Merge llvm, clang, lld and lldb release_40 branch r292009. Also update

build glue.

contrib/llvm/LICENSE.TXT

@@ -4,7 +4,7 @@ LLVM Release License
 University of Illinois/NCSA
 Open Source License
 
-Copyright (c) 2003-2016 University of Illinois at Urbana-Champaign.
+Copyright (c) 2003-2017 University of Illinois at Urbana-Champaign.
 All rights reserved.
 
 Developed by:

contrib/llvm/include/llvm/ADT/PointerSumType.h

@@ -94,7 +94,7 @@ template <typename TagT, typename... MemberTs> class PointerSumType {
     return HelperT::template Lookup<N>::TraitsT::getFromVoidPointer(getImpl());
   }
 
-  operator bool() const { return Value & HelperT::PointerMask; }
+  explicit operator bool() const { return Value & HelperT::PointerMask; }
   bool operator==(const PointerSumType &R) const { return Value == R.Value; }
   bool operator!=(const PointerSumType &R) const { return Value != R.Value; }
   bool operator<(const PointerSumType &R) const { return Value < R.Value; }

contrib/llvm/include/llvm/ADT/iterator.h

@@ -33,6 +33,32 @@ namespace llvm {
 /// Another abstraction that this doesn't provide is implementing increment in
 /// terms of addition of one. These aren't equivalent for all iterator
 /// categories, and respecting that adds a lot of complexity for little gain.
+///
+/// Classes wishing to use `iterator_facade_base` should implement the following
+/// methods:
+///
+/// Forward Iterators:
+///   (All of the following methods)
+///   - DerivedT &operator=(const DerivedT &R);
+///   - bool operator==(const DerivedT &R) const;
+///   - const T &operator*() const;
+///   - T &operator*();
+///   - DerivedT &operator++();
+///
+/// Bidirectional Iterators:
+///   (All methods of forward iterators, plus the following)
+///   - DerivedT &operator--();
+///
+/// Random-access Iterators:
+///   (All methods of bidirectional iterators excluding the following)
+///   - DerivedT &operator++();
+///   - DerivedT &operator--();
+///   (and plus the following)
+///   - bool operator<(const DerivedT &RHS) const;
+///   - DifferenceTypeT operator-(const DerivedT &R) const;
+///   - DerivedT &operator+=(DifferenceTypeT N);
+///   - DerivedT &operator-=(DifferenceTypeT N);
+///
 template <typename DerivedT, typename IteratorCategoryT, typename T,
           typename DifferenceTypeT = std::ptrdiff_t, typename PointerT = T *,
           typename ReferenceT = T &>

contrib/llvm/include/llvm/Analysis/AssumptionCache.h

@@ -46,6 +46,30 @@ class AssumptionCache {
   /// intrinsic.
   SmallVector<WeakVH, 4> AssumeHandles;
 
+  class AffectedValueCallbackVH final : public CallbackVH {
+    AssumptionCache *AC;
+    void deleted() override;
+    void allUsesReplacedWith(Value *) override;
+
+  public:
+    using DMI = DenseMapInfo<Value *>;
+
+    AffectedValueCallbackVH(Value *V, AssumptionCache *AC = nullptr)
+        : CallbackVH(V), AC(AC) {}
+  };
+
+  friend AffectedValueCallbackVH;
+
+  /// \brief A map of values about which an assumption might be providing
+  /// information to the relevant set of assumptions.
+  using AffectedValuesMap =
+    DenseMap<AffectedValueCallbackVH, SmallVector<WeakVH, 1>,
+             AffectedValueCallbackVH::DMI>;
+  AffectedValuesMap AffectedValues;
+
+  /// Get the vector of assumptions which affect a value from the cache.
+  SmallVector<WeakVH, 1> &getAffectedValues(Value *V);
+
   /// \brief Flag tracking whether we have scanned the function yet.
   ///
   /// We want to be as lazy about this as possible, and so we scan the function
@@ -66,11 +90,16 @@ class AssumptionCache {
   /// not already be in the cache.
   void registerAssumption(CallInst *CI);
 
+  /// \brief Update the cache of values being affected by this assumption (i.e.
+  /// the values about which this assumption provides information).
+  void updateAffectedValues(CallInst *CI);
+
   /// \brief Clear the cache of @llvm.assume intrinsics for a function.
   ///
   /// It will be re-scanned the next time it is requested.
   void clear() {
     AssumeHandles.clear();
+    AffectedValues.clear();
     Scanned = false;
   }
 
@@ -87,6 +116,18 @@ class AssumptionCache {
       scanFunction();
     return AssumeHandles;
   }
+
+  /// \brief Access the list of assumptions which affect this value.
+  MutableArrayRef<WeakVH> assumptionsFor(const Value *V) {
+    if (!Scanned)
+      scanFunction();
+
+    auto AVI = AffectedValues.find_as(const_cast<Value *>(V));
+    if (AVI == AffectedValues.end())
+      return MutableArrayRef<WeakVH>();
+
+    return AVI->second;
+  }
 };
 
 /// \brief A function analysis which provides an \c AssumptionCache.

contrib/llvm/include/llvm/Analysis/IVUsers.h

@@ -15,8 +15,8 @@
 #ifndef LLVM_ANALYSIS_IVUSERS_H
 #define LLVM_ANALYSIS_IVUSERS_H
 
+#include "llvm/Analysis/LoopAnalysisManager.h"
 #include "llvm/Analysis/LoopPass.h"
-#include "llvm/Analysis/LoopPassManager.h"
 #include "llvm/Analysis/ScalarEvolutionNormalization.h"
 #include "llvm/IR/ValueHandle.h"
 
@@ -193,17 +193,10 @@ class IVUsersAnalysis : public AnalysisInfoMixin<IVUsersAnalysis> {
 public:
   typedef IVUsers Result;
 
-  IVUsers run(Loop &L, LoopAnalysisManager &AM);
+  IVUsers run(Loop &L, LoopAnalysisManager &AM,
+              LoopStandardAnalysisResults &AR);
 };
 
-/// Printer pass for the \c IVUsers for a loop.
-class IVUsersPrinterPass : public PassInfoMixin<IVUsersPrinterPass> {
-  raw_ostream &OS;
-
-public:
-  explicit IVUsersPrinterPass(raw_ostream &OS) : OS(OS) {}
-  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM);
-};
 }
 
 #endif

contrib/llvm/include/llvm/Analysis/LazyCallGraph.h

@@ -148,7 +148,7 @@ class LazyCallGraph {
     ///
     /// This happens when an edge has been deleted. We leave the edge objects
     /// around but clear them.
-    operator bool() const;
+    explicit operator bool() const;
 
     /// Returnss the \c Kind of the edge.
     Kind getKind() const;

contrib/llvm/include/llvm/Analysis/LoopAccessAnalysis.h

@@ -20,7 +20,7 @@
 #include "llvm/ADT/SetVector.h"
 #include "llvm/Analysis/AliasAnalysis.h"
 #include "llvm/Analysis/AliasSetTracker.h"
-#include "llvm/Analysis/LoopPassManager.h"
+#include "llvm/Analysis/LoopAnalysisManager.h"
 #include "llvm/Analysis/ScalarEvolutionExpressions.h"
 #include "llvm/IR/DiagnosticInfo.h"
 #include "llvm/IR/ValueHandle.h"
@@ -753,18 +753,8 @@ class LoopAccessAnalysis
 
 public:
   typedef LoopAccessInfo Result;
-  Result run(Loop &, LoopAnalysisManager &);
-  static StringRef name() { return "LoopAccessAnalysis"; }
-};
 
-/// \brief Printer pass for the \c LoopAccessInfo results.
-class LoopAccessInfoPrinterPass
-    : public PassInfoMixin<LoopAccessInfoPrinterPass> {
-  raw_ostream &OS;
-
-public:
-  explicit LoopAccessInfoPrinterPass(raw_ostream &OS) : OS(OS) {}
-  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM);
+  Result run(Loop &L, LoopAnalysisManager &AM, LoopStandardAnalysisResults &AR);
 };
 
 inline Instruction *MemoryDepChecker::Dependence::getSource(

contrib/llvm/include/llvm/Analysis/LoopAnalysisManager.h

@@ -0,0 +1,155 @@
+//===- LoopAnalysisManager.h - Loop analysis management ---------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+/// \file
+///
+/// This header provides classes for managing per-loop analyses. These are
+/// typically used as part of a loop pass pipeline over the loop nests of
+/// a function.
+///
+/// Loop analyses are allowed to make some simplifying assumptions:
+/// 1) Loops are, where possible, in simplified form.
+/// 2) Loops are *always* in LCSSA form.
+/// 3) A collection of analysis results are available:
+///    - LoopInfo
+///    - DominatorTree
+///    - ScalarEvolution
+///    - AAManager
+///
+/// The primary mechanism to provide these invariants is the loop pass manager,
+/// but they can also be manually provided in order to reason about a loop from
+/// outside of a dedicated pass manager.
+///
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_ANALYSIS_LOOPANALYSISMANAGER_H
+#define LLVM_ANALYSIS_LOOPANALYSISMANAGER_H
+
+#include "llvm/ADT/PostOrderIterator.h"
+#include "llvm/ADT/PriorityWorklist.h"
+#include "llvm/ADT/STLExtras.h"
+#include "llvm/Analysis/AliasAnalysis.h"
+#include "llvm/Analysis/BasicAliasAnalysis.h"
+#include "llvm/Analysis/GlobalsModRef.h"
+#include "llvm/Analysis/LoopInfo.h"
+#include "llvm/Analysis/ScalarEvolution.h"
+#include "llvm/Analysis/ScalarEvolutionAliasAnalysis.h"
+#include "llvm/Analysis/TargetLibraryInfo.h"
+#include "llvm/Analysis/TargetTransformInfo.h"
+#include "llvm/IR/Dominators.h"
+#include "llvm/IR/PassManager.h"
+
+namespace llvm {
+
+/// The adaptor from a function pass to a loop pass computes these analyses and
+/// makes them available to the loop passes "for free". Each loop pass is
+/// expected expected to update these analyses if necessary to ensure they're
+/// valid after it runs.
+struct LoopStandardAnalysisResults {
+  AAResults &AA;
+  AssumptionCache ∾
+  DominatorTree &DT;
+  LoopInfo &LI;
+  ScalarEvolution &SE;
+  TargetLibraryInfo &TLI;
+  TargetTransformInfo &TTI;
+};
+
+/// Extern template declaration for the analysis set for this IR unit.
+extern template class AllAnalysesOn<Loop>;
+
+extern template class AnalysisManager<Loop, LoopStandardAnalysisResults &>;
+/// \brief The loop analysis manager.
+///
+/// See the documentation for the AnalysisManager template for detail
+/// documentation. This typedef serves as a convenient way to refer to this
+/// construct in the adaptors and proxies used to integrate this into the larger
+/// pass manager infrastructure.
+typedef AnalysisManager<Loop, LoopStandardAnalysisResults &>
+    LoopAnalysisManager;
+
+/// A proxy from a \c LoopAnalysisManager to a \c Function.
+typedef InnerAnalysisManagerProxy<LoopAnalysisManager, Function>
+    LoopAnalysisManagerFunctionProxy;
+
+/// A specialized result for the \c LoopAnalysisManagerFunctionProxy which
+/// retains a \c LoopInfo reference.
+///
+/// This allows it to collect loop objects for which analysis results may be
+/// cached in the \c LoopAnalysisManager.
+template <> class LoopAnalysisManagerFunctionProxy::Result {
+public:
+  explicit Result(LoopAnalysisManager &InnerAM, LoopInfo &LI)
+      : InnerAM(&InnerAM), LI(&LI) {}
+  Result(Result &&Arg) : InnerAM(std::move(Arg.InnerAM)), LI(Arg.LI) {
+    // We have to null out the analysis manager in the moved-from state
+    // because we are taking ownership of the responsibilty to clear the
+    // analysis state.
+    Arg.InnerAM = nullptr;
+  }
+  Result &operator=(Result &&RHS) {
+    InnerAM = RHS.InnerAM;
+    LI = RHS.LI;
+    // We have to null out the analysis manager in the moved-from state
+    // because we are taking ownership of the responsibilty to clear the
+    // analysis state.
+    RHS.InnerAM = nullptr;
+    return *this;
+  }
+  ~Result() {
+    // InnerAM is cleared in a moved from state where there is nothing to do.
+    if (!InnerAM)
+      return;
+
+    // Clear out the analysis manager if we're being destroyed -- it means we
+    // didn't even see an invalidate call when we got invalidated.
+    InnerAM->clear();
+  }
+
+  /// Accessor for the analysis manager.
+  LoopAnalysisManager &getManager() { return *InnerAM; }
+
+  /// Handler for invalidation of the proxy for a particular function.
+  ///
+  /// If the proxy, \c LoopInfo, and associated analyses are preserved, this
+  /// will merely forward the invalidation event to any cached loop analysis
+  /// results for loops within this function.
+  ///
+  /// If the necessary loop infrastructure is not preserved, this will forcibly
+  /// clear all of the cached analysis results that are keyed on the \c
+  /// LoopInfo for this function.
+  bool invalidate(Function &F, const PreservedAnalyses &PA,
+                  FunctionAnalysisManager::Invalidator &Inv);
+
+private:
+  LoopAnalysisManager *InnerAM;
+  LoopInfo *LI;
+};
+
+/// Provide a specialized run method for the \c LoopAnalysisManagerFunctionProxy
+/// so it can pass the \c LoopInfo to the result.
+template <>
+LoopAnalysisManagerFunctionProxy::Result
+LoopAnalysisManagerFunctionProxy::run(Function &F, FunctionAnalysisManager &AM);
+
+// Ensure the \c LoopAnalysisManagerFunctionProxy is provided as an extern
+// template.
+extern template class InnerAnalysisManagerProxy<LoopAnalysisManager, Function>;
+
+extern template class OuterAnalysisManagerProxy<FunctionAnalysisManager, Loop,
+                                                LoopStandardAnalysisResults &>;
+/// A proxy from a \c FunctionAnalysisManager to a \c Loop.
+typedef OuterAnalysisManagerProxy<FunctionAnalysisManager, Loop,
+                                  LoopStandardAnalysisResults &>
+    FunctionAnalysisManagerLoopProxy;
+
+/// Returns the minimum set of Analyses that all loop passes must preserve.
+PreservedAnalyses getLoopPassPreservedAnalyses();
+}
+
+#endif // LLVM_ANALYSIS_LOOPANALYSISMANAGER_H

contrib/llvm/include/llvm/Analysis/LoopInfo.h

@@ -853,17 +853,8 @@ class LoopInfoWrapperPass : public FunctionPass {
   void getAnalysisUsage(AnalysisUsage &AU) const override;
 };
 
-/// \brief Pass for printing a loop's contents as LLVM's text IR assembly.
-class PrintLoopPass : public PassInfoMixin<PrintLoopPass> {
-  raw_ostream &OS;
-  std::string Banner;
-
-public:
-  PrintLoopPass();
-  PrintLoopPass(raw_ostream &OS, const std::string &Banner = "");
-
-  PreservedAnalyses run(Loop &L, AnalysisManager<Loop> &);
-};
+/// Function to print a loop's contents as LLVM's text IR assembly.
+void printLoop(Loop &L, raw_ostream &OS, const std::string &Banner = "");
 
 } // End llvm namespace
 

contrib/llvm/include/llvm/Analysis/LoopPassManager.h

@@ -1,149 +0,0 @@
-//===- LoopPassManager.h - Loop pass management -----------------*- C++ -*-===//
-//
-//                     The LLVM Compiler Infrastructure
-//
-// This file is distributed under the University of Illinois Open Source
-// License. See LICENSE.TXT for details.
-//
-//===----------------------------------------------------------------------===//
-/// \file
-///
-/// This header provides classes for managing passes over loops in LLVM IR.
-///
-//===----------------------------------------------------------------------===//
-
-#ifndef LLVM_ANALYSIS_LOOPPASSMANAGER_H
-#define LLVM_ANALYSIS_LOOPPASSMANAGER_H
-
-#include "llvm/ADT/STLExtras.h"
-#include "llvm/Analysis/AliasAnalysis.h"
-#include "llvm/Analysis/LoopInfo.h"
-#include "llvm/Analysis/ScalarEvolution.h"
-#include "llvm/Analysis/TargetLibraryInfo.h"
-#include "llvm/IR/Dominators.h"
-#include "llvm/IR/PassManager.h"
-
-namespace llvm {
-
-extern template class PassManager<Loop>;
-/// \brief The loop pass manager.
-///
-/// See the documentation for the PassManager template for details. It runs a
-/// sequency of loop passes over each loop that the manager is run over. This
-/// typedef serves as a convenient way to refer to this construct.
-typedef PassManager<Loop> LoopPassManager;
-
-extern template class AnalysisManager<Loop>;
-/// \brief The loop analysis manager.
-///
-/// See the documentation for the AnalysisManager template for detail
-/// documentation. This typedef serves as a convenient way to refer to this
-/// construct in the adaptors and proxies used to integrate this into the larger
-/// pass manager infrastructure.
-typedef AnalysisManager<Loop> LoopAnalysisManager;
-
-/// A proxy from a \c LoopAnalysisManager to a \c Function.
-typedef InnerAnalysisManagerProxy<LoopAnalysisManager, Function>
-    LoopAnalysisManagerFunctionProxy;
-
-/// Specialization of the invalidate method for the \c
-/// LoopAnalysisManagerFunctionProxy's result.
-template <>
-bool LoopAnalysisManagerFunctionProxy::Result::invalidate(
-    Function &F, const PreservedAnalyses &PA,
-    FunctionAnalysisManager::Invalidator &Inv);
-
-// Ensure the \c LoopAnalysisManagerFunctionProxy is provided as an extern
-// template.
-extern template class InnerAnalysisManagerProxy<LoopAnalysisManager, Function>;
-
-extern template class OuterAnalysisManagerProxy<FunctionAnalysisManager, Loop>;
-/// A proxy from a \c FunctionAnalysisManager to a \c Loop.
-typedef OuterAnalysisManagerProxy<FunctionAnalysisManager, Loop>
-    FunctionAnalysisManagerLoopProxy;
-
-/// Returns the minimum set of Analyses that all loop passes must preserve.
-PreservedAnalyses getLoopPassPreservedAnalyses();
-
-/// \brief Adaptor that maps from a function to its loops.
-///
-/// Designed to allow composition of a LoopPass(Manager) and a
-/// FunctionPassManager. Note that if this pass is constructed with a \c
-/// FunctionAnalysisManager it will run the \c LoopAnalysisManagerFunctionProxy
-/// analysis prior to running the loop passes over the function to enable a \c
-/// LoopAnalysisManager to be used within this run safely.
-template <typename LoopPassT>
-class FunctionToLoopPassAdaptor
-    : public PassInfoMixin<FunctionToLoopPassAdaptor<LoopPassT>> {
-public:
-  explicit FunctionToLoopPassAdaptor(LoopPassT Pass)
-      : Pass(std::move(Pass)) {}
-
-  /// \brief Runs the loop passes across every loop in the function.
-  PreservedAnalyses run(Function &F, FunctionAnalysisManager &AM) {
-    // Setup the loop analysis manager from its proxy.
-    LoopAnalysisManager &LAM =
-        AM.getResult<LoopAnalysisManagerFunctionProxy>(F).getManager();
-    // Get the loop structure for this function
-    LoopInfo &LI = AM.getResult<LoopAnalysis>(F);
-
-    // Also precompute all of the function analyses used by loop passes.
-    // FIXME: These should be handed into the loop passes when the loop pass
-    // management layer is reworked to follow the design of CGSCC.
-    (void)AM.getResult<AAManager>(F);
-    (void)AM.getResult<DominatorTreeAnalysis>(F);
-    (void)AM.getResult<ScalarEvolutionAnalysis>(F);
-    (void)AM.getResult<TargetLibraryAnalysis>(F);
-
-    PreservedAnalyses PA = PreservedAnalyses::all();
-
-    // We want to visit the loops in reverse post-order. We'll build the stack
-    // of loops to visit in Loops by first walking the loops in pre-order.
-    SmallVector<Loop *, 2> Loops;
-    SmallVector<Loop *, 2> WorkList(LI.begin(), LI.end());
-    while (!WorkList.empty()) {
-      Loop *L = WorkList.pop_back_val();
-      WorkList.insert(WorkList.end(), L->begin(), L->end());
-      Loops.push_back(L);
-    }
-
-    // Now pop each element off of the stack to visit the loops in reverse
-    // post-order.
-    for (auto *L : reverse(Loops)) {
-      PreservedAnalyses PassPA = Pass.run(*L, LAM);
-      // FIXME: We should verify the set of analyses relevant to Loop passes
-      // are preserved.
-
-      // We know that the loop pass couldn't have invalidated any other loop's
-      // analyses (that's the contract of a loop pass), so directly handle the
-      // loop analysis manager's invalidation here.
-      LAM.invalidate(*L, PassPA);
-
-      // Then intersect the preserved set so that invalidation of module
-      // analyses will eventually occur when the module pass completes.
-      PA.intersect(std::move(PassPA));
-    }
-
-    // By definition we preserve the proxy. We also preserve all analyses on
-    // Loops. This precludes *any* invalidation of loop analyses by the proxy,
-    // but that's OK because we've taken care to invalidate analyses in the
-    // loop analysis manager incrementally above.
-    PA.preserveSet<AllAnalysesOn<Loop>>();
-    PA.preserve<LoopAnalysisManagerFunctionProxy>();
-    return PA;
-  }
-
-private:
-  LoopPassT Pass;
-};
-
-/// \brief A function to deduce a loop pass type and wrap it in the templated
-/// adaptor.
-template <typename LoopPassT>
-FunctionToLoopPassAdaptor<LoopPassT>
-createFunctionToLoopPassAdaptor(LoopPassT Pass) {
-  return FunctionToLoopPassAdaptor<LoopPassT>(std::move(Pass));
-}
-}
-
-#endif // LLVM_ANALYSIS_LOOPPASSMANAGER_H

