned I = 0; I < NumElems; ++I)
 857     Res[I] = Builder.CreatePHI(VT->getElementType(), NumOps,
 858                                PHI.getName() + ".i" + Twine(I));
 859
 860   for (unsigned I = 0; I < NumOps; ++I) {
 861     Scatterer Op = scatter(&PHI, PHI.getIncomingValue(I));
 862     BasicBlock *IncomingBlock = PHI.getIncomingBlock(I);
 863     for (unsigned J = 0; J < NumElems; ++J)
 864       cast<PHINode>(Res[J])->addIncoming(Op[J], IncomingBlock);
 865   }
 866   gather(&PHI, Res);
 867   return true;
 868 }
 869
 870 bool ScalarizerVisitor::visitLoadInst(LoadInst &LI) {
 871   if (!ScalarizeLoadStore)
 872     return false;
 873   if (!LI.isSimple())
 874     return false;
 875
 876   Optional<VectorLayout> Layout = getVectorLayout(
 877       LI.getType(), LI.getAlign(), LI.getModule()->getDataLayout());
 878   if (!Layout)
 879     return false;
 880
 881   unsigned NumElems = cast<FixedVectorType>(Layout->VecTy)->getNumElements();
 882   IRBuilder<> Builder(&LI);
 883   Scatterer Ptr = scatter(&LI, LI.getPointerOperand());
 884   ValueVector Res;
 885   Res.resize(NumElems);
 886
 887   for (unsigned I = 0; I < NumElems; ++I)
 888     Res[I] = Builder.CreateAlignedLoad(Layout->VecTy->getElementType(), Ptr[I],
 889                                        Align(Layout->getElemAlign(I)),
 890                                        LI.getName() + ".i" + Twine(I));
 891   gather(&LI, Res);
 892   return true;
 893 }
 894
 895 bool ScalarizerVisitor::visitStoreInst(StoreInst &SI) {
 896   if (!ScalarizeLoadStore)
 897     return false;
 898   if (!SI.isSimple())
 899     return false;
 900
 901   Value *FullValue = SI.getValueOperand();
 902   Optional<VectorLayout> Layout = getVectorLayout(
 903       FullValue->getType(), SI.getAlign(), SI.getModule()->getDataLayout());
 904   if (!Layout)
 905     return false;
 906
 907   unsigned NumElems = cast<FixedVectorType>(Layout->VecTy)->getNumElements();
 908   IRBuilder<> Builder(&SI);
 909   Scatterer VPtr = scatter(&SI, SI.getPointerOperand());
 910   Scatterer VVal = scatter(&SI, FullValue);
 911
 912   ValueVector Stores;
 913   Stores.resize(NumElems);
 914   for (unsigned I = 0; I < NumElems; ++I) {
 915     Value *Val = VVal[I];
 916     Value *Ptr = VPtr[I];
 917     Stores[I] = Builder.CreateAlignedStore(Val, Ptr, Layout->getElemAlign(I));
 918   }
 919   transferMetadataAndIRFlags(&SI, Stores);
 920   return true;
 921 }
 922
 923 bool ScalarizerVisitor::visitCallInst(CallInst &CI) {
 924   return splitCall(CI);
 925 }
 926
 927 // Delete the instructions that we scalarized.  If a full vector result
 928 // is still needed, recreate it using InsertElements.
 929 bool ScalarizerVisitor::finish() {
 930   // The presence of data in Gathered or Scattered indicates changes
 931   // made to the Function.
 932   if (Gathered.empty() && Scattered.empty())
 933     return false;
 934   for (const auto &GMI : Gathered) {
 935     Instruction *Op = GMI.first;
 936     ValueVector &CV = *GMI.second;
 937     if (!Op->use_empty()) {
 938       // The value is still needed, so recreate it using a series of
 939       // InsertElements.
 940       Value *Res = PoisonValue::get(Op->getType());
 941       if (auto *Ty = dyn_cast<VectorType>(Op->getType())) {
 942         BasicBlock *BB = Op->getParent();
 943         unsigned Count = cast<FixedVectorType>(Ty)->getNumElements();
 944         IRBuilder<> Builder(Op);
 945         if (isa<PHINode>(Op))
 946           Builder.SetInsertPoint(BB, BB->getFirstInsertionPt());
 947         for (unsigned I = 0; I < Count; ++I)
 948           Res = Builder.CreateInsertElement(Res, CV[I], Builder.getInt32(I),
 949                                             Op->getName() + ".upto" + Twine(I));
 950         Res->takeName(Op);
 951       } else {
 952         assert(CV.size() == 1 && Op->getType() == CV[0]->getType());
 953         Res = CV[0];
 954         if (Op == Res)
 955           continue;
 956       }
 957       Op->replaceAllUsesWith(Res);
 958     }
 959     PotentiallyDeadInstrs.emplace_back(Op);
 960   }
 961   Gathered.clear();
 962   Scattered.clear();
 963
 964   RecursivelyDeleteTriviallyDeadInstructionsPermissive(PotentiallyDeadInstrs);
 965
 966   return true;
 967 }
 968
 969 PreservedAnalyses ScalarizerPass::run(Function &F, FunctionAnalysisManager &AM) {
 970   Module &M = *F.getParent();
 971   unsigned ParallelLoopAccessMDKind =
 972       M.getContext().getMDKindID("llvm.mem.parallel_loop_access");
 973   DominatorTree *DT = &AM.getResult<DominatorTreeAnalysis>(F);
 974   ScalarizerVisitor Impl(ParallelLoopAccessMDKind, DT);
 975   bool Changed = Impl.visit(F);
 976   PreservedAnalyses PA;
 977   PA.preserve<DominatorTreeAnalysis>();
 978   return Changed ? PA : PreservedAnalyses::all();
 979 }