contrib/llvm/include/llvm/Analysis/MemoryDependenceAnalysis.h

@@ -302,6 +302,10 @@ class MemoryDependenceResults {
     NonLocalPointerInfo() : Size(MemoryLocation::UnknownSize) {}
   };
 
+  /// Cache storing single nonlocal def for the instruction.
+  /// It is set when nonlocal def would be found in function returning only
+  /// local dependencies.
+  DenseMap<Instruction *, NonLocalDepResult> NonLocalDefsCache;
   /// This map stores the cached results of doing a pointer lookup at the
   /// bottom of a block.
   ///
@@ -441,9 +445,9 @@ class MemoryDependenceResults {
   /// This analysis looks for other loads and stores with invariant.group
   /// metadata and the same pointer operand. Returns Unknown if it does not
   /// find anything, and Def if it can be assumed that 2 instructions load or
-  /// store the same value.
-  /// FIXME: This analysis works only on single block because of restrictions
-  /// at the call site.
+  /// store the same value and NonLocal which indicate that non-local Def was
+  /// found, which can be retrieved by calling getNonLocalPointerDependency
+  /// with the same queried instruction.
   MemDepResult getInvariantGroupPointerDependency(LoadInst *LI, BasicBlock *BB);
 
   /// Looks at a memory location for a load (specified by MemLocBase, Offs, and

contrib/llvm/include/llvm/Analysis/TargetTransformInfo.h

@@ -55,6 +55,11 @@ struct MemIntrinsicInfo {
   // Same Id is set by the target for corresponding load/store intrinsics.
   unsigned short MatchingId;
   int NumMemRefs;
+
+  /// This is the pointer that the intrinsic is loading from or storing to.
+  /// If this is non-null, then analysis/optimization passes can assume that
+  /// this intrinsic is functionally equivalent to a load/store from this
+  /// pointer.
   Value *PtrVal;
 };
 
@@ -518,11 +523,15 @@ class TargetTransformInfo {
   unsigned getMaxInterleaveFactor(unsigned VF) const;
 
   /// \return The expected cost of arithmetic ops, such as mul, xor, fsub, etc.
+  /// \p Args is an optional argument which holds the instruction operands  
+  /// values so the TTI can analyize those values searching for special 
+  /// cases\optimizations based on those values.
   int getArithmeticInstrCost(
       unsigned Opcode, Type *Ty, OperandValueKind Opd1Info = OK_AnyValue,
       OperandValueKind Opd2Info = OK_AnyValue,
       OperandValueProperties Opd1PropInfo = OP_None,
-      OperandValueProperties Opd2PropInfo = OP_None) const;
+      OperandValueProperties Opd2PropInfo = OP_None,
+      ArrayRef<const Value *> Args = ArrayRef<const Value *>()) const;
 
   /// \return The cost of a shuffle instruction of kind Kind and of type Tp.
   /// The index and subtype parameters are used by the subvector insertion and
@@ -763,7 +772,8 @@ class TargetTransformInfo::Concept {
   getArithmeticInstrCost(unsigned Opcode, Type *Ty, OperandValueKind Opd1Info,
                          OperandValueKind Opd2Info,
                          OperandValueProperties Opd1PropInfo,
-                         OperandValueProperties Opd2PropInfo) = 0;
+                         OperandValueProperties Opd2PropInfo,
+                         ArrayRef<const Value *> Args) = 0;
   virtual int getShuffleCost(ShuffleKind Kind, Type *Tp, int Index,
                              Type *SubTp) = 0;
   virtual int getCastInstrCost(unsigned Opcode, Type *Dst, Type *Src) = 0;
@@ -984,9 +994,10 @@ class TargetTransformInfo::Model final : public TargetTransformInfo::Concept {
   getArithmeticInstrCost(unsigned Opcode, Type *Ty, OperandValueKind Opd1Info,
                          OperandValueKind Opd2Info,
                          OperandValueProperties Opd1PropInfo,
-                         OperandValueProperties Opd2PropInfo) override {
+                         OperandValueProperties Opd2PropInfo,
+                         ArrayRef<const Value *> Args) override {
     return Impl.getArithmeticInstrCost(Opcode, Ty, Opd1Info, Opd2Info,
-                                       Opd1PropInfo, Opd2PropInfo);
+                                       Opd1PropInfo, Opd2PropInfo, Args);
   }
   int getShuffleCost(ShuffleKind Kind, Type *Tp, int Index,
                      Type *SubTp) override {

contrib/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h

@@ -306,7 +306,8 @@ class TargetTransformInfoImplBase {
                                   TTI::OperandValueKind Opd1Info,
                                   TTI::OperandValueKind Opd2Info,
                                   TTI::OperandValueProperties Opd1PropInfo,
-                                  TTI::OperandValueProperties Opd2PropInfo) {
+                                  TTI::OperandValueProperties Opd2PropInfo,
+                                  ArrayRef<const Value *> Args) {
     return 1;
   }
 
@@ -427,6 +428,63 @@ class TargetTransformInfoImplBase {
     return VF;
   }
 protected:
+  // Obtain the minimum required size to hold the value (without the sign)
+  // In case of a vector it returns the min required size for one element.
+  unsigned minRequiredElementSize(const Value* Val, bool &isSigned) {
+    if (isa<ConstantDataVector>(Val) || isa<ConstantVector>(Val)) {
+      const auto* VectorValue = cast<Constant>(Val);
+
+      // In case of a vector need to pick the max between the min
+      // required size for each element
+      auto *VT = cast<VectorType>(Val->getType());
+
+      // Assume unsigned elements
+      isSigned = false;
+
+      // The max required size is the total vector width divided by num
+      // of elements in the vector
+      unsigned MaxRequiredSize = VT->getBitWidth() / VT->getNumElements();
+
+      unsigned MinRequiredSize = 0;
+      for(unsigned i = 0, e = VT->getNumElements(); i < e; ++i) {
+        if (auto* IntElement =
+              dyn_cast<ConstantInt>(VectorValue->getAggregateElement(i))) {
+          bool signedElement = IntElement->getValue().isNegative();
+          // Get the element min required size.
+          unsigned ElementMinRequiredSize =
+            IntElement->getValue().getMinSignedBits() - 1;
+          // In case one element is signed then all the vector is signed.
+          isSigned |= signedElement;
+          // Save the max required bit size between all the elements.
+          MinRequiredSize = std::max(MinRequiredSize, ElementMinRequiredSize);
+        }
+        else {
+          // not an int constant element
+          return MaxRequiredSize;
+        }
+      }
+      return MinRequiredSize;
+    }
+
+    if (const auto* CI = dyn_cast<ConstantInt>(Val)) {
+      isSigned = CI->getValue().isNegative();
+      return CI->getValue().getMinSignedBits() - 1;
+    }
+
+    if (const auto* Cast = dyn_cast<SExtInst>(Val)) {
+      isSigned = true;
+      return Cast->getSrcTy()->getScalarSizeInBits() - 1;
+    }
+
+    if (const auto* Cast = dyn_cast<ZExtInst>(Val)) {
+      isSigned = false;
+      return Cast->getSrcTy()->getScalarSizeInBits();
+    }
+
+    isSigned = false;
+    return Val->getType()->getScalarSizeInBits();
+  }
+
   bool isStridedAccess(const SCEV *Ptr) {
     return Ptr && isa<SCEVAddRecExpr>(Ptr);
   }

contrib/llvm/include/llvm/Analysis/ValueTracking.h

@@ -169,8 +169,12 @@ template <typename T> class ArrayRef;
 
   /// Return true if we can prove that the specified FP value is either a NaN or
   /// never less than 0.0.
-  bool CannotBeOrderedLessThanZero(const Value *V, const TargetLibraryInfo *TLI,
-                                   unsigned Depth = 0);
+  /// If \p IncludeNeg0 is false, -0.0 is considered less than 0.0.
+  bool CannotBeOrderedLessThanZero(const Value *V, const TargetLibraryInfo *TLI);
+
+  /// \returns true if we can prove that the specified FP value has a 0 sign
+  /// bit.
+  bool SignBitMustBeZero(const Value *V, const TargetLibraryInfo *TLI);
 
   /// If the specified value can be set by repeating the same byte in memory,
   /// return the i8 value that it is represented with. This is true for all i8

contrib/llvm/include/llvm/CodeGen/BasicTTIImpl.h

@@ -308,7 +308,8 @@ class BasicTTIImplBase : public TargetTransformInfoImplCRTPBase<T> {
       TTI::OperandValueKind Opd1Info = TTI::OK_AnyValue,
       TTI::OperandValueKind Opd2Info = TTI::OK_AnyValue,
       TTI::OperandValueProperties Opd1PropInfo = TTI::OP_None,
-      TTI::OperandValueProperties Opd2PropInfo = TTI::OP_None) {
+      TTI::OperandValueProperties Opd2PropInfo = TTI::OP_None,
+      ArrayRef<const Value *> Args = ArrayRef<const Value *>()) {
     // Check if any of the operands are vector operands.
     const TargetLoweringBase *TLI = getTLI();
     int ISD = TLI->InstructionOpcodeToISD(Opcode);

contrib/llvm/include/llvm/CodeGen/DIE.h

@@ -52,13 +52,20 @@ class DIEAbbrevData {
   /// Dwarf form code.
   dwarf::Form Form;
 
+  /// Dwarf attribute value for DW_FORM_implicit_const
+  int64_t Value;
+
 public:
-  DIEAbbrevData(dwarf::Attribute A, dwarf::Form F) : Attribute(A), Form(F) {}
+  DIEAbbrevData(dwarf::Attribute A, dwarf::Form F)
+      : Attribute(A), Form(F), Value(0) {}
+  DIEAbbrevData(dwarf::Attribute A, int64_t V)
+      : Attribute(A), Form(dwarf::DW_FORM_implicit_const), Value(V) {}
 
   /// Accessors.
   /// @{
   dwarf::Attribute getAttribute() const { return Attribute; }
   dwarf::Form getForm() const { return Form; }
+  int64_t getValue() const { return Value; }
   /// @}
 
   /// Used to gather unique data for the abbreviation folding set.
@@ -102,6 +109,11 @@ class DIEAbbrev : public FoldingSetNode {
     Data.push_back(DIEAbbrevData(Attribute, Form));
   }
 
+  /// Adds attribute with DW_FORM_implicit_const value
+  void AddImplicitConstAttribute(dwarf::Attribute Attribute, int64_t Value) {
+    Data.push_back(DIEAbbrevData(Attribute, Value));
+  }
+
   /// Used to gather unique data for the abbreviation folding set.
   void Profile(FoldingSetNodeID &ID) const;
 

contrib/llvm/include/llvm/CodeGen/GlobalISel/RegBankSelect.h

@@ -76,6 +76,7 @@ class MachineBlockFrequencyInfo;
 class MachineRegisterInfo;
 class TargetPassConfig;
 class TargetRegisterInfo;
+class raw_ostream;
 
 /// This pass implements the reg bank selector pass used in the GlobalISel
 /// pipeline. At the end of this pass, all register operands have been assigned
@@ -450,6 +451,18 @@ class RegBankSelect : public MachineFunctionPass {
     bool operator>(const MappingCost &Cost) const {
       return *this != Cost && Cost < *this;
     }
+
+    /// Print this on dbgs() stream.
+    void dump() const;
+
+    /// Print this on \p OS;
+    void print(raw_ostream &OS) const;
+
+    /// Overload the stream operator for easy debug printing.
+    friend raw_ostream &operator<<(raw_ostream &OS, const MappingCost &Cost) {
+      Cost.print(OS);
+      return OS;
+    }
   };
 
   /// Interface to the target lowering info related
@@ -626,6 +639,7 @@ class RegBankSelect : public MachineFunctionPass {
   /// \endcode
   bool runOnMachineFunction(MachineFunction &MF) override;
 };
+
 } // End namespace llvm.
 
 #endif

contrib/llvm/include/llvm/CodeGen/GlobalISel/RegisterBank.h

@@ -41,11 +41,8 @@ class RegisterBank {
   friend RegisterBankInfo;
 
 public:
-  /// The default constructor will leave the object in
-  /// an invalid state. I.e. isValid() == false.
-  /// The fields must be updated to fix that and only
-  /// RegisterBankInfo instances are allowed to do that
-  RegisterBank();
+  RegisterBank(unsigned ID, const char *Name, unsigned Size,
+               const uint32_t *ContainedRegClasses);
 
   /// Get the identifier of this register bank.
   unsigned getID() const { return ID; }

contrib/llvm/include/llvm/CodeGen/GlobalISel/RegisterBankInfo.h

@@ -384,10 +384,6 @@ class RegisterBankInfo {
 
   /// Create a RegisterBankInfo that can accomodate up to \p NumRegBanks
   /// RegisterBank instances.
-  ///
-  /// \note For the verify method to succeed all the \p NumRegBanks
-  /// must be initialized by createRegisterBank and updated with
-  /// addRegBankCoverage RegisterBank.
   RegisterBankInfo(RegisterBank **RegBanks, unsigned NumRegBanks);
 
   /// This constructor is meaningless.
@@ -400,31 +396,6 @@ class RegisterBankInfo {
     llvm_unreachable("This constructor should not be executed");
   }
 
-  /// Create a new register bank with the given parameter and add it
-  /// to RegBanks.
-  /// \pre \p ID must not already be used.
-  /// \pre \p ID < NumRegBanks.
-  void createRegisterBank(unsigned ID, const char *Name);
-
-  /// Add \p RCId to the set of register class that the register bank,
-  /// identified \p ID, covers.
-  /// This method transitively adds all the sub classes and the subreg-classes
-  /// of \p RCId to the set of covered register classes.
-  /// It also adjusts the size of the register bank to reflect the maximal
-  /// size of a value that can be hold into that register bank.
-  ///
-  /// \note This method does *not* add the super classes of \p RCId.
-  /// The rationale is if \p ID covers the registers of \p RCId, that
-  /// does not necessarily mean that \p ID covers the set of registers
-  /// of RCId's superclasses.
-  /// This method does *not* add the superreg classes as well for consistents.
-  /// The expected use is to add the coverage top-down with respect to the
-  /// register hierarchy.
-  ///
-  /// \todo TableGen should just generate the BitSet vector for us.
-  void addRegBankCoverage(unsigned ID, unsigned RCId,
-                          const TargetRegisterInfo &TRI);
-
   /// Get the register bank identified by \p ID.
   RegisterBank &getRegBank(unsigned ID) {
     assert(ID < getNumRegBanks() && "Accessing an unknown register bank");

contrib/llvm/include/llvm/CodeGen/ISDOpcodes.h

@@ -503,19 +503,6 @@ namespace ISD {
     /// address spaces.
     ADDRSPACECAST,
 
-    /// CONVERT_RNDSAT - This operator is used to support various conversions
-    /// between various types (float, signed, unsigned and vectors of those
-    /// types) with rounding and saturation. NOTE: Avoid using this operator as
-    /// most target don't support it and the operator might be removed in the
-    /// future. It takes the following arguments:
-    ///   0) value
-    ///   1) dest type (type to convert to)
-    ///   2) src type (type to convert from)
-    ///   3) rounding imm
-    ///   4) saturation imm
-    ///   5) ISD::CvtCode indicating the type of conversion to do
-    CONVERT_RNDSAT,
-
     /// FP16_TO_FP, FP_TO_FP16 - These operators are used to perform promotions
     /// and truncation for half-precision (16 bit) floating numbers. These nodes
     /// form a semi-softened interface for dealing with f16 (as an i16), which
@@ -927,21 +914,6 @@ namespace ISD {
   /// SETCC_INVALID if it is not possible to represent the resultant comparison.
   CondCode getSetCCAndOperation(CondCode Op1, CondCode Op2, bool isInteger);
 
-  //===--------------------------------------------------------------------===//
-  /// This enum defines the various converts CONVERT_RNDSAT supports.
-  enum CvtCode {
-    CVT_FF,     /// Float from Float
-    CVT_FS,     /// Float from Signed
-    CVT_FU,     /// Float from Unsigned
-    CVT_SF,     /// Signed from Float
-    CVT_UF,     /// Unsigned from Float
-    CVT_SS,     /// Signed from Signed
-    CVT_SU,     /// Signed from Unsigned
-    CVT_US,     /// Unsigned from Signed
-    CVT_UU,     /// Unsigned from Unsigned
-    CVT_INVALID /// Marker - Invalid opcode
-  };
-
 } // end llvm::ISD namespace
 
 } // end llvm namespace

contrib/llvm/include/llvm/CodeGen/SelectionDAG.h

@@ -626,12 +626,6 @@ class SelectionDAG {
 
   SDValue getCondCode(ISD::CondCode Cond);
 
-  /// Returns the ConvertRndSat Note: Avoid using this node because it may
-  /// disappear in the future and most targets don't support it.
-  SDValue getConvertRndSat(EVT VT, const SDLoc &dl, SDValue Val, SDValue DTy,
-                           SDValue STy, SDValue Rnd, SDValue Sat,
-                           ISD::CvtCode Code);
-
   /// Return an ISD::VECTOR_SHUFFLE node. The number of elements in VT,
   /// which must be a vector type, must match the number of mask elements
   /// NumElts. An integer mask element equal to -1 is treated as undefined.

contrib/llvm/include/llvm/CodeGen/SelectionDAGNodes.h

@@ -1860,26 +1860,6 @@ class CondCodeSDNode : public SDNode {
   }
 };
 
-/// NOTE: avoid using this node as this may disappear in the
-/// future and most targets don't support it.
-class CvtRndSatSDNode : public SDNode {
-  ISD::CvtCode CvtCode;
-
-  friend class SelectionDAG;
-
-  explicit CvtRndSatSDNode(EVT VT, unsigned Order, const DebugLoc &dl,
-                           ISD::CvtCode Code)
-      : SDNode(ISD::CONVERT_RNDSAT, Order, dl, getSDVTList(VT)), CvtCode(Code) {
-  }
-
-public:
-  ISD::CvtCode getCvtCode() const { return CvtCode; }
-
-  static bool classof(const SDNode *N) {
-    return N->getOpcode() == ISD::CONVERT_RNDSAT;
-  }
-};
-
 /// This class is used to represent EVT's, which are used
 /// to parameterize some operations.
 class VTSDNode : public SDNode {
@@ -2041,7 +2021,7 @@ class MaskedStoreSDNode : public MaskedLoadStoreSDNode {
   friend class SelectionDAG;
 
   MaskedStoreSDNode(unsigned Order, const DebugLoc &dl, SDVTList VTs,
-                    bool isTrunc, bool isCompressing, EVT MemVT, 
+                    bool isTrunc, bool isCompressing, EVT MemVT,
                     MachineMemOperand *MMO)
       : MaskedLoadStoreSDNode(ISD::MSTORE, Order, dl, VTs, MemVT, MMO) {
     StoreSDNodeBits.IsTruncating = isTrunc;
@@ -2054,8 +2034,8 @@ class MaskedStoreSDNode : public MaskedLoadStoreSDNode {
   bool isTruncatingStore() const { return StoreSDNodeBits.IsTruncating; }
 
   /// Returns true if the op does a compression to the vector before storing.
-  /// The node contiguously stores the active elements (integers or floats) 
-  /// in src (those with their respective bit set in writemask k) to unaligned 
+  /// The node contiguously stores the active elements (integers or floats)
+  /// in src (those with their respective bit set in writemask k) to unaligned
   /// memory at base_addr.
   bool isCompressingStore() const { return StoreSDNodeBits.IsCompressing; }
 

contrib/llvm/include/llvm/DebugInfo/CodeView/CVTypeDumper.h

@@ -0,0 +1,56 @@
+//===-- CVTypeDumper.h - CodeView type info dumper --------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_DEBUGINFO_CODEVIEW_CVTYPEDUMPER_H
+#define LLVM_DEBUGINFO_CODEVIEW_CVTYPEDUMPER_H
+
+#include "llvm/ADT/ArrayRef.h"
+#include "llvm/ADT/StringSet.h"
+#include "llvm/DebugInfo/CodeView/TypeDatabase.h"
+#include "llvm/DebugInfo/CodeView/TypeIndex.h"
+#include "llvm/DebugInfo/CodeView/TypeRecord.h"
+#include "llvm/DebugInfo/CodeView/TypeVisitorCallbacks.h"
+#include "llvm/Support/ScopedPrinter.h"
+
+namespace llvm {
+
+namespace codeview {
+
+/// Dumper for CodeView type streams found in COFF object files and PDB files.
+class CVTypeDumper {
+public:
+  explicit CVTypeDumper(TypeDatabase &TypeDB) : TypeDB(TypeDB) {}
+
+  /// Dumps one type record.  Returns false if there was a type parsing error,
+  /// and true otherwise.  This should be called in order, since the dumper
+  /// maintains state about previous records which are necessary for cross
+  /// type references.
+  Error dump(const CVType &Record, TypeVisitorCallbacks &Dumper);
+
+  /// Dumps the type records in Types. Returns false if there was a type stream
+  /// parse error, and true otherwise.
+  Error dump(const CVTypeArray &Types, TypeVisitorCallbacks &Dumper);
+
+  /// Dumps the type records in Data. Returns false if there was a type stream
+  /// parse error, and true otherwise. Use this method instead of the
+  /// CVTypeArray overload when type records are laid out contiguously in
+  /// memory.
+  Error dump(ArrayRef<uint8_t> Data, TypeVisitorCallbacks &Dumper);
+
+  static void printTypeIndex(ScopedPrinter &Printer, StringRef FieldName,
+                             TypeIndex TI, TypeDatabase &DB);
+
+private:
+  TypeDatabase &TypeDB;
+};
+
+} // end namespace codeview
+} // end namespace llvm
+
+#endif // LLVM_DEBUGINFO_CODEVIEW_TYPEDUMPER_H

contrib/llvm/include/llvm/DebugInfo/CodeView/SymbolDumper.h

@@ -20,15 +20,15 @@ namespace llvm {
 class ScopedPrinter;
 
 namespace codeview {
-class CVTypeDumper;
+class TypeDatabase;
 
 /// Dumper for CodeView symbol streams found in COFF object files and PDB files.
 class CVSymbolDumper {
 public:
-  CVSymbolDumper(ScopedPrinter &W, CVTypeDumper &CVTD,
+  CVSymbolDumper(ScopedPrinter &W, TypeDatabase &TypeDB,
                  std::unique_ptr<SymbolDumpDelegate> ObjDelegate,
                  bool PrintRecordBytes)
-      : W(W), CVTD(CVTD), ObjDelegate(std::move(ObjDelegate)),
+      : W(W), TypeDB(TypeDB), ObjDelegate(std::move(ObjDelegate)),
         PrintRecordBytes(PrintRecordBytes) {}
 
   /// Dumps one type record.  Returns false if there was a type parsing error,
@@ -43,7 +43,7 @@ class CVSymbolDumper {
 
 private:
   ScopedPrinter &W;
-  CVTypeDumper &CVTD;
+  TypeDatabase &TypeDB;
   std::unique_ptr<SymbolDumpDelegate> ObjDelegate;
 
   bool PrintRecordBytes;

contrib/llvm/include/llvm/DebugInfo/CodeView/TypeDatabase.h

@@ -0,0 +1,55 @@
+//===- TypeDatabase.h - A collection of CodeView type records ---*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_DEBUGINFO_CODEVIEW_TYPEDATABASE_H
+#define LLVM_DEBUGINFO_CODEVIEW_TYPEDATABASE_H
+
+#include "llvm/ADT/SmallVector.h"
+#include "llvm/ADT/StringRef.h"
+#include "llvm/DebugInfo/CodeView/TypeIndex.h"
+#include "llvm/DebugInfo/CodeView/TypeRecord.h"
+#include "llvm/Support/Allocator.h"
+#include "llvm/Support/StringSaver.h"
+
+namespace llvm {
+namespace codeview {
+class TypeDatabase {
+public:
+  TypeDatabase() : TypeNameStorage(Allocator) {}
+
+  /// Gets the type index for the next type record.
+  TypeIndex getNextTypeIndex() const;
+
+  /// Records the name of a type, and reserves its type index.
+  void recordType(StringRef Name, CVType Data);
+
+  /// Saves the name in a StringSet and creates a stable StringRef.
+  StringRef saveTypeName(StringRef TypeName);
+
+  StringRef getTypeName(TypeIndex Index) const;
+
+  bool containsTypeIndex(TypeIndex Index) const;
+
+  uint32_t size() const;
+
+private:
+  BumpPtrAllocator Allocator;
+
+  /// All user defined type records in .debug$T live in here. Type indices
+  /// greater than 0x1000 are user defined. Subtract 0x1000 from the index to
+  /// index into this vector.
+  SmallVector<StringRef, 10> CVUDTNames;
+  SmallVector<CVType, 10> TypeRecords;
+
+  StringSaver TypeNameStorage;
+};
+}
+}
+
+#endif

contrib/llvm/include/llvm/DebugInfo/CodeView/TypeDatabaseVisitor.h

@@ -0,0 +1,53 @@
+//===-- TypeDatabaseVisitor.h -----------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_DEBUGINFO_CODEVIEW_TYPEDATABASEVISITOR_H
+#define LLVM_DEBUGINFO_CODEVIEW_TYPEDATABASEVISITOR_H
+
+#include "llvm/DebugInfo/CodeView/TypeDatabase.h"
+#include "llvm/DebugInfo/CodeView/TypeIndex.h"
+#include "llvm/DebugInfo/CodeView/TypeRecord.h"
+#include "llvm/DebugInfo/CodeView/TypeVisitorCallbacks.h"
+
+namespace llvm {
+namespace codeview {
+
+/// Dumper for CodeView type streams found in COFF object files and PDB files.
+class TypeDatabaseVisitor : public TypeVisitorCallbacks {
+public:
+  explicit TypeDatabaseVisitor(TypeDatabase &TypeDB) : TypeDB(TypeDB) {}
+
+  /// Paired begin/end actions for all types. Receives all record data,
+  /// including the fixed-length record prefix.
+  Error visitTypeBegin(CVType &Record) override;
+  Error visitTypeEnd(CVType &Record) override;
+  Error visitMemberBegin(CVMemberRecord &Record) override;
+  Error visitMemberEnd(CVMemberRecord &Record) override;
+
+#define TYPE_RECORD(EnumName, EnumVal, Name)                                   \
+  Error visitKnownRecord(CVType &CVR, Name##Record &Record) override;
+#define MEMBER_RECORD(EnumName, EnumVal, Name)                                 \
+  Error visitKnownMember(CVMemberRecord &CVR, Name##Record &Record) override;
+#define TYPE_RECORD_ALIAS(EnumName, EnumVal, Name, AliasName)
+#define MEMBER_RECORD_ALIAS(EnumName, EnumVal, Name, AliasName)
+#include "TypeRecords.def"
+
+private:
+  bool IsInFieldList = false;
+
+  /// Name of the current type. Only valid before visitTypeEnd.
+  StringRef Name;
+
+  TypeDatabase &TypeDB;
+};
+
+} // end namespace codeview
+} // end namespace llvm
+
+#endif // LLVM_DEBUGINFO_CODEVIEW_TYPEDUMPER_H

contrib/llvm/include/llvm/DebugInfo/CodeView/TypeDumpVisitor.h

@@ -0,0 +1,67 @@
+//===-- TypeDumpVisitor.h - CodeView type info dumper -----------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_DEBUGINFO_CODEVIEW_TYPEDUMPVISITOR_H
+#define LLVM_DEBUGINFO_CODEVIEW_TYPEDUMPVISITOR_H
+
+#include "llvm/ADT/ArrayRef.h"
+#include "llvm/ADT/StringSet.h"
+#include "llvm/DebugInfo/CodeView/TypeDatabase.h"
+#include "llvm/DebugInfo/CodeView/TypeIndex.h"
+#include "llvm/DebugInfo/CodeView/TypeRecord.h"
+#include "llvm/DebugInfo/CodeView/TypeVisitorCallbacks.h"
+
+namespace llvm {
+class ScopedPrinter;
+
+namespace codeview {
+
+/// Dumper for CodeView type streams found in COFF object files and PDB files.
+class TypeDumpVisitor : public TypeVisitorCallbacks {
+public:
+  TypeDumpVisitor(TypeDatabase &TypeDB, ScopedPrinter *W, bool PrintRecordBytes)
+      : W(W), PrintRecordBytes(PrintRecordBytes), TypeDB(TypeDB) {}
+
+  void printTypeIndex(StringRef FieldName, TypeIndex TI) const;
+
+  /// Action to take on unknown types. By default, they are ignored.
+  Error visitUnknownType(CVType &Record) override;
+  Error visitUnknownMember(CVMemberRecord &Record) override;
+
+  /// Paired begin/end actions for all types. Receives all record data,
+  /// including the fixed-length record prefix.
+  Error visitTypeBegin(CVType &Record) override;
+  Error visitTypeEnd(CVType &Record) override;
+  Error visitMemberBegin(CVMemberRecord &Record) override;
+  Error visitMemberEnd(CVMemberRecord &Record) override;
+
+#define TYPE_RECORD(EnumName, EnumVal, Name)                                   \
+  Error visitKnownRecord(CVType &CVR, Name##Record &Record) override;
+#define MEMBER_RECORD(EnumName, EnumVal, Name)                                 \
+  Error visitKnownMember(CVMemberRecord &CVR, Name##Record &Record) override;
+#define TYPE_RECORD_ALIAS(EnumName, EnumVal, Name, AliasName)
+#define MEMBER_RECORD_ALIAS(EnumName, EnumVal, Name, AliasName)
+#include "TypeRecords.def"
+
+private:
+  void printMemberAttributes(MemberAttributes Attrs);
+  void printMemberAttributes(MemberAccess Access, MethodKind Kind,
+                             MethodOptions Options);
+
+  ScopedPrinter *W;
+
+  bool PrintRecordBytes = false;
+
+  TypeDatabase &TypeDB;
+};
+
+} // end namespace codeview
+} // end namespace llvm
+
+#endif

contrib/llvm/include/llvm/DebugInfo/CodeView/TypeDumper.h

@@ -1,108 +0,0 @@
-//===-- TypeDumper.h - CodeView type info dumper ----------------*- C++ -*-===//
-//
-//                     The LLVM Compiler Infrastructure
-//
-// This file is distributed under the University of Illinois Open Source
-// License. See LICENSE.TXT for details.
-//
-//===----------------------------------------------------------------------===//
-
-#ifndef LLVM_DEBUGINFO_CODEVIEW_TYPEDUMPER_H
-#define LLVM_DEBUGINFO_CODEVIEW_TYPEDUMPER_H
-
-#include "llvm/ADT/ArrayRef.h"
-#include "llvm/ADT/StringSet.h"
-#include "llvm/DebugInfo/CodeView/TypeIndex.h"
-#include "llvm/DebugInfo/CodeView/TypeRecord.h"
-#include "llvm/DebugInfo/CodeView/TypeVisitorCallbacks.h"
-
-namespace llvm {
-class ScopedPrinter;
-
-namespace codeview {
-
-/// Dumper for CodeView type streams found in COFF object files and PDB files.
-class CVTypeDumper : public TypeVisitorCallbacks {
-public:
-  CVTypeDumper(ScopedPrinter *W, bool PrintRecordBytes)
-      : W(W), PrintRecordBytes(PrintRecordBytes) {}
-
-  StringRef getTypeName(TypeIndex TI);
-  void printTypeIndex(StringRef FieldName, TypeIndex TI);
-
-  /// Dumps one type record.  Returns false if there was a type parsing error,
-  /// and true otherwise.  This should be called in order, since the dumper
-  /// maintains state about previous records which are necessary for cross
-  /// type references.
-  Error dump(const CVRecord<TypeLeafKind> &Record);
-
-  /// Dumps the type records in Types. Returns false if there was a type stream
-  /// parse error, and true otherwise.
-  Error dump(const CVTypeArray &Types);
-
-  /// Dumps the type records in Data. Returns false if there was a type stream
-  /// parse error, and true otherwise. Use this method instead of the
-  /// CVTypeArray overload when type records are laid out contiguously in
-  /// memory.
-  Error dump(ArrayRef<uint8_t> Data);
-
-  /// Gets the type index for the next type record.
-  unsigned getNextTypeIndex() const {
-    return 0x1000 + CVUDTNames.size();
-  }
-
-  /// Records the name of a type, and reserves its type index.
-  void recordType(StringRef Name) { CVUDTNames.push_back(Name); }
-
-  /// Saves the name in a StringSet and creates a stable StringRef.
-  StringRef saveName(StringRef TypeName) {
-    return TypeNames.insert(TypeName).first->getKey();
-  }
-
-  void setPrinter(ScopedPrinter *P);
-  ScopedPrinter *getPrinter() { return W; }
-
-  /// Action to take on unknown types. By default, they are ignored.
-  Error visitUnknownType(CVType &Record) override;
-  Error visitUnknownMember(CVMemberRecord &Record) override;
-
-  /// Paired begin/end actions for all types. Receives all record data,
-  /// including the fixed-length record prefix.
-  Error visitTypeBegin(CVType &Record) override;
-  Error visitTypeEnd(CVType &Record) override;
-  Error visitMemberBegin(CVMemberRecord &Record) override;
-  Error visitMemberEnd(CVMemberRecord &Record) override;
-
-#define TYPE_RECORD(EnumName, EnumVal, Name)                                   \
-  Error visitKnownRecord(CVType &CVR, Name##Record &Record) override;
-#define MEMBER_RECORD(EnumName, EnumVal, Name)                                 \
-  Error visitKnownMember(CVMemberRecord &CVR, Name##Record &Record) override;
-#define TYPE_RECORD_ALIAS(EnumName, EnumVal, Name, AliasName)
-#define MEMBER_RECORD_ALIAS(EnumName, EnumVal, Name, AliasName)
-#include "TypeRecords.def"
-
-private:
-  void printMemberAttributes(MemberAttributes Attrs);
-  void printMemberAttributes(MemberAccess Access, MethodKind Kind,
-                             MethodOptions Options);
-
-  ScopedPrinter *W;
-
-  bool IsInFieldList = false;
-  bool PrintRecordBytes = false;
-
-  /// Name of the current type. Only valid before visitTypeEnd.
-  StringRef Name;
-
-  /// All user defined type records in .debug$T live in here. Type indices
-  /// greater than 0x1000 are user defined. Subtract 0x1000 from the index to
-  /// index into this vector.
-  SmallVector<StringRef, 10> CVUDTNames;
-
-  StringSet<> TypeNames;
-};
-
-} // end namespace codeview
-} // end namespace llvm
-
-#endif // LLVM_DEBUGINFO_CODEVIEW_TYPEDUMPER_H

contrib/llvm/include/llvm/DebugInfo/DWARF/DWARFAbbreviationDeclaration.h

@@ -23,21 +23,32 @@ class raw_ostream;
 class DWARFAbbreviationDeclaration {
 public:
   struct AttributeSpec {
-    AttributeSpec(dwarf::Attribute A, dwarf::Form F, Optional<uint8_t> S)
-        : Attr(A), Form(F), ByteSize(S) {}
+    AttributeSpec(dwarf::Attribute A, dwarf::Form F, Optional<int64_t> V)
+        : Attr(A), Form(F), ByteSizeOrValue(V) {}
     dwarf::Attribute Attr;
     dwarf::Form Form;
-    /// If ByteSize has a value, then it contains the fixed size in bytes for
-    /// the Form in this object. If ByteSize doesn't have a value, then the
-    /// byte size of Form either varies according to the DWARFUnit that it is
-    /// contained in or the value size varies and must be decoded from the
-    /// debug information in order to determine its size.
-    Optional<uint8_t> ByteSize;
+    /// The following field is used for ByteSize for non-implicit_const
+    /// attributes and as value for implicit_const ones, indicated by
+    /// Form == DW_FORM_implicit_const.
+    /// The following cases are distinguished:
+    /// * Form != DW_FORM_implicit_const and ByteSizeOrValue has a value:
+    ///     ByteSizeOrValue contains the fixed size in bytes
+    ///     for the Form in this object.
+    /// * Form != DW_FORM_implicit_const and ByteSizeOrValue is None:
+    ///     byte size of Form either varies according to the DWARFUnit
+    ///     that it is contained in or the value size varies and must be
+    ///     decoded from the debug information in order to determine its size.
+    /// * Form == DW_FORM_implicit_const:
+    ///     ByteSizeOrValue contains value for the implicit_const attribute.
+    Optional<int64_t> ByteSizeOrValue;
+    bool isImplicitConst() const {
+      return Form == dwarf::DW_FORM_implicit_const;
+    }
     /// Get the fixed byte size of this Form if possible. This function might
     /// use the DWARFUnit to calculate the size of the Form, like for
     /// DW_AT_address and DW_AT_ref_addr, so this isn't just an accessor for
     /// the ByteSize member.
-    Optional<uint8_t> getByteSize(const DWARFUnit &U) const;
+    Optional<int64_t> getByteSize(const DWARFUnit &U) const;
   };
   typedef SmallVector<AttributeSpec, 8> AttributeSpecVector;
 

contrib/llvm/include/llvm/DebugInfo/DWARF/DWARFDie.h

@@ -140,21 +140,6 @@ class DWARFDie {
   const char *getAttributeValueAsString(dwarf::Attribute Attr,
                                         const char *FailValue) const;
   
-  /// Extract the specified attribute from this DIE as an address.
-  ///
-  /// Extract an attribute value from this DIE only. This call doesn't look
-  /// for the attribute value in any DW_AT_specification or
-  /// DW_AT_abstract_origin referenced DIEs.
-  ///
-  /// \param Attr the attribute to extract.
-  /// \param FailValue the value to return if this DIE doesn't have this
-  /// attribute.
-  /// \returns the address value of the attribute or FailValue if the
-  /// attribute doesn't exist or if the attribute's form isn't a form that
-  /// describes an address.
-  uint64_t getAttributeValueAsAddress(dwarf::Attribute Attr,
-                                      uint64_t FailValue) const;
-
   /// Extract the specified attribute from this DIE as an address.
   ///
   /// Extract an attribute value from this DIE only. This call doesn't look
@@ -165,21 +150,6 @@ class DWARFDie {
   /// \returns an optional value for the attribute.
   Optional<uint64_t> getAttributeValueAsAddress(dwarf::Attribute Attr) const;
   
-  /// Extract the specified attribute from this DIE as a signed integer.
-  ///
-  /// Extract an attribute value from this DIE only. This call doesn't look
-  /// for the attribute value in any DW_AT_specification or
-  /// DW_AT_abstract_origin referenced DIEs.
-  ///
-  /// \param Attr the attribute to extract.
-  /// \param FailValue the value to return if this DIE doesn't have this
-  /// attribute.
-  /// \returns the signed integer constant value of the attribute or FailValue
-  /// if the attribute doesn't exist or if the attribute's form isn't a form
-  /// that describes a signed integer.
-  int64_t getAttributeValueAsSignedConstant(dwarf::Attribute Attr,
-                                            int64_t FailValue) const;
-
   /// Extract the specified attribute from this DIE as a signed integer.
   ///
   /// Extract an attribute value from this DIE only. This call doesn't look
@@ -191,21 +161,6 @@ class DWARFDie {
   Optional<int64_t>
   getAttributeValueAsSignedConstant(dwarf::Attribute Attr) const;
   
-  /// Extract the specified attribute from this DIE as an unsigned integer.
-  ///
-  /// Extract an attribute value from this DIE only. This call doesn't look
-  /// for the attribute value in any DW_AT_specification or
-  /// DW_AT_abstract_origin referenced DIEs.
-  ///
-  /// \param Attr the attribute to extract.
-  /// \param FailValue the value to return if this DIE doesn't have this
-  /// attribute.
-  /// \returns the unsigned integer constant value of the attribute or FailValue
-  /// if the attribute doesn't exist or if the attribute's form isn't a form
-  /// that describes an unsigned integer.
-  uint64_t getAttributeValueAsUnsignedConstant(dwarf::Attribute Attr,
-                                               uint64_t FailValue) const;
-  
   /// Extract the specified attribute from this DIE as an unsigned integer.
   ///
   /// Extract an attribute value from this DIE only. This call doesn't look
@@ -217,21 +172,6 @@ class DWARFDie {
   Optional<uint64_t>
   getAttributeValueAsUnsignedConstant(dwarf::Attribute Attr) const;
 
-  /// Extract the specified attribute from this DIE as absolute DIE Offset.
-  ///
-  /// Extract an attribute value from this DIE only. This call doesn't look
-  /// for the attribute value in any DW_AT_specification or
-  /// DW_AT_abstract_origin referenced DIEs.
-  ///
-  /// \param Attr the attribute to extract.
-  /// \param FailValue the value to return if this DIE doesn't have this
-  /// attribute.
-  /// \returns the unsigned integer constant value of the attribute or FailValue
-  /// if the attribute doesn't exist or if the attribute's form isn't a form
-  /// that describes a reference.
-  uint64_t getAttributeValueAsReference(dwarf::Attribute Attr,
-                                        uint64_t FailValue) const;
-  
   /// Extract the specified attribute from this DIE as absolute DIE Offset.
   ///
   /// Extract an attribute value from this DIE only. This call doesn't look
@@ -242,20 +182,6 @@ class DWARFDie {
   /// \returns an optional value for the attribute.
   Optional<uint64_t> getAttributeValueAsReference(dwarf::Attribute Attr) const;
   
-  /// Extract the specified attribute from this DIE as absolute section offset.
-  ///
-  /// Extract an attribute value from this DIE only. This call doesn't look
-  /// for the attribute value in any DW_AT_specification or
-  /// DW_AT_abstract_origin referenced DIEs.
-  ///
-  /// \param Attr the attribute to extract.
-  /// \param FailValue the value to return if this DIE doesn't have this
-  /// attribute.
-  /// \returns the unsigned integer constant value of the attribute or FailValue
-  /// if the attribute doesn't exist or if the attribute's form isn't a form
-  /// that describes a section offset.
-  uint64_t getAttributeValueAsSectionOffset(dwarf::Attribute Attr,
-                                            uint64_t FailValue) const;
   /// Extract the specified attribute from this DIE as absolute section offset.
   ///
   /// Extract an attribute value from this DIE only. This call doesn't look

contrib/llvm/include/llvm/DebugInfo/DWARF/DWARFFormValue.h

@@ -57,6 +57,9 @@ class DWARFFormValue {
   DWARFFormValue(dwarf::Form F = dwarf::Form(0)) : Form(F), U(nullptr) {}
   dwarf::Form getForm() const { return Form; }
   void setForm(dwarf::Form F) { Form = F; }
+  void setUValue(uint64_t V) { Value.uval = V; }
+  void setSValue(int64_t V) { Value.sval = V; }
+  void setPValue(const char *V) { Value.cstr = V; }
   bool isFormClass(FormClass FC) const;
   const DWARFUnit *getUnit() const { return U; }
   void dump(raw_ostream &OS) const;

contrib/llvm/include/llvm/DebugInfo/MSF/StreamArray.h

@@ -153,30 +153,24 @@ class VarStreamArrayIterator
     return ThisValue;
   }
 
-  IterType &operator+=(std::ptrdiff_t N) {
-    while (N > 0) {
-      // We are done with the current record, discard it so that we are
-      // positioned at the next record.
-      IterRef = IterRef.drop_front(ThisLen);
-      if (IterRef.getLength() == 0) {
-        // There is nothing after the current record, we must make this an end
-        // iterator.
+  IterType &operator++() {
+    // We are done with the current record, discard it so that we are
+    // positioned at the next record.
+    IterRef = IterRef.drop_front(ThisLen);
+    if (IterRef.getLength() == 0) {
+      // There is nothing after the current record, we must make this an end
+      // iterator.
+      moveToEnd();
+    } else {
+      // There is some data after the current record.
+      auto EC = Extract(IterRef, ThisLen, ThisValue);
+      if (EC) {
+        consumeError(std::move(EC));
+        markError();
+      } else if (ThisLen == 0) {
+        // An empty record? Make this an end iterator.
         moveToEnd();
-        return *this;
-      } else {
-        // There is some data after the current record.
-        auto EC = Extract(IterRef, ThisLen, ThisValue);
-        if (EC) {
-          consumeError(std::move(EC));
-          markError();
-          return *this;
-        } else if (ThisLen == 0) {
-          // An empty record? Make this an end iterator.
-          moveToEnd();
-          return *this;
-        }
       }
-      --N;
     }
     return *this;
   }

contrib/llvm/include/llvm/IR/DIBuilder.h

@@ -17,7 +17,9 @@
 
 #include "llvm/ADT/ArrayRef.h"
 #include "llvm/ADT/DenseMap.h"
+#include "llvm/ADT/MapVector.h"
 #include "llvm/ADT/None.h"
+#include "llvm/ADT/SetVector.h"
 #include "llvm/ADT/SmallVector.h"
 #include "llvm/ADT/StringRef.h"
 #include "llvm/IR/DebugInfo.h"
@@ -51,6 +53,10 @@ namespace llvm {
     SmallVector<Metadata *, 4> AllSubprograms;
     SmallVector<Metadata *, 4> AllGVs;
     SmallVector<TrackingMDNodeRef, 4> AllImportedModules;
+    /// Map Macro parent (which can be DIMacroFile or nullptr) to a list of
+    /// Metadata all of type DIMacroNode.
+    /// DIMacroNode's with nullptr parent are DICompileUnit direct children.
+    MapVector<MDNode *, SetVector<Metadata *>> AllMacrosPerParent;
 
     /// Track nodes that may be unresolved.
     SmallVector<TrackingMDNodeRef, 4> UnresolvedNodes;
@@ -116,6 +122,24 @@ namespace llvm {
                        DIFile::ChecksumKind CSKind = DIFile::CSK_None,
                        StringRef Checksum = StringRef());
 
+    /// Create debugging information entry for a macro.
+    /// \param Parent     Macro parent (could be nullptr).
+    /// \param Line       Source line number where the macro is defined.
+    /// \param MacroType  DW_MACINFO_define or DW_MACINFO_undef.
+    /// \param Name       Macro name.
+    /// \param Value      Macro value.
+    DIMacro *createMacro(DIMacroFile *Parent, unsigned Line, unsigned MacroType,
+                         StringRef Name, StringRef Value = StringRef());
+
+    /// Create debugging information temporary entry for a macro file.
+    /// List of macro node direct children will be calculated by DIBuilder,
+    /// using the \p Parent relationship.
+    /// \param Parent     Macro file parent (could be nullptr).
+    /// \param Line       Source line number where the macro file is included.
+    /// \param File       File descriptor containing the name of the macro file.
+    DIMacroFile *createTempMacroFile(DIMacroFile *Parent, unsigned Line,
+                                     DIFile *File);
+
     /// Create a single enumerator value.
     DIEnumerator *createEnumerator(StringRef Name, int64_t Val);
 
@@ -447,6 +471,9 @@ namespace llvm {
     /// Get a DINodeArray, create one if required.
     DINodeArray getOrCreateArray(ArrayRef<Metadata *> Elements);
 
+    /// Get a DIMacroNodeArray, create one if required.
+    DIMacroNodeArray getOrCreateMacroArray(ArrayRef<Metadata *> Elements);
+
     /// Get a DITypeRefArray, create one if required.
     DITypeRefArray getOrCreateTypeArray(ArrayRef<Metadata *> Elements);
 

contrib/llvm/include/llvm/IR/DebugInfoMetadata.h

@@ -1295,16 +1295,12 @@ class DILocation : public MDNode {
   /// Check \c this can be discriminated from \c RHS in a linetable entry.
   /// Scope and inlined-at chains are not recorded in the linetable, so they
   /// cannot be used to distinguish basic blocks.
-  ///
-  /// The current implementation is weaker than it should be, since it just
-  /// checks filename and line.
-  ///
-  /// FIXME: Add a check for getDiscriminator().
-  /// FIXME: Add a check for getColumn().
-  /// FIXME: Change the getFilename() check to getFile() (or add one for
-  /// getDirectory()).
   bool canDiscriminate(const DILocation &RHS) const {
-    return getFilename() != RHS.getFilename() || getLine() != RHS.getLine();
+    return getLine() != RHS.getLine() ||
+           getColumn() != RHS.getColumn() ||
+           getDiscriminator() != RHS.getDiscriminator() ||
+           getFilename() != RHS.getFilename() ||
+           getDirectory() != RHS.getDirectory();
   }
 
   /// Get the DWARF discriminator.
@@ -1327,10 +1323,13 @@ class DILocation : public MDNode {
   /// represented in a single line entry.  In this case, no location
   /// should be set.
   ///
-  /// Currently this function is simply a stub, and no location will be
-  /// used for all cases.
-  static DILocation *getMergedLocation(const DILocation *LocA,
-                                       const DILocation *LocB) {
+  /// Currently the function does not create a new location. If the locations
+  /// are the same, or cannot be discriminated, the first location is returned.
+  /// Otherwise an empty location will be used.
+  static const DILocation *getMergedLocation(const DILocation *LocA,
+                                             const DILocation *LocB) {
+    if (LocA && LocB && (LocA == LocB || !LocA->canDiscriminate(*LocB)))
+      return LocA;
     return nullptr;
   }
 

contrib/llvm/include/llvm/IR/GlobalObject.h

@@ -37,11 +37,11 @@ class GlobalObject : public GlobalValue {
     setGlobalValueSubClassData(0);
   }
 
-  std::string Section;     // Section to emit this into, empty means default
   Comdat *ObjComdat;
   enum {
     LastAlignmentBit = 4,
     HasMetadataHashEntryBit,
+    HasSectionHashEntryBit,
 
     GlobalObjectBits,
   };
@@ -66,8 +66,26 @@ class GlobalObject : public GlobalValue {
   unsigned getGlobalObjectSubClassData() const;
   void setGlobalObjectSubClassData(unsigned Val);
 
-  bool hasSection() const { return !getSection().empty(); }
-  StringRef getSection() const { return Section; }
+  /// Check if this global has a custom object file section.
+  ///
+  /// This is more efficient than calling getSection() and checking for an empty
+  /// string.
+  bool hasSection() const {
+    return getGlobalValueSubClassData() & (1 << HasSectionHashEntryBit);
+  }
+
+  /// Get the custom section of this global if it has one.
+  ///
+  /// If this global does not have a custom section, this will be empty and the
+  /// default object file section (.text, .data, etc) will be used.
+  StringRef getSection() const {
+    return hasSection() ? getSectionImpl() : StringRef();
+  }
+
+  /// Change the section for this global.
+  ///
+  /// Setting the section to the empty string tells LLVM to choose an
+  /// appropriate default object file section.
   void setSection(StringRef S);
 
   bool hasComdat() const { return getComdat() != nullptr; }
@@ -134,14 +152,20 @@ class GlobalObject : public GlobalValue {
   void clearMetadata();
 
 private:
+  void setGlobalObjectFlag(unsigned Bit, bool Val) {
+    unsigned Mask = 1 << Bit;
+    setGlobalValueSubClassData((~Mask & getGlobalValueSubClassData()) |
+                               (Val ? Mask : 0u));
+  }
+
   bool hasMetadataHashEntry() const {
     return getGlobalValueSubClassData() & (1 << HasMetadataHashEntryBit);
   }
   void setHasMetadataHashEntry(bool HasEntry) {
-    unsigned Mask = 1 << HasMetadataHashEntryBit;
-    setGlobalValueSubClassData((~Mask & getGlobalValueSubClassData()) |
-                               (HasEntry ? Mask : 0u));
+    setGlobalObjectFlag(HasMetadataHashEntryBit, HasEntry);
   }
+
+  StringRef getSectionImpl() const;
 };
 
 } // end namespace llvm

contrib/llvm/include/llvm/IR/Intrinsics.td

@@ -578,8 +578,8 @@ def int_invariant_end   : Intrinsic<[],
                                      llvm_anyptr_ty],
                                     [IntrArgMemOnly, NoCapture<2>]>;
 
-def int_invariant_group_barrier : Intrinsic<[llvm_ptr_ty], 
-                                            [llvm_ptr_ty], 
+def int_invariant_group_barrier : Intrinsic<[llvm_ptr_ty],
+                                            [llvm_ptr_ty],
                                             [IntrNoMem]>;
 
 //===------------------------ Stackmap Intrinsics -------------------------===//
@@ -683,29 +683,6 @@ def int_convert_to_fp16   : Intrinsic<[llvm_i16_ty], [llvm_anyfloat_ty]>;
 def int_convert_from_fp16 : Intrinsic<[llvm_anyfloat_ty], [llvm_i16_ty]>;
 }
 
-// These convert intrinsics are to support various conversions between
-// various types with rounding and saturation. NOTE: avoid using these
-// intrinsics as they might be removed sometime in the future and
-// most targets don't support them.
-def int_convertff  : Intrinsic<[llvm_anyfloat_ty],
-                               [llvm_anyfloat_ty, llvm_i32_ty, llvm_i32_ty]>;
-def int_convertfsi : Intrinsic<[llvm_anyfloat_ty],
-                               [llvm_anyint_ty, llvm_i32_ty, llvm_i32_ty]>;
-def int_convertfui : Intrinsic<[llvm_anyfloat_ty],
-                               [llvm_anyint_ty, llvm_i32_ty, llvm_i32_ty]>;
-def int_convertsif : Intrinsic<[llvm_anyint_ty],
-                               [llvm_anyfloat_ty, llvm_i32_ty, llvm_i32_ty]>;
-def int_convertuif : Intrinsic<[llvm_anyint_ty],
-                               [llvm_anyfloat_ty, llvm_i32_ty, llvm_i32_ty]>;
-def int_convertss  : Intrinsic<[llvm_anyint_ty],
-                               [llvm_anyint_ty, llvm_i32_ty, llvm_i32_ty]>;
-def int_convertsu  : Intrinsic<[llvm_anyint_ty],
-                               [llvm_anyint_ty, llvm_i32_ty, llvm_i32_ty]>;
-def int_convertus  : Intrinsic<[llvm_anyint_ty],
-                               [llvm_anyint_ty, llvm_i32_ty, llvm_i32_ty]>;
-def int_convertuu  : Intrinsic<[llvm_anyint_ty],
-                               [llvm_anyint_ty, llvm_i32_ty, llvm_i32_ty]>;
-
 // Clear cache intrinsic, default to ignore (ie. emit nothing)
 // maps to void __clear_cache() on supporting platforms
 def int_clear_cache : Intrinsic<[], [llvm_ptr_ty, llvm_ptr_ty],

contrib/llvm/include/llvm/IR/IntrinsicsAArch64.td

@@ -37,12 +37,6 @@ def int_aarch64_udiv : Intrinsic<[llvm_anyint_ty], [LLVMMatchType<0>,
 
 def int_aarch64_hint : Intrinsic<[], [llvm_i32_ty]>;
 
-//===----------------------------------------------------------------------===//
-// RBIT
-
-def int_aarch64_rbit : Intrinsic<[llvm_anyint_ty], [LLVMMatchType<0>],
-                                 [IntrNoMem]>;
-
 //===----------------------------------------------------------------------===//
 // Data Barrier Instructions
 

contrib/llvm/include/llvm/IR/IntrinsicsARM.td

@@ -155,11 +155,6 @@ def int_arm_crc32cw : Intrinsic<[llvm_i32_ty], [llvm_i32_ty, llvm_i32_ty],
 def int_arm_hint : Intrinsic<[], [llvm_i32_ty]>;
 def int_arm_dbg : Intrinsic<[], [llvm_i32_ty]>;
 
-//===----------------------------------------------------------------------===//
-// RBIT
-
-def int_arm_rbit : Intrinsic<[llvm_i32_ty], [llvm_i32_ty], [IntrNoMem]>;
-
 //===----------------------------------------------------------------------===//
 // UND (reserved undefined sequence)
 

contrib/llvm/include/llvm/IR/ModuleSummaryIndex.h

@@ -317,10 +317,10 @@ struct TypeTestResolution {
                ///  All-Ones Bit Vectors")
   } TheKind = Unsat;
 
-  /// Range of the size expressed as a bit width. For example, if the size is in
-  /// range [0,256), this number will be 8. This helps generate the most compact
+  /// Range of size-1 expressed as a bit width. For example, if the size is in
+  /// range [1,256], this number will be 8. This helps generate the most compact
   /// instruction sequences.
-  unsigned SizeBitWidth = 0;
+  unsigned SizeM1BitWidth = 0;
 };
 
 struct TypeIdSummary {

contrib/llvm/include/llvm/IR/ModuleSummaryIndexYAML.h

@@ -29,7 +29,7 @@ template <> struct ScalarEnumerationTraits<TypeTestResolution::Kind> {
 template <> struct MappingTraits<TypeTestResolution> {
   static void mapping(IO &io, TypeTestResolution &res) {
     io.mapOptional("Kind", res.TheKind);
-    io.mapOptional("SizeBitWidth", res.SizeBitWidth);
+    io.mapOptional("SizeM1BitWidth", res.SizeM1BitWidth);
   }
 };
 

contrib/llvm/include/llvm/Object/Decompressor.h

@@ -0,0 +1,64 @@
+//===-- Decompressor.h ------------------------------------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===/
+
+#ifndef LLVM_OBJECT_DECOMPRESSOR_H
+#define LLVM_OBJECT_DECOMPRESSOR_H
+
+#include "llvm/ADT/StringRef.h"
+#include "llvm/ADT/SmallString.h"
+#include "llvm/Object/ObjectFile.h"
+
+namespace llvm {
+namespace object {
+
+/// @brief Decompressor helps to handle decompression of compressed sections.
+class Decompressor {
+public:
+  /// @brief Create decompressor object.
+  /// @param Name        Section name.
+  /// @param Data        Section content.
+  /// @param IsLE        Flag determines if Data is in little endian form.
+  /// @param Is64Bit     Flag determines if object is 64 bit.
+  static Expected<Decompressor> create(StringRef Name, StringRef Data,
+                                       bool IsLE, bool Is64Bit);
+
+  /// @brief Resize the buffer and uncompress section data into it.
+  /// @param Out         Destination buffer.
+  Error decompress(SmallString<32> &Out);
+
+  /// @brief Uncompress section data to raw buffer provided.
+  /// @param Buffer      Destination buffer.
+  Error decompress(MutableArrayRef<char> Buffer);
+
+  /// @brief Return memory buffer size required for decompression.
+  uint64_t getDecompressedSize() { return DecompressedSize; }
+
+  /// @brief Return true if section is compressed, including gnu-styled case.
+  static bool isCompressed(const object::SectionRef &Section);
+
+  /// @brief Return true if section is a ELF compressed one.
+  static bool isCompressedELFSection(uint64_t Flags, StringRef Name);
+
+  /// @brief Return true if section name matches gnu style compressed one.
+  static bool isGnuStyle(StringRef Name);
+
+private:
+  Decompressor(StringRef Data);
+
+  Error consumeCompressedGnuHeader();
+  Error consumeCompressedZLibHeader(bool Is64Bit, bool IsLittleEndian);
+
+  StringRef SectionData;
+  uint64_t DecompressedSize;
+};
+
+} // end namespace object
+} // end namespace llvm
+
+#endif // LLVM_OBJECT_DECOMPRESSOR_H

contrib/llvm/include/llvm/ObjectYAML/DWARFYAML.h

@@ -85,6 +85,41 @@ struct Unit {
   std::vector<Entry> Entries;
 };
 
+struct File {
+  StringRef Name;
+  uint64_t DirIdx;
+  uint64_t ModTime;
+  uint64_t Length;
+};
+
+struct LineTableOpcode {
+  dwarf::LineNumberOps Opcode;
+  uint64_t ExtLen;
+  dwarf::LineNumberExtendedOps SubOpcode;
+  uint64_t Data;
+  int64_t SData;
+  File FileEntry;
+  std::vector<llvm::yaml::Hex8> UnknownOpcodeData;
+  std::vector<llvm::yaml::Hex64> StandardOpcodeData;
+};
+
+struct LineTable {
+  uint32_t TotalLength;
+  uint64_t TotalLength64;
+  uint16_t Version;
+  uint64_t PrologueLength;
+  uint8_t MinInstLength;
+  uint8_t MaxOpsPerInst;
+  uint8_t DefaultIsStmt;
+  uint8_t LineBase;
+  uint8_t LineRange;
+  uint8_t OpcodeBase;
+  std::vector<uint8_t> StandardOpcodeLengths;
+  std::vector<StringRef> IncludeDirs;
+  std::vector<File> Files;
+  std::vector<LineTableOpcode> Opcodes;
+};
+
 struct Data {
   bool IsLittleEndian;
   std::vector<Abbrev> AbbrevDecls;
@@ -98,6 +133,8 @@ struct Data {
   
   std::vector<Unit> CompileUnits;
 
+  std::vector<LineTable> DebugLines;
+
   bool isEmpty() const;
 };
 
@@ -105,6 +142,7 @@ struct Data {
 } // namespace llvm
 
 LLVM_YAML_IS_SEQUENCE_VECTOR(uint8_t)
+LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::yaml::Hex64)
 LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::StringRef)
 LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::yaml::Hex8)
 LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::DWARFYAML::AttributeAbbrev)
@@ -115,6 +153,9 @@ LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::DWARFYAML::PubEntry)
 LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::DWARFYAML::Unit)
 LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::DWARFYAML::FormValue)
 LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::DWARFYAML::Entry)
+LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::DWARFYAML::File)
+LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::DWARFYAML::LineTable)
+LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::DWARFYAML::LineTableOpcode)
 
 namespace llvm {
 namespace yaml {
@@ -159,6 +200,18 @@ template <> struct MappingTraits<DWARFYAML::FormValue> {
   static void mapping(IO &IO, DWARFYAML::FormValue &FormValue);
 };
 
+template <> struct MappingTraits<DWARFYAML::File> {
+  static void mapping(IO &IO, DWARFYAML::File &File);
+};
+  
+template <> struct MappingTraits<DWARFYAML::LineTableOpcode> {
+  static void mapping(IO &IO, DWARFYAML::LineTableOpcode &LineTableOpcode);
+};
+
+template <> struct MappingTraits<DWARFYAML::LineTable> {
+  static void mapping(IO &IO, DWARFYAML::LineTable &LineTable);
+};
+
 #define HANDLE_DW_TAG(unused, name)                                            \
   io.enumCase(value, "DW_TAG_" #name, dwarf::DW_TAG_##name);
 
@@ -169,6 +222,26 @@ template <> struct ScalarEnumerationTraits<dwarf::Tag> {
   }
 };
 
+#define HANDLE_DW_LNS(unused, name)                                            \
+  io.enumCase(value, "DW_LNS_" #name, dwarf::DW_LNS_##name);
+
+template <> struct ScalarEnumerationTraits<dwarf::LineNumberOps> {
+  static void enumeration(IO &io, dwarf::LineNumberOps &value) {
+#include "llvm/Support/Dwarf.def"
+    io.enumFallback<Hex8>(value);
+  }
+};
+
+#define HANDLE_DW_LNE(unused, name)                                            \
+  io.enumCase(value, "DW_LNE_" #name, dwarf::DW_LNE_##name);
+
+template <> struct ScalarEnumerationTraits<dwarf::LineNumberExtendedOps> {
+  static void enumeration(IO &io, dwarf::LineNumberExtendedOps &value) {
+#include "llvm/Support/Dwarf.def"
+    io.enumFallback<Hex16>(value);
+  }
+};
+
 #define HANDLE_DW_AT(unused, name)                                             \
   io.enumCase(value, "DW_AT_" #name, dwarf::DW_AT_##name);
 

contrib/llvm/include/llvm/ObjectYAML/MachOYAML.h

@@ -139,7 +139,6 @@ struct UniversalBinary {
 
 LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::MachOYAML::LoadCommand)
 LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::MachOYAML::Section)
-LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::yaml::Hex64)
 LLVM_YAML_IS_SEQUENCE_VECTOR(int64_t)
 LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::MachOYAML::RebaseOpcode)
 LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::MachOYAML::BindOpcode)

contrib/llvm/include/llvm/Passes/PassBuilder.h

@@ -18,8 +18,8 @@
 
 #include "llvm/ADT/Optional.h"
 #include "llvm/Analysis/CGSCCPassManager.h"
-#include "llvm/Analysis/LoopPassManager.h"
 #include "llvm/IR/PassManager.h"
+#include "llvm/Transforms/Scalar/LoopPassManager.h"
 #include <vector>
 
 namespace llvm {

contrib/llvm/include/llvm/ProfileData/InstrProf.h

@@ -230,6 +230,15 @@ class InstrProfSymtab;
 /// bytes. This method decodes the string and populates the \c Symtab.
 Error readPGOFuncNameStrings(StringRef NameStrings, InstrProfSymtab &Symtab);
 
+/// Check if INSTR_PROF_RAW_VERSION_VAR is defined. This global is only being
+/// set in IR PGO compilation.
+bool isIRPGOFlagSet(const Module *M);
+
+/// Check if we can safely rename this Comdat function. Instances of the same
+/// comdat function may have different control flows thus can not share the
+/// same counter variable.
+bool canRenameComdatFunc(const Function &F, bool CheckAddressTaken = false);
+
 enum InstrProfValueKind : uint32_t {
 #define VALUE_PROF_KIND(Enumerator, Value) Enumerator = Value,
 #include "llvm/ProfileData/InstrProfData.inc"

contrib/llvm/include/llvm/Support/CommandLine.h

@@ -201,7 +201,7 @@ class SubCommand {
 
   void reset();
 
-  operator bool() const;
+  explicit operator bool() const;
 
   StringRef getName() const { return Name; }
   StringRef getDescription() const { return Description; }

contrib/llvm/include/llvm/Support/Dwarf.h

@@ -207,7 +207,7 @@ enum DiscriminantList {
 };
 
 /// Line Number Standard Opcode Encodings.
-enum LineNumberOps {
+enum LineNumberOps : uint8_t {
 #define HANDLE_DW_LNS(ID, NAME) DW_LNS_##NAME = ID,
 #include "llvm/Support/Dwarf.def"
 };

contrib/llvm/include/llvm/Support/FileOutputBuffer.h

@@ -78,11 +78,12 @@ class FileOutputBuffer {
   FileOutputBuffer &operator=(const FileOutputBuffer &) = delete;
 
   FileOutputBuffer(std::unique_ptr<llvm::sys::fs::mapped_file_region> R,
-                   StringRef Path, StringRef TempPath);
+                   StringRef Path, StringRef TempPath, bool IsRegular);
 
   std::unique_ptr<llvm::sys::fs::mapped_file_region> Region;
   SmallString<128>    FinalPath;
   SmallString<128>    TempPath;
+  bool IsRegular;
 };
 } // end namespace llvm
 

contrib/llvm/include/llvm/Support/GenericDomTree.h

@@ -571,9 +571,15 @@ template <class NodeT> class DominatorTreeBase : public DominatorBase<NodeT> {
   // API to update (Post)DominatorTree information based on modifications to
   // the CFG...
 
-  /// addNewBlock - Add a new node to the dominator tree information.  This
-  /// creates a new node as a child of DomBB dominator node,linking it into
-  /// the children list of the immediate dominator.
+  /// Add a new node to the dominator tree information.
+  ///
+  /// This creates a new node as a child of DomBB dominator node, linking it
+  /// into the children list of the immediate dominator.
+  ///
+  /// \param BB New node in CFG.
+  /// \param DomBB CFG node that is dominator for BB.
+  /// \returns New dominator tree node that represents new CFG node.
+  ///
   DomTreeNodeBase<NodeT> *addNewBlock(NodeT *BB, NodeT *DomBB) {
     assert(getNode(BB) == nullptr && "Block already in dominator tree!");
     DomTreeNodeBase<NodeT> *IDomNode = getNode(DomBB);
@@ -583,6 +589,31 @@ template <class NodeT> class DominatorTreeBase : public DominatorBase<NodeT> {
                 llvm::make_unique<DomTreeNodeBase<NodeT>>(BB, IDomNode))).get();
   }
 
+  /// Add a new node to the forward dominator tree and make it a new root.
+  ///
+  /// \param BB New node in CFG.
+  /// \returns New dominator tree node that represents new CFG node.
+  ///
+  DomTreeNodeBase<NodeT> *setNewRoot(NodeT *BB) {
+    assert(getNode(BB) == nullptr && "Block already in dominator tree!");
+    assert(!this->isPostDominator() &&
+           "Cannot change root of post-dominator tree");
+    DFSInfoValid = false;
+    auto &Roots = DominatorBase<NodeT>::Roots;
+    DomTreeNodeBase<NodeT> *NewNode = (DomTreeNodes[BB] =
+      llvm::make_unique<DomTreeNodeBase<NodeT>>(BB, nullptr)).get();
+    if (Roots.empty()) {
+      addRoot(BB);
+    } else {
+      assert(Roots.size() == 1);
+      NodeT *OldRoot = Roots.front();
+      DomTreeNodes[OldRoot] =
+        NewNode->addChild(std::move(DomTreeNodes[OldRoot]));
+      Roots[0] = BB;
+    }
+    return RootNode = NewNode;
+  }
+
   /// changeImmediateDominator - This method is used to update the dominator
   /// tree information when a node's immediate dominator changes.
   ///

contrib/llvm/include/llvm/Target/TargetLowering.h

@@ -23,66 +23,80 @@
 #ifndef LLVM_TARGET_TARGETLOWERING_H
 #define LLVM_TARGET_TARGETLOWERING_H
 
+#include "llvm/ADT/ArrayRef.h"
 #include "llvm/ADT/DenseMap.h"
+#include "llvm/ADT/SmallVector.h"
+#include "llvm/ADT/STLExtras.h"
+#include "llvm/ADT/StringRef.h"
 #include "llvm/CodeGen/DAGCombine.h"
+#include "llvm/CodeGen/ISDOpcodes.h"
+#include "llvm/CodeGen/MachineValueType.h"
 #include "llvm/CodeGen/RuntimeLibcalls.h"
 #include "llvm/CodeGen/SelectionDAGNodes.h"
+#include "llvm/CodeGen/ValueTypes.h"
 #include "llvm/IR/Attributes.h"
 #include "llvm/IR/CallSite.h"
 #include "llvm/IR/CallingConv.h"
+#include "llvm/IR/DataLayout.h"
+#include "llvm/IR/DerivedTypes.h"
 #include "llvm/IR/IRBuilder.h"
 #include "llvm/IR/InlineAsm.h"
+#include "llvm/IR/Instruction.h"
 #include "llvm/IR/Instructions.h"
+#include "llvm/IR/Type.h"
 #include "llvm/MC/MCRegisterInfo.h"
+#include "llvm/Support/AtomicOrdering.h"
+#include "llvm/Support/Casting.h"
+#include "llvm/Support/ErrorHandling.h"
 #include "llvm/Target/TargetCallingConv.h"
 #include "llvm/Target/TargetMachine.h"
+#include <algorithm>
+#include <cassert>
 #include <climits>
+#include <cstdint>
+#include <iterator>
 #include <map>
+#include <string>
+#include <utility>
 #include <vector>
 
 namespace llvm {
-  class BranchProbability;
-  class CallInst;
-  class CCState;
-  class CCValAssign;
-  class FastISel;
-  class FunctionLoweringInfo;
-  class ImmutableCallSite;
-  class IntrinsicInst;
-  class MachineBasicBlock;
-  class MachineFunction;
-  class MachineInstr;
-  class MachineJumpTableInfo;
-  class MachineLoop;
-  class MachineRegisterInfo;
-  class Mangler;
-  class MCContext;
-  class MCExpr;
-  class MCSymbol;
-  template<typename T> class SmallVectorImpl;
-  class DataLayout;
-  class TargetRegisterClass;
-  class TargetLibraryInfo;
-  class TargetLoweringObjectFile;
-  class Value;
 
-  namespace Sched {
-    enum Preference {
-      None,             // No preference
-      Source,           // Follow source order.
-      RegPressure,      // Scheduling for lowest register pressure.
-      Hybrid,           // Scheduling for both latency and register pressure.
-      ILP,              // Scheduling for ILP in low register pressure mode.
-      VLIW              // Scheduling for VLIW targets.
-    };
-  }
+class BranchProbability;
+class CCState;
+class CCValAssign;
+class FastISel;
+class FunctionLoweringInfo;
+class IntrinsicInst;
+class MachineBasicBlock;
+class MachineFunction;
+class MachineInstr;
+class MachineJumpTableInfo;
+class MachineLoop;
+class MachineRegisterInfo;
+class MCContext;
+class MCExpr;
+class TargetRegisterClass;
+class TargetLibraryInfo;
+class TargetRegisterInfo;
+class Value;
+
+namespace Sched {
+
+  enum Preference {
+    None,             // No preference
+    Source,           // Follow source order.
+    RegPressure,      // Scheduling for lowest register pressure.
+    Hybrid,           // Scheduling for both latency and register pressure.
+    ILP,              // Scheduling for ILP in low register pressure mode.
+    VLIW              // Scheduling for VLIW targets.
+  };
+
+} // end namespace Sched
 
 /// This base class for TargetLowering contains the SelectionDAG-independent
 /// parts that can be used from the rest of CodeGen.
 class TargetLoweringBase {
-  TargetLoweringBase(const TargetLoweringBase&) = delete;
-  void operator=(const TargetLoweringBase&) = delete;
-
 public:
   /// This enum indicates whether operations are valid for a target, and if not,
   /// what action should be used to make them valid.
@@ -166,7 +180,9 @@ class TargetLoweringBase {
 
   /// NOTE: The TargetMachine owns TLOF.
   explicit TargetLoweringBase(const TargetMachine &TM);
-  virtual ~TargetLoweringBase() {}
+  TargetLoweringBase(const TargetLoweringBase&) = delete;
+  void operator=(const TargetLoweringBase&) = delete;
+  virtual ~TargetLoweringBase() = default;
 
 protected:
   /// \brief Initialize all of the actions to default values.
@@ -599,19 +615,18 @@ class TargetLoweringBase {
                                   MVT &RegisterVT) const;
 
   struct IntrinsicInfo {
-    unsigned     opc;         // target opcode
-    EVT          memVT;       // memory VT
-    const Value* ptrVal;      // value representing memory location
-    int          offset;      // offset off of ptrVal
-    unsigned     size;        // the size of the memory location
-                              // (taken from memVT if zero)
-    unsigned     align;       // alignment
-    bool         vol;         // is volatile?
-    bool         readMem;     // reads memory?
-    bool         writeMem;    // writes memory?
+    unsigned     opc = 0;          // target opcode
+    EVT          memVT;            // memory VT
+    const Value* ptrVal = nullptr; // value representing memory location
+    int          offset = 0;       // offset off of ptrVal
+    unsigned     size = 0;         // the size of the memory location
+                                   // (taken from memVT if zero)
+    unsigned     align = 1;        // alignment
+    bool         vol = false;      // is volatile?
+    bool         readMem = false;  // reads memory?
+    bool         writeMem = false; // writes memory?
 
-    IntrinsicInfo() : opc(0), ptrVal(nullptr), offset(0), size(0), align(1),
-                      vol(false), readMem(false), writeMem(false) {}
+    IntrinsicInfo() = default;
   };
 
   /// Given an intrinsic, checks if on the target the intrinsic will need to map
@@ -823,7 +838,6 @@ class TargetLoweringBase {
       getCondCodeAction(CC, VT) == Custom;
   }
 
-
   /// If the action for this operation is to promote, this method returns the
   /// ValueType to promote to.
   MVT getTypeToPromoteTo(unsigned Op, MVT VT) const {
@@ -1643,11 +1657,11 @@ class TargetLoweringBase {
   /// If Scale is zero, there is no ScaleReg.  Scale of 1 indicates a reg with
   /// no scale.
   struct AddrMode {
-    GlobalValue *BaseGV;
-    int64_t      BaseOffs;
-    bool         HasBaseReg;
-    int64_t      Scale;
-    AddrMode() : BaseGV(nullptr), BaseOffs(0), HasBaseReg(false), Scale(0) {}
+    GlobalValue *BaseGV = nullptr;
+    int64_t      BaseOffs = 0;
+    bool         HasBaseReg = false;
+    int64_t      Scale = 0;
+    AddrMode() = default;
   };
 
   /// Return true if the addressing mode represented by AM is legal for this
@@ -2093,8 +2107,6 @@ class TargetLoweringBase {
 private:
   LegalizeKind getTypeConversion(LLVMContext &Context, EVT VT) const;
 
-private:
-
   /// Targets can specify ISD nodes that they would like PerformDAGCombine
   /// callbacks for by calling setTargetDAGCombine(), which sets a bit in this
   /// array.
@@ -2192,7 +2204,6 @@ class TargetLoweringBase {
   /// \see enableExtLdPromotion.
   bool EnableExtLdPromotion;
 
-protected:
   /// Return true if the value types that can be represented by the specified
   /// register class are all legal.
   bool isLegalRC(const TargetRegisterClass *RC) const;
@@ -2209,12 +2220,12 @@ class TargetLoweringBase {
 /// This class also defines callbacks that targets must implement to lower
 /// target-specific constructs to SelectionDAG operators.
 class TargetLowering : public TargetLoweringBase {
-  TargetLowering(const TargetLowering&) = delete;
-  void operator=(const TargetLowering&) = delete;
-
 public:
   struct DAGCombinerInfo;
 
+  TargetLowering(const TargetLowering&) = delete;
+  void operator=(const TargetLowering&) = delete;
+
   /// NOTE: The TargetMachine owns TLOF.
   explicit TargetLowering(const TargetMachine &TM);
 
@@ -2376,6 +2387,7 @@ class TargetLowering : public TargetLoweringBase {
     void *DC;  // The DAG Combiner object.
     CombineLevel Level;
     bool CalledByLegalizer;
+
   public:
     SelectionDAG &DAG;
 
@@ -2542,7 +2554,7 @@ class TargetLowering : public TargetLoweringBase {
     ArgListEntry() : isSExt(false), isZExt(false), isInReg(false),
       isSRet(false), isNest(false), isByVal(false), isInAlloca(false),
       isReturned(false), isSwiftSelf(false), isSwiftError(false),
-      Alignment(0) { }
+      Alignment(0) {}
 
     void setAttributes(ImmutableCallSite *CS, unsigned AttrIdx);
   };
@@ -2681,7 +2693,6 @@ class TargetLowering : public TargetLoweringBase {
     ArgListTy &getArgs() {
       return Args;
     }
-
   };
 
   /// This function lowers an abstract call to a function into an actual call.
@@ -3118,6 +3129,13 @@ class TargetLowering : public TargetLoweringBase {
                                  EVT DataVT, SelectionDAG &DAG,
                                  bool IsCompressedMemory) const;
 
+  /// Get a pointer to vector element \p Idx located in memory for a vector of
+  /// type \p VecVT starting at a base address of \p VecPtr. If \p Idx is out of
+  /// bounds the returned pointer is unspecified, but will be within the vector
+  /// bounds.
+  SDValue getVectorElementPointer(SelectionDAG &DAG, SDValue VecPtr, EVT VecVT,
+                                  SDValue Idx) const;
+
   //===--------------------------------------------------------------------===//
   // Instruction Emitting Hooks
   //
@@ -3169,6 +3187,6 @@ void GetReturnInfo(Type *ReturnType, AttributeSet attr,
                    SmallVectorImpl<ISD::OutputArg> &Outs,
                    const TargetLowering &TLI, const DataLayout &DL);
 
-} // end llvm namespace
+} // end namespace llvm
 
-#endif
+#endif // LLVM_TARGET_TARGETLOWERING_H

contrib/llvm/include/llvm/Target/TargetMachine.h

@@ -20,36 +20,26 @@
 #include "llvm/Pass.h"
 #include "llvm/Support/CodeGen.h"
 #include "llvm/Target/TargetOptions.h"
-#include <cassert>
 #include <string>
 
 namespace llvm {
 
-class InstrItineraryData;
 class GlobalValue;
-class Mangler;
 class MachineFunctionInitializer;
-class MachineModuleInfo;
+class Mangler;
 class MCAsmInfo;
 class MCContext;
 class MCInstrInfo;
 class MCRegisterInfo;
 class MCSubtargetInfo;
 class MCSymbol;
-class Target;
-class TargetLibraryInfo;
-class TargetFrameLowering;
-class TargetIRAnalysis;
-class TargetIntrinsicInfo;
-class TargetLowering;
-class TargetPassConfig;
-class TargetRegisterInfo;
-class TargetSubtargetInfo;
-class TargetTransformInfo;
-class formatted_raw_ostream;
-class raw_ostream;
 class raw_pwrite_stream;
+class Target;
+class TargetIntrinsicInfo;
+class TargetIRAnalysis;
 class TargetLoweringObjectFile;
+class TargetPassConfig;
+class TargetSubtargetInfo;
 
 // The old pass manager infrastructure is hidden in a legacy namespace now.
 namespace legacy {
@@ -64,8 +54,6 @@ using legacy::PassManagerBase;
 /// interface.
 ///
 class TargetMachine {
-  TargetMachine(const TargetMachine &) = delete;
-  void operator=(const TargetMachine &) = delete;
 protected: // Can only create subclasses.
   TargetMachine(const Target &T, StringRef DataLayoutString,
                 const Triple &TargetTriple, StringRef CPU, StringRef FS,
@@ -103,8 +91,11 @@ class TargetMachine {
   unsigned O0WantsFastISel : 1;
 
 public:
+  const TargetOptions DefaultOptions;
   mutable TargetOptions Options;
 
+  TargetMachine(const TargetMachine &) = delete;
+  void operator=(const TargetMachine &) = delete;
   virtual ~TargetMachine();
 
   const Target &getTarget() const { return TheTarget; }
@@ -310,6 +301,6 @@ class LLVMTargetMachine : public TargetMachine {
                          bool DisableVerify = true) override;
 };
 
-} // End llvm namespace
+} // end namespace llvm
 
-#endif
+#endif // LLVM_TARGET_TARGETMACHINE_H

contrib/llvm/include/llvm/Target/TargetSelectionDAG.td

@@ -575,9 +575,6 @@ def intrinsic_w_chain : SDNode<"ISD::INTRINSIC_W_CHAIN",
 def intrinsic_wo_chain : SDNode<"ISD::INTRINSIC_WO_CHAIN",
                                 SDTypeProfile<1, -1, [SDTCisPtrTy<1>]>, []>;
 
-// Do not use cvt directly. Use cvt forms below
-def cvt : SDNode<"ISD::CONVERT_RNDSAT", SDTConvertOp>;
-
 def SDT_assertext : SDTypeProfile<1, 1,
   [SDTCisInt<0>, SDTCisInt<1>, SDTCisSameAs<1, 0>]>;
 def assertsext : SDNode<"ISD::AssertSext", SDT_assertext>;
@@ -1084,54 +1081,6 @@ def atomic_load_64 :
   return cast<AtomicSDNode>(N)->getMemoryVT() == MVT::i64;
 }]>;
 
-//===----------------------------------------------------------------------===//
-// Selection DAG CONVERT_RNDSAT patterns
-
-def cvtff : PatFrag<(ops node:$val, node:$dty, node:$sty, node:$rd, node:$sat),
-    (cvt node:$val, node:$dty, node:$sty, node:$rd, node:$sat), [{
-       return cast<CvtRndSatSDNode>(N)->getCvtCode() == ISD::CVT_FF;
-    }]>;
-
-def cvtss : PatFrag<(ops node:$val, node:$dty, node:$sty, node:$rd, node:$sat),
-    (cvt node:$val, node:$dty, node:$sty, node:$rd, node:$sat), [{
-       return cast<CvtRndSatSDNode>(N)->getCvtCode() == ISD::CVT_SS;
-    }]>;
-
-def cvtsu : PatFrag<(ops node:$val, node:$dty, node:$sty, node:$rd, node:$sat),
-    (cvt node:$val, node:$dty, node:$sty, node:$rd, node:$sat), [{
-       return cast<CvtRndSatSDNode>(N)->getCvtCode() == ISD::CVT_SU;
-    }]>;
-
-def cvtus : PatFrag<(ops node:$val, node:$dty, node:$sty, node:$rd, node:$sat),
-    (cvt node:$val, node:$dty, node:$sty, node:$rd, node:$sat), [{
-       return cast<CvtRndSatSDNode>(N)->getCvtCode() == ISD::CVT_US;
-    }]>;
-
-def cvtuu : PatFrag<(ops node:$val, node:$dty, node:$sty, node:$rd, node:$sat),
-    (cvt node:$val, node:$dty, node:$sty, node:$rd, node:$sat), [{
-       return cast<CvtRndSatSDNode>(N)->getCvtCode() == ISD::CVT_UU;
-    }]>;
-
-def cvtsf : PatFrag<(ops node:$val, node:$dty, node:$sty, node:$rd, node:$sat),
-    (cvt node:$val, node:$dty, node:$sty, node:$rd, node:$sat), [{
-       return cast<CvtRndSatSDNode>(N)->getCvtCode() == ISD::CVT_SF;
-    }]>;
-
-def cvtuf : PatFrag<(ops node:$val, node:$dty, node:$sty, node:$rd, node:$sat),
-    (cvt node:$val, node:$dty, node:$sty, node:$rd, node:$sat), [{
-       return cast<CvtRndSatSDNode>(N)->getCvtCode() == ISD::CVT_UF;
-    }]>;
-
-def cvtfs : PatFrag<(ops node:$val, node:$dty, node:$sty, node:$rd, node:$sat),
-    (cvt node:$val, node:$dty, node:$sty, node:$rd, node:$sat), [{
-       return cast<CvtRndSatSDNode>(N)->getCvtCode() == ISD::CVT_FS;
-    }]>;
-
-def cvtfu : PatFrag<(ops node:$val, node:$dty, node:$sty, node:$rd, node:$sat),
-    (cvt node:$val, node:$dty, node:$sty, node:$rd, node:$sat), [{
-       return cast<CvtRndSatSDNode>(N)->getCvtCode() == ISD::CVT_FU;
-    }]>;
-
 //===----------------------------------------------------------------------===//
 // Selection DAG Pattern Support.
 //

contrib/llvm/include/llvm/Target/TargetSubtargetInfo.h

@@ -14,23 +14,26 @@
 #ifndef LLVM_TARGET_TARGETSUBTARGETINFO_H
 #define LLVM_TARGET_TARGETSUBTARGETINFO_H
 
+#include "llvm/ADT/ArrayRef.h"
+#include "llvm/ADT/SmallVector.h"
+#include "llvm/ADT/StringRef.h"
 #include "llvm/CodeGen/PBQPRAConstraint.h"
 #include "llvm/CodeGen/SchedulerRegistry.h"
 #include "llvm/CodeGen/ScheduleDAGMutation.h"
 #include "llvm/MC/MCSubtargetInfo.h"
 #include "llvm/Support/CodeGen.h"
+#include <memory>
 #include <vector>
 
 namespace llvm {
 
 class CallLowering;
-class DataLayout;
 class InstructionSelector;
 class LegalizerInfo;
-class MachineFunction;
 class MachineInstr;
 class RegisterBankInfo;
 class SDep;
+class SelectionDAGTargetInfo;
 class SUnit;
 class TargetFrameLowering;
 class TargetInstrInfo;
@@ -38,9 +41,7 @@ class TargetLowering;
 class TargetRegisterClass;
 class TargetRegisterInfo;
 class TargetSchedModel;
-class SelectionDAGTargetInfo;
 struct MachineSchedPolicy;
-template <typename T> class SmallVectorImpl;
 
 //===----------------------------------------------------------------------===//
 ///
@@ -49,10 +50,6 @@ template <typename T> class SmallVectorImpl;
 /// be exposed through a TargetSubtargetInfo-derived class.
 ///
 class TargetSubtargetInfo : public MCSubtargetInfo {
-  TargetSubtargetInfo(const TargetSubtargetInfo &) = delete;
-  void operator=(const TargetSubtargetInfo &) = delete;
-  TargetSubtargetInfo() = delete;
-
 protected: // Can only create subclasses...
   TargetSubtargetInfo(const Triple &TT, StringRef CPU, StringRef FS,
                       ArrayRef<SubtargetFeatureKV> PF,
@@ -69,6 +66,9 @@ class TargetSubtargetInfo : public MCSubtargetInfo {
   typedef enum { ANTIDEP_NONE, ANTIDEP_CRITICAL, ANTIDEP_ALL } AntiDepBreakMode;
   typedef SmallVectorImpl<const TargetRegisterClass *> RegClassVector;
 
+  TargetSubtargetInfo() = delete;
+  TargetSubtargetInfo(const TargetSubtargetInfo &) = delete;
+  void operator=(const TargetSubtargetInfo &) = delete;
   virtual ~TargetSubtargetInfo();
 
   virtual bool isXRaySupported() const { return false; }
@@ -229,6 +229,6 @@ class TargetSubtargetInfo : public MCSubtargetInfo {
   virtual bool enableSubRegLiveness() const { return false; }
 };
 
-} // End llvm namespace
+} // end namespace llvm
 
-#endif
+#endif // LLVM_TARGET_TARGETSUBTARGETINFO_H

contrib/llvm/include/llvm/Transforms/Scalar/IVUsersPrinter.h

@@ -0,0 +1,30 @@
+//===- IVUsersPrinter.h - Induction Variable Users Printing -----*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_TRANSFORMS_SCALAR_IVUSERSPRINTER_H
+#define LLVM_TRANSFORMS_SCALAR_IVUSERSPRINTER_H
+
+#include "llvm/Analysis/IVUsers.h"
+#include "llvm/Support/raw_ostream.h"
+#include "llvm/Transforms/Scalar/LoopPassManager.h"
+
+namespace llvm {
+
+/// Printer pass for the \c IVUsers for a loop.
+class IVUsersPrinterPass : public PassInfoMixin<IVUsersPrinterPass> {
+  raw_ostream &OS;
+
+public:
+  explicit IVUsersPrinterPass(raw_ostream &OS) : OS(OS) {}
+  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM,
+                        LoopStandardAnalysisResults &AR, LPMUpdater &U);
+};
+}
+
+#endif

contrib/llvm/include/llvm/Transforms/Scalar/IndVarSimplify.h

@@ -16,14 +16,15 @@
 #define LLVM_TRANSFORMS_SCALAR_INDVARSIMPLIFY_H
 
 #include "llvm/Analysis/LoopInfo.h"
-#include "llvm/Analysis/LoopPassManager.h"
 #include "llvm/IR/PassManager.h"
+#include "llvm/Transforms/Scalar/LoopPassManager.h"
 
 namespace llvm {
 
 class IndVarSimplifyPass : public PassInfoMixin<IndVarSimplifyPass> {
 public:
-  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM);
+  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM,
+                        LoopStandardAnalysisResults &AR, LPMUpdater &U);
 };
 }
 

contrib/llvm/include/llvm/Transforms/Scalar/LICM.h

@@ -34,15 +34,16 @@
 #define LLVM_TRANSFORMS_SCALAR_LICM_H
 
 #include "llvm/Analysis/LoopInfo.h"
-#include "llvm/Analysis/LoopPassManager.h"
 #include "llvm/IR/PassManager.h"
+#include "llvm/Transforms/Scalar/LoopPassManager.h"
 
 namespace llvm {
 
 /// Performs Loop Invariant Code Motion Pass.
 class LICMPass : public PassInfoMixin<LICMPass> {
 public:
-  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM);
+  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM,
+                        LoopStandardAnalysisResults &AR, LPMUpdater &U);
 };
 } // end namespace llvm
 

contrib/llvm/include/llvm/Transforms/Scalar/LoopAccessAnalysisPrinter.h

@@ -0,0 +1,31 @@
+//===- llvm/Analysis/LoopAccessAnalysisPrinter.h ----------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_TRANSFORMS_SCALAR_LOOPACCESSANALYSISPRINTER_H
+#define LLVM_TRANSFORMS_SCALAR_LOOPACCESSANALYSISPRINTER_H
+
+#include "llvm/Support/raw_ostream.h"
+#include "llvm/Transforms/Scalar/LoopPassManager.h"
+
+namespace llvm {
+
+/// \brief Printer pass for the \c LoopAccessInfo results.
+class LoopAccessInfoPrinterPass
+    : public PassInfoMixin<LoopAccessInfoPrinterPass> {
+  raw_ostream &OS;
+
+public:
+  explicit LoopAccessInfoPrinterPass(raw_ostream &OS) : OS(OS) {}
+  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM,
+                        LoopStandardAnalysisResults &AR, LPMUpdater &U);
+};
+
+} // End llvm namespace
+
+#endif

contrib/llvm/include/llvm/Transforms/Scalar/LoopDeletion.h

@@ -15,16 +15,17 @@
 #define LLVM_TRANSFORMS_SCALAR_LOOPDELETION_H
 
 #include "llvm/Analysis/LoopInfo.h"
-#include "llvm/Analysis/LoopPassManager.h"
 #include "llvm/Analysis/ScalarEvolution.h"
 #include "llvm/IR/PassManager.h"
+#include "llvm/Transforms/Scalar/LoopPassManager.h"
 
 namespace llvm {
 
 class LoopDeletionPass : public PassInfoMixin<LoopDeletionPass> {
 public:
   LoopDeletionPass() {}
-  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM);
+  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM,
+                        LoopStandardAnalysisResults &AR, LPMUpdater &U);
   bool runImpl(Loop *L, DominatorTree &DT, ScalarEvolution &SE,
               LoopInfo &loopInfo);
 

contrib/llvm/include/llvm/Transforms/Scalar/LoopIdiomRecognize.h

@@ -17,15 +17,16 @@
 #define LLVM_TRANSFORMS_SCALAR_LOOPIDIOMRECOGNIZE_H
 
 #include "llvm/Analysis/LoopInfo.h"
-#include "llvm/Analysis/LoopPassManager.h"
 #include "llvm/IR/PassManager.h"
+#include "llvm/Transforms/Scalar/LoopPassManager.h"
 
 namespace llvm {
 
 /// Performs Loop Idiom Recognize Pass.
 class LoopIdiomRecognizePass : public PassInfoMixin<LoopIdiomRecognizePass> {
 public:
-  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM);
+  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM,
+                        LoopStandardAnalysisResults &AR, LPMUpdater &U);
 };
 } // end namespace llvm
 

contrib/llvm/include/llvm/Transforms/Scalar/LoopInstSimplify.h

@@ -15,15 +15,16 @@
 #define LLVM_TRANSFORMS_SCALAR_LOOPINSTSIMPLIFY_H
 
 #include "llvm/Analysis/LoopInfo.h"
-#include "llvm/Analysis/LoopPassManager.h"
 #include "llvm/IR/PassManager.h"
+#include "llvm/Transforms/Scalar/LoopPassManager.h"
 
 namespace llvm {
 
 /// Performs Loop Inst Simplify Pass.
 class LoopInstSimplifyPass : public PassInfoMixin<LoopInstSimplifyPass> {
 public:
-  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM);
+  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM,
+                        LoopStandardAnalysisResults &AR, LPMUpdater &U);
 };
 } // end namespace llvm
 

contrib/llvm/include/llvm/Transforms/Scalar/LoopPassManager.h

@@ -0,0 +1,363 @@
+//===- LoopPassManager.h - Loop pass management -----------------*- C++ -*-===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+/// \file
+///
+/// This header provides classes for managing a pipeline of passes over loops
+/// in LLVM IR.
+///
+/// The primary loop pass pipeline is managed in a very particular way to
+/// provide a set of core guarantees:
+/// 1) Loops are, where possible, in simplified form.
+/// 2) Loops are *always* in LCSSA form.
+/// 3) A collection of Loop-specific analysis results are available:
+///    - LoopInfo
+///    - DominatorTree
+///    - ScalarEvolution
+///    - AAManager
+/// 4) All loop passes preserve #1 (where possible), #2, and #3.
+/// 5) Loop passes run over each loop in the loop nest from the innermost to
+///    the outermost. Specifically, all inner loops are processed before
+///    passes run over outer loops. When running the pipeline across an inner
+///    loop creates new inner loops, those are added and processed in this
+///    order as well.
+///
+/// This process is designed to facilitate transformations which simplify,
+/// reduce, and remove loops. For passes which are more oriented towards
+/// optimizing loops, especially optimizing loop *nests* instead of single
+/// loops in isolation, this framework is less interesting.
+///
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_TRANSFORMS_SCALAR_LOOPPASSMANAGER_H
+#define LLVM_TRANSFORMS_SCALAR_LOOPPASSMANAGER_H
+
+#include "llvm/ADT/PostOrderIterator.h"
+#include "llvm/ADT/PriorityWorklist.h"
+#include "llvm/ADT/STLExtras.h"
+#include "llvm/Analysis/AliasAnalysis.h"
+#include "llvm/Analysis/BasicAliasAnalysis.h"
+#include "llvm/Analysis/GlobalsModRef.h"
+#include "llvm/Analysis/LoopAnalysisManager.h"
+#include "llvm/Analysis/LoopInfo.h"
+#include "llvm/Analysis/ScalarEvolution.h"
+#include "llvm/Analysis/ScalarEvolutionAliasAnalysis.h"
+#include "llvm/Analysis/TargetLibraryInfo.h"
+#include "llvm/Analysis/TargetTransformInfo.h"
+#include "llvm/IR/Dominators.h"
+#include "llvm/IR/PassManager.h"
+
+namespace llvm {
+
+// Forward declarations of an update tracking API used in the pass manager.
+class LPMUpdater;
+
+// Explicit specialization and instantiation declarations for the pass manager.
+// See the comments on the definition of the specialization for details on how
+// it differs from the primary template.
+template <>
+PreservedAnalyses
+PassManager<Loop, LoopAnalysisManager, LoopStandardAnalysisResults &,
+            LPMUpdater &>::run(Loop &InitialL, LoopAnalysisManager &AM,
+                               LoopStandardAnalysisResults &AnalysisResults,
+                               LPMUpdater &U);
+extern template class PassManager<Loop, LoopAnalysisManager,
+                                  LoopStandardAnalysisResults &, LPMUpdater &>;
+
+/// \brief The Loop pass manager.
+///
+/// See the documentation for the PassManager template for details. It runs
+/// a sequence of Loop passes over each Loop that the manager is run over. This
+/// typedef serves as a convenient way to refer to this construct.
+typedef PassManager<Loop, LoopAnalysisManager, LoopStandardAnalysisResults &,
+                    LPMUpdater &>
+    LoopPassManager;
+
+/// A partial specialization of the require analysis template pass to forward
+/// the extra parameters from a transformation's run method to the
+/// AnalysisManager's getResult.
+template <typename AnalysisT>
+struct RequireAnalysisPass<AnalysisT, Loop, LoopAnalysisManager,
+                           LoopStandardAnalysisResults &, LPMUpdater &>
+    : PassInfoMixin<
+          RequireAnalysisPass<AnalysisT, Loop, LoopAnalysisManager,
+                              LoopStandardAnalysisResults &, LPMUpdater &>> {
+  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM,
+                        LoopStandardAnalysisResults &AR, LPMUpdater &) {
+    (void)AM.template getResult<AnalysisT>(L, AR);
+    return PreservedAnalyses::all();
+  }
+};
+
+/// An alias template to easily name a require analysis loop pass.
+template <typename AnalysisT>
+using RequireAnalysisLoopPass =
+    RequireAnalysisPass<AnalysisT, Loop, LoopAnalysisManager,
+                        LoopStandardAnalysisResults &, LPMUpdater &>;
+
+namespace internal {
+/// Helper to implement appending of loops onto a worklist.
+///
+/// We want to process loops in postorder, but the worklist is a LIFO data
+/// structure, so we append to it in *reverse* postorder.
+///
+/// For trees, a preorder traversal is a viable reverse postorder, so we
+/// actually append using a preorder walk algorithm.
+template <typename RangeT>
+inline void appendLoopsToWorklist(RangeT &&Loops,
+                                  SmallPriorityWorklist<Loop *, 4> &Worklist) {
+  // We use an internal worklist to build up the preorder traversal without
+  // recursion.
+  SmallVector<Loop *, 4> PreOrderLoops, PreOrderWorklist;
+
+  // We walk the initial sequence of loops in reverse because we generally want
+  // to visit defs before uses and the worklist is LIFO.
+  for (Loop *RootL : reverse(Loops)) {
+    assert(PreOrderLoops.empty() && "Must start with an empty preorder walk.");
+    assert(PreOrderWorklist.empty() &&
+           "Must start with an empty preorder walk worklist.");
+    PreOrderWorklist.push_back(RootL);
+    do {
+      Loop *L = PreOrderWorklist.pop_back_val();
+      PreOrderWorklist.append(L->begin(), L->end());
+      PreOrderLoops.push_back(L);
+    } while (!PreOrderWorklist.empty());
+
+    Worklist.insert(std::move(PreOrderLoops));
+    PreOrderLoops.clear();
+  }
+}
+}
+
+template <typename LoopPassT> class FunctionToLoopPassAdaptor;
+
+/// This class provides an interface for updating the loop pass manager based
+/// on mutations to the loop nest.
+///
+/// A reference to an instance of this class is passed as an argument to each
+/// Loop pass, and Loop passes should use it to update LPM infrastructure if
+/// they modify the loop nest structure.
+class LPMUpdater {
+public:
+  /// This can be queried by loop passes which run other loop passes (like pass
+  /// managers) to know whether the loop needs to be skipped due to updates to
+  /// the loop nest.
+  ///
+  /// If this returns true, the loop object may have been deleted, so passes
+  /// should take care not to touch the object.
+  bool skipCurrentLoop() const { return SkipCurrentLoop; }
+
+  /// Loop passes should use this method to indicate they have deleted a loop
+  /// from the nest.
+  ///
+  /// Note that this loop must either be the current loop or a subloop of the
+  /// current loop. This routine must be called prior to removing the loop from
+  /// the loop nest.
+  ///
+  /// If this is called for the current loop, in addition to clearing any
+  /// state, this routine will mark that the current loop should be skipped by
+  /// the rest of the pass management infrastructure.
+  void markLoopAsDeleted(Loop &L) {
+    LAM.clear(L);
+    assert(CurrentL->contains(&L) && "Cannot delete a loop outside of the "
+                                     "subloop tree currently being processed.");
+    if (&L == CurrentL)
+      SkipCurrentLoop = true;
+  }
+
+  /// Loop passes should use this method to indicate they have added new child
+  /// loops of the current loop.
+  ///
+  /// \p NewChildLoops must contain only the immediate children. Any nested
+  /// loops within them will be visited in postorder as usual for the loop pass
+  /// manager.
+  void addChildLoops(ArrayRef<Loop *> NewChildLoops) {
+    // Insert ourselves back into the worklist first, as this loop should be
+    // revisited after all the children have been processed.
+    Worklist.insert(CurrentL);
+
+#ifndef NDEBUG
+    for (Loop *NewL : NewChildLoops)
+      assert(NewL->getParentLoop() == CurrentL && "All of the new loops must "
+                                                  "be immediate children of "
+                                                  "the current loop!");
+#endif
+
+    internal::appendLoopsToWorklist(NewChildLoops, Worklist);
+
+    // Also skip further processing of the current loop--it will be revisited
+    // after all of its newly added children are accounted for.
+    SkipCurrentLoop = true;
+  }
+
+  /// Loop passes should use this method to indicate they have added new
+  /// sibling loops to the current loop.
+  ///
+  /// \p NewSibLoops must only contain the immediate sibling loops. Any nested
+  /// loops within them will be visited in postorder as usual for the loop pass
+  /// manager.
+  void addSiblingLoops(ArrayRef<Loop *> NewSibLoops) {
+#ifndef NDEBUG
+    for (Loop *NewL : NewSibLoops)
+      assert(NewL->getParentLoop() == ParentL &&
+             "All of the new loops must be siblings of the current loop!");
+#endif
+
+    internal::appendLoopsToWorklist(NewSibLoops, Worklist);
+
+    // No need to skip the current loop or revisit it, as sibling loops
+    // shouldn't impact anything.
+  }
+
+private:
+  template <typename LoopPassT> friend class llvm::FunctionToLoopPassAdaptor;
+
+  /// The \c FunctionToLoopPassAdaptor's worklist of loops to process.
+  SmallPriorityWorklist<Loop *, 4> &Worklist;
+
+  /// The analysis manager for use in the current loop nest.
+  LoopAnalysisManager &LAM;
+
+  Loop *CurrentL;
+  bool SkipCurrentLoop;
+
+#ifndef NDEBUG
+  // In debug builds we also track the parent loop to implement asserts even in
+  // the face of loop deletion.
+  Loop *ParentL;
+#endif
+
+  LPMUpdater(SmallPriorityWorklist<Loop *, 4> &Worklist,
+             LoopAnalysisManager &LAM)
+      : Worklist(Worklist), LAM(LAM) {}
+};
+
+/// \brief Adaptor that maps from a function to its loops.
+///
+/// Designed to allow composition of a LoopPass(Manager) and a
+/// FunctionPassManager. Note that if this pass is constructed with a \c
+/// FunctionAnalysisManager it will run the \c LoopAnalysisManagerFunctionProxy
+/// analysis prior to running the loop passes over the function to enable a \c
+/// LoopAnalysisManager to be used within this run safely.
+template <typename LoopPassT>
+class FunctionToLoopPassAdaptor
+    : public PassInfoMixin<FunctionToLoopPassAdaptor<LoopPassT>> {
+public:
+  explicit FunctionToLoopPassAdaptor(LoopPassT Pass) : Pass(std::move(Pass)) {}
+
+  /// \brief Runs the loop passes across every loop in the function.
+  PreservedAnalyses run(Function &F, FunctionAnalysisManager &AM) {
+    // Setup the loop analysis manager from its proxy.
+    LoopAnalysisManager &LAM =
+        AM.getResult<LoopAnalysisManagerFunctionProxy>(F).getManager();
+    // Get the loop structure for this function
+    LoopInfo &LI = AM.getResult<LoopAnalysis>(F);
+
+    // If there are no loops, there is nothing to do here.
+    if (LI.empty())
+      return PreservedAnalyses::all();
+
+    // Get the analysis results needed by loop passes.
+    LoopStandardAnalysisResults LAR = {AM.getResult<AAManager>(F),
+                                       AM.getResult<AssumptionAnalysis>(F),
+                                       AM.getResult<DominatorTreeAnalysis>(F),
+                                       AM.getResult<LoopAnalysis>(F),
+                                       AM.getResult<ScalarEvolutionAnalysis>(F),
+                                       AM.getResult<TargetLibraryAnalysis>(F),
+                                       AM.getResult<TargetIRAnalysis>(F)};
+
+    PreservedAnalyses PA = PreservedAnalyses::all();
+
+    // A postorder worklist of loops to process.
+    SmallPriorityWorklist<Loop *, 4> Worklist;
+
+    // Register the worklist and loop analysis manager so that loop passes can
+    // update them when they mutate the loop nest structure.
+    LPMUpdater Updater(Worklist, LAM);
+
+    // Add the loop nests in the reverse order of LoopInfo. For some reason,
+    // they are stored in RPO w.r.t. the control flow graph in LoopInfo. For
+    // the purpose of unrolling, loop deletion, and LICM, we largely want to
+    // work forward across the CFG so that we visit defs before uses and can
+    // propagate simplifications from one loop nest into the next.
+    // FIXME: Consider changing the order in LoopInfo.
+    internal::appendLoopsToWorklist(reverse(LI), Worklist);
+
+    do {
+      Loop *L = Worklist.pop_back_val();
+
+      // Reset the update structure for this loop.
+      Updater.CurrentL = L;
+      Updater.SkipCurrentLoop = false;
+#ifndef NDEBUG
+      Updater.ParentL = L->getParentLoop();
+#endif
+
+      PreservedAnalyses PassPA = Pass.run(*L, LAM, LAR, Updater);
+      // FIXME: We should verify the set of analyses relevant to Loop passes
+      // are preserved.
+
+      // If the loop hasn't been deleted, we need to handle invalidation here.
+      if (!Updater.skipCurrentLoop())
+        // We know that the loop pass couldn't have invalidated any other
+        // loop's analyses (that's the contract of a loop pass), so directly
+        // handle the loop analysis manager's invalidation here.
+        LAM.invalidate(*L, PassPA);
+
+      // Then intersect the preserved set so that invalidation of module
+      // analyses will eventually occur when the module pass completes.
+      PA.intersect(std::move(PassPA));
+    } while (!Worklist.empty());
+
+    // By definition we preserve the proxy. We also preserve all analyses on
+    // Loops. This precludes *any* invalidation of loop analyses by the proxy,
+    // but that's OK because we've taken care to invalidate analyses in the
+    // loop analysis manager incrementally above.
+    PA.preserveSet<AllAnalysesOn<Loop>>();
+    PA.preserve<LoopAnalysisManagerFunctionProxy>();
+    // We also preserve the set of standard analyses.
+    PA.preserve<AssumptionAnalysis>();
+    PA.preserve<DominatorTreeAnalysis>();
+    PA.preserve<LoopAnalysis>();
+    PA.preserve<ScalarEvolutionAnalysis>();
+    // FIXME: What we really want to do here is preserve an AA category, but
+    // that concept doesn't exist yet.
+    PA.preserve<AAManager>();
+    PA.preserve<BasicAA>();
+    PA.preserve<GlobalsAA>();
+    PA.preserve<SCEVAA>();
+    return PA;
+  }
+
+private:
+  LoopPassT Pass;
+};
+
+/// \brief A function to deduce a loop pass type and wrap it in the templated
+/// adaptor.
+template <typename LoopPassT>
+FunctionToLoopPassAdaptor<LoopPassT>
+createFunctionToLoopPassAdaptor(LoopPassT Pass) {
+  return FunctionToLoopPassAdaptor<LoopPassT>(std::move(Pass));
+}
+
+/// \brief Pass for printing a loop's contents as textual IR.
+class PrintLoopPass : public PassInfoMixin<PrintLoopPass> {
+  raw_ostream &OS;
+  std::string Banner;
+
+public:
+  PrintLoopPass();
+  PrintLoopPass(raw_ostream &OS, const std::string &Banner = "");
+
+  PreservedAnalyses run(Loop &L, LoopAnalysisManager &,
+                        LoopStandardAnalysisResults &, LPMUpdater &);
+};
+}
+
+#endif // LLVM_TRANSFORMS_SCALAR_LOOPPASSMANAGER_H

contrib/llvm/include/llvm/Transforms/Scalar/LoopRotation.h

@@ -15,8 +15,8 @@
 #define LLVM_TRANSFORMS_SCALAR_LOOPROTATION_H
 
 #include "llvm/Analysis/LoopInfo.h"
-#include "llvm/Analysis/LoopPassManager.h"
 #include "llvm/IR/PassManager.h"
+#include "llvm/Transforms/Scalar/LoopPassManager.h"
 
 namespace llvm {
 
@@ -24,7 +24,8 @@ namespace llvm {
 class LoopRotatePass : public PassInfoMixin<LoopRotatePass> {
 public:
   LoopRotatePass(bool EnableHeaderDuplication = true);
-  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM);
+  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM,
+                        LoopStandardAnalysisResults &AR, LPMUpdater &U);
 
 private:
   const bool EnableHeaderDuplication;

contrib/llvm/include/llvm/Transforms/Scalar/LoopSimplifyCFG.h

@@ -18,15 +18,16 @@
 #define LLVM_TRANSFORMS_SCALAR_LOOPSIMPLIFYCFG_H
 
 #include "llvm/Analysis/LoopInfo.h"
-#include "llvm/Analysis/LoopPassManager.h"
 #include "llvm/IR/PassManager.h"
+#include "llvm/Transforms/Scalar/LoopPassManager.h"
 
 namespace llvm {
 
 /// Performs basic CFG simplifications to assist other loop passes.
 class LoopSimplifyCFGPass : public PassInfoMixin<LoopSimplifyCFGPass> {
 public:
-  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM);
+  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM,
+                        LoopStandardAnalysisResults &AR, LPMUpdater &U);
 };
 } // end namespace llvm
 

contrib/llvm/include/llvm/Transforms/Scalar/LoopStrengthReduce.h

@@ -23,15 +23,16 @@
 #define LLVM_TRANSFORMS_SCALAR_LOOPSTRENGTHREDUCE_H
 
 #include "llvm/Analysis/LoopInfo.h"
-#include "llvm/Analysis/LoopPassManager.h"
 #include "llvm/IR/PassManager.h"
+#include "llvm/Transforms/Scalar/LoopPassManager.h"
 
 namespace llvm {
 
 /// Performs Loop Strength Reduce Pass.
 class LoopStrengthReducePass : public PassInfoMixin<LoopStrengthReducePass> {
 public:
-  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM);
+  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM,
+                        LoopStandardAnalysisResults &AR, LPMUpdater &U);
 };
 } // end namespace llvm
 

contrib/llvm/include/llvm/Transforms/Scalar/LoopUnrollPass.h

@@ -11,8 +11,8 @@
 #define LLVM_TRANSFORMS_SCALAR_LOOPUNROLLPASS_H
 
 #include "llvm/Analysis/LoopInfo.h"
-#include "llvm/Analysis/LoopPassManager.h"
 #include "llvm/IR/PassManager.h"
+#include "llvm/Transforms/Scalar/LoopPassManager.h"
 
 namespace llvm {
 
@@ -23,7 +23,8 @@ struct LoopUnrollPass : public PassInfoMixin<LoopUnrollPass> {
   Optional<bool> ProvidedRuntime;
   Optional<bool> ProvidedUpperBound;
 
-  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM);
+  PreservedAnalyses run(Loop &L, LoopAnalysisManager &AM,
+                        LoopStandardAnalysisResults &AR, LPMUpdater &U);
 };
 } // end namespace llvm
 

contrib/llvm/include/llvm/Transforms/Utils/LoopUtils.h

@@ -29,6 +29,7 @@ class DataLayout;
 class DominatorTree;
 class Loop;
 class LoopInfo;
+class OptimizationRemarkEmitter;
 class Pass;
 class PredicatedScalarEvolution;
 class PredIteratorCache;
@@ -404,11 +405,11 @@ bool formLCSSARecursively(Loop &L, DominatorTree &DT, LoopInfo *LI,
 /// uses before definitions, allowing us to sink a loop body in one pass without
 /// iteration. Takes DomTreeNode, AliasAnalysis, LoopInfo, DominatorTree,
 /// DataLayout, TargetLibraryInfo, Loop, AliasSet information for all
-/// instructions of the loop and loop safety information as arguments.
-/// It returns changed status.
+/// instructions of the loop and loop safety information as
+/// arguments. Diagnostics is emitted via \p ORE. It returns changed status.
 bool sinkRegion(DomTreeNode *, AliasAnalysis *, LoopInfo *, DominatorTree *,
                 TargetLibraryInfo *, Loop *, AliasSetTracker *,
-                LoopSafetyInfo *);
+                LoopSafetyInfo *, OptimizationRemarkEmitter *ORE);
 
 /// \brief Walk the specified region of the CFG (defined by all blocks
 /// dominated by the specified block, and that are in the current loop) in depth
@@ -416,10 +417,11 @@ bool sinkRegion(DomTreeNode *, AliasAnalysis *, LoopInfo *, DominatorTree *,
 /// before uses, allowing us to hoist a loop body in one pass without iteration.
 /// Takes DomTreeNode, AliasAnalysis, LoopInfo, DominatorTree, DataLayout,
 /// TargetLibraryInfo, Loop, AliasSet information for all instructions of the
-/// loop and loop safety information as arguments. It returns changed status.
+/// loop and loop safety information as arguments. Diagnostics is emitted via \p
+/// ORE. It returns changed status.
 bool hoistRegion(DomTreeNode *, AliasAnalysis *, LoopInfo *, DominatorTree *,
                  TargetLibraryInfo *, Loop *, AliasSetTracker *,
-                 LoopSafetyInfo *);
+                 LoopSafetyInfo *, OptimizationRemarkEmitter *ORE);
 
 /// \brief Try to promote memory values to scalars by sinking stores out of
 /// the loop and moving loads to before the loop.  We do this by looping over
@@ -427,12 +429,14 @@ bool hoistRegion(DomTreeNode *, AliasAnalysis *, LoopInfo *, DominatorTree *,
 /// loop invariant. It takes AliasSet, Loop exit blocks vector, loop exit blocks
 /// insertion point vector, PredIteratorCache, LoopInfo, DominatorTree, Loop,
 /// AliasSet information for all instructions of the loop and loop safety
-/// information as arguments. It returns changed status.
+/// information as arguments. Diagnostics is emitted via \p ORE. It returns
+/// changed status.
 bool promoteLoopAccessesToScalars(AliasSet &, SmallVectorImpl<BasicBlock *> &,
                                   SmallVectorImpl<Instruction *> &,
                                   PredIteratorCache &, LoopInfo *,
                                   DominatorTree *, const TargetLibraryInfo *,
-                                  Loop *, AliasSetTracker *, LoopSafetyInfo *);
+                                  Loop *, AliasSetTracker *, LoopSafetyInfo *,
+                                  OptimizationRemarkEmitter *);
 
 /// \brief Computes safety information for a loop
 /// checks loop body & header for the possibility of may throw
@@ -478,11 +482,12 @@ void getLoopAnalysisUsage(AnalysisUsage &AU);
 /// preheader to loop body (no speculation).
 /// If SafetyInfo is not null, we are checking for hoisting/sinking
 /// instructions from loop body to preheader/exit. Check if the instruction
-/// can execute specultatively.
-///
+/// can execute speculatively.
+/// If \p ORE is set use it to emit optimization remarks.
 bool canSinkOrHoistInst(Instruction &I, AAResults *AA, DominatorTree *DT,
                         Loop *CurLoop, AliasSetTracker *CurAST,
-                        LoopSafetyInfo *SafetyInfo);
+                        LoopSafetyInfo *SafetyInfo,
+                        OptimizationRemarkEmitter *ORE = nullptr);
 }
 
 #endif

contrib/llvm/include/llvm/Transforms/Utils/UnrollLoop.h

@@ -33,6 +33,12 @@ class Pass;
 class OptimizationRemarkEmitter;
 class ScalarEvolution;
 
+typedef SmallDenseMap<const Loop *, Loop *, 4> NewLoopsMap;
+
+const Loop* addClonedBlockToLoopInfo(BasicBlock *OriginalBB,
+                                     BasicBlock *ClonedBB, LoopInfo *LI,
+                                     NewLoopsMap &NewLoops);
+
 bool UnrollLoop(Loop *L, unsigned Count, unsigned TripCount, bool Force,
                 bool AllowRuntime, bool AllowExpensiveTripCount,
                 bool PreserveCondBr, bool PreserveOnlyFirst,

contrib/llvm/include/llvm/Transforms/Vectorize/LoopVectorize.h

@@ -57,12 +57,12 @@
 #include "llvm/Analysis/DemandedBits.h"
 #include "llvm/Analysis/LoopAccessAnalysis.h"
 #include "llvm/Analysis/LoopInfo.h"
-#include "llvm/Analysis/LoopPassManager.h"
 #include "llvm/Analysis/OptimizationDiagnosticInfo.h"
 #include "llvm/Analysis/ScalarEvolution.h"
 #include "llvm/Analysis/TargetTransformInfo.h"
 #include "llvm/IR/Function.h"
 #include "llvm/IR/PassManager.h"
+#include "llvm/Transforms/Scalar/LoopPassManager.h"
 #include <functional>
 
 namespace llvm {

contrib/llvm/include/llvm/XRay/Trace.h

@@ -0,0 +1,71 @@
+//===- Trace.h - XRay Trace Abstraction -----------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// Defines the XRay Trace class representing records in an XRay trace file.
+//
+//===----------------------------------------------------------------------===//
+#ifndef LLVM_XRAY_TRACE_H
+#define LLVM_XRAY_TRACE_H
+
+#include <cstdint>
+#include <vector>
+
+#include "llvm/ADT/StringRef.h"
+#include "llvm/Support/Error.h"
+#include "llvm/Support/FileSystem.h"
+#include "llvm/XRay/XRayRecord.h"
+
+namespace llvm {
+namespace xray {
+
+/// A Trace object represents the records that have been loaded from XRay
+/// log files generated by instrumented binaries. We encapsulate the logic of
+/// reading the traces in factory functions that populate the Trace object
+/// appropriately.
+///
+/// Trace objects provide an accessor to an XRayFileHeader which says more about
+/// details of the file from which the XRay trace was loaded from.
+///
+/// Usage:
+///
+///   if (auto TraceOrErr = loadTraceFile("xray-log.something.xray")) {
+///     auto& T = *TraceOrErr;
+///     // T.getFileHeader() will provide information from the trace header.
+///     for (const XRayRecord &R : T) {
+///       // ... do something with R here.
+///     }
+///   } else {
+///     // Handle the error here.
+///   }
+///
+class Trace {
+  XRayFileHeader FileHeader;
+  std::vector<XRayRecord> Records;
+
+  typedef std::vector<XRayRecord>::const_iterator citerator;
+
+  friend Expected<Trace> loadTraceFile(StringRef, bool);
+
+public:
+  /// Provides access to the loaded XRay trace file header.
+  const XRayFileHeader &getFileHeader() const { return FileHeader; }
+
+  citerator begin() const { return Records.begin(); }
+  citerator end() const { return Records.end(); }
+  size_t size() const { return Records.size(); }
+};
+
+/// This function will attempt to load XRay trace records from the provided
+/// |Filename|.
+Expected<Trace> loadTraceFile(StringRef Filename, bool Sort = false);
+
+} // namespace xray
+} // namespace llvm
+
+#endif // LLVM_XRAY_TRACE_H

contrib/llvm/include/llvm/XRay/XRayRecord.h

@@ -0,0 +1,76 @@
+//===- XRayRecord.h - XRay Trace Record -----------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// This file replicates the record definition for XRay log entries. This should
+// follow the evolution of the log record versions supported in the compiler-rt
+// xray project.
+//
+//===----------------------------------------------------------------------===//
+#ifndef LLVM_XRAY_XRAY_RECORD_H
+#define LLVM_XRAY_XRAY_RECORD_H
+
+#include <cstdint>
+
+namespace llvm {
+namespace xray {
+
+/// XRay traces all have a header providing some top-matter information useful
+/// to help tools determine how to interpret the information available in the
+/// trace.
+struct XRayFileHeader {
+  /// Version of the XRay implementation that produced this file.
+  uint16_t Version = 0;
+
+  /// A numeric identifier for the type of file this is. Best used in
+  /// combination with Version.
+  uint16_t Type = 0;
+
+  /// Whether the CPU that produced the timestamp counters (TSC) move at a
+  /// constant rate.
+  bool ConstantTSC;
+
+  /// Whether the CPU that produced the timestamp counters (TSC) do not stop.
+  bool NonstopTSC;
+
+  /// The number of cycles per second for the CPU that produced the timestamp
+  /// counter (TSC) values. Useful for estimating the amount of time that
+  /// elapsed between two TSCs on some platforms.
+  uint64_t CycleFrequency = 0;
+};
+
+/// Determines the supported types of records that could be seen in XRay traces.
+/// This may or may not correspond to actual record types in the raw trace (as
+/// the loader implementation may synthesize this information in the process of
+/// of loading).
+enum class RecordTypes { ENTER, EXIT };
+
+struct XRayRecord {
+  /// The type of record.
+  uint16_t RecordType;
+
+  /// The CPU where the thread is running. We assume number of CPUs <= 256.
+  uint8_t CPU;
+
+  /// Identifies the type of record.
+  RecordTypes Type;
+
+  /// The function ID for the record.
+  int32_t FuncId;
+
+  /// Get the full 8 bytes of the TSC when we get the log record.
+  uint64_t TSC;
+
+  /// The thread ID for the currently running thread.
+  uint32_t TId;
+};
+
+} // namespace xray
+} // namespace llvm
+
+#endif // LLVM_XRAY_XRAY_RECORD_H

contrib/llvm/include/llvm/XRay/YAMLXRayRecord.h

@@ -0,0 +1,99 @@
+//===- YAMLXRayRecord.h - XRay Record YAML Support Definitions ------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// Types and traits specialisations for YAML I/O of XRay log entries.
+//
+//===----------------------------------------------------------------------===//
+#ifndef LLVM_XRAY_YAML_XRAY_RECORD_H
+#define LLVM_XRAY_YAML_XRAY_RECORD_H
+
+#include <type_traits>
+
+#include "llvm/Support/YAMLTraits.h"
+#include "llvm/XRay/XRayRecord.h"
+
+namespace llvm {
+namespace xray {
+
+struct YAMLXRayFileHeader {
+  uint16_t Version;
+  uint16_t Type;
+  bool ConstantTSC;
+  bool NonstopTSC;
+  uint64_t CycleFrequency;
+};
+
+struct YAMLXRayRecord {
+  uint16_t RecordType;
+  uint8_t CPU;
+  RecordTypes Type;
+  int32_t FuncId;
+  std::string Function;
+  uint64_t TSC;
+  uint32_t TId;
+};
+
+struct YAMLXRayTrace {
+  YAMLXRayFileHeader Header;
+  std::vector<YAMLXRayRecord> Records;
+};
+
+} // namespace xray
+
+namespace yaml {
+
+// YAML Traits
+// -----------
+template <> struct ScalarEnumerationTraits<xray::RecordTypes> {
+  static void enumeration(IO &IO, xray::RecordTypes &Type) {
+    IO.enumCase(Type, "function-enter", xray::RecordTypes::ENTER);
+    IO.enumCase(Type, "function-exit", xray::RecordTypes::EXIT);
+  }
+};
+
+template <> struct MappingTraits<xray::YAMLXRayFileHeader> {
+  static void mapping(IO &IO, xray::YAMLXRayFileHeader &Header) {
+    IO.mapRequired("version", Header.Version);
+    IO.mapRequired("type", Header.Type);
+    IO.mapRequired("constant-tsc", Header.ConstantTSC);
+    IO.mapRequired("nonstop-tsc", Header.NonstopTSC);
+    IO.mapRequired("cycle-frequency", Header.CycleFrequency);
+  }
+};
+
+template <> struct MappingTraits<xray::YAMLXRayRecord> {
+  static void mapping(IO &IO, xray::YAMLXRayRecord &Record) {
+    // FIXME: Make this type actually be descriptive
+    IO.mapRequired("type", Record.RecordType);
+    IO.mapRequired("func-id", Record.FuncId);
+    IO.mapOptional("function", Record.Function);
+    IO.mapRequired("cpu", Record.CPU);
+    IO.mapRequired("thread", Record.TId);
+    IO.mapRequired("kind", Record.Type);
+    IO.mapRequired("tsc", Record.TSC);
+  }
+
+  static constexpr bool flow = true;
+};
+
+template <> struct MappingTraits<xray::YAMLXRayTrace> {
+  static void mapping(IO &IO, xray::YAMLXRayTrace &Trace) {
+    // A trace file contains two parts, the header and the list of all the
+    // trace records.
+    IO.mapRequired("header", Trace.Header);
+    IO.mapRequired("records", Trace.Records);
+  }
+};
+
+} // namespace yaml
+} // namespace llvm
+
+LLVM_YAML_IS_SEQUENCE_VECTOR(xray::YAMLXRayRecord)
+
+#endif // LLVM_XRAY_YAML_XRAY_RECORD_H

contrib/llvm/lib/Analysis/AssumptionCache.cpp

@@ -24,6 +24,109 @@
 using namespace llvm;
 using namespace llvm::PatternMatch;
 
+SmallVector<WeakVH, 1> &AssumptionCache::getAffectedValues(Value *V) {
+  // Try using find_as first to avoid creating extra value handles just for the
+  // purpose of doing the lookup.
+  auto AVI = AffectedValues.find_as(V);
+  if (AVI != AffectedValues.end())
+    return AVI->second;
+
+  auto AVIP = AffectedValues.insert({
+      AffectedValueCallbackVH(V, this), SmallVector<WeakVH, 1>()});
+  return AVIP.first->second;
+}
+
+void AssumptionCache::updateAffectedValues(CallInst *CI) {
+  // Note: This code must be kept in-sync with the code in
+  // computeKnownBitsFromAssume in ValueTracking.
+
+  SmallVector<Value *, 16> Affected;
+  auto AddAffected = [&Affected](Value *V) {
+    if (isa<Argument>(V)) {
+      Affected.push_back(V);
+    } else if (auto *I = dyn_cast<Instruction>(V)) {
+      Affected.push_back(I);
+
+      if (I->getOpcode() == Instruction::BitCast ||
+          I->getOpcode() == Instruction::PtrToInt) {
+        auto *Op = I->getOperand(0);
+        if (isa<Instruction>(Op) || isa<Argument>(Op))
+          Affected.push_back(Op);
+      }
+    }
+  };
+
+  Value *Cond = CI->getArgOperand(0), *A, *B;
+  AddAffected(Cond);
+
+  CmpInst::Predicate Pred;
+  if (match(Cond, m_ICmp(Pred, m_Value(A), m_Value(B)))) {
+    AddAffected(A);
+    AddAffected(B);
+
+    if (Pred == ICmpInst::ICMP_EQ) {
+      // For equality comparisons, we handle the case of bit inversion.
+      auto AddAffectedFromEq = [&AddAffected](Value *V) {
+        Value *A;
+        if (match(V, m_Not(m_Value(A)))) {
+          AddAffected(A);
+          V = A;
+        }
+
+        Value *B;
+        ConstantInt *C;
+        // (A & B) or (A | B) or (A ^ B).
+        if (match(V,
+                  m_CombineOr(m_And(m_Value(A), m_Value(B)),
+                    m_CombineOr(m_Or(m_Value(A), m_Value(B)),
+                                m_Xor(m_Value(A), m_Value(B)))))) {
+          AddAffected(A);
+          AddAffected(B);
+        // (A << C) or (A >>_s C) or (A >>_u C) where C is some constant.
+        } else if (match(V,
+                         m_CombineOr(m_Shl(m_Value(A), m_ConstantInt(C)),
+                           m_CombineOr(m_LShr(m_Value(A), m_ConstantInt(C)),
+                                       m_AShr(m_Value(A),
+                                              m_ConstantInt(C)))))) {
+          AddAffected(A);
+        }
+      };
+
+      AddAffectedFromEq(A);
+      AddAffectedFromEq(B);
+    }
+  }
+
+  for (auto &AV : Affected) {
+    auto &AVV = getAffectedValues(AV);
+    if (std::find(AVV.begin(), AVV.end(), CI) == AVV.end())
+      AVV.push_back(CI);
+  }
+}
+
+void AssumptionCache::AffectedValueCallbackVH::deleted() {
+  auto AVI = AC->AffectedValues.find(getValPtr());
+  if (AVI != AC->AffectedValues.end())
+    AC->AffectedValues.erase(AVI);
+  // 'this' now dangles!
+}
+
+void AssumptionCache::AffectedValueCallbackVH::allUsesReplacedWith(Value *NV) {
+  if (!isa<Instruction>(NV) && !isa<Argument>(NV))
+    return;
+
+  // Any assumptions that affected this value now affect the new value.
+
+  auto &NAVV = AC->getAffectedValues(NV);
+  auto AVI = AC->AffectedValues.find(getValPtr());
+  if (AVI == AC->AffectedValues.end())
+    return;
+
+  for (auto &A : AVI->second)
+    if (std::find(NAVV.begin(), NAVV.end(), A) == NAVV.end())
+      NAVV.push_back(A);
+}
+
 void AssumptionCache::scanFunction() {
   assert(!Scanned && "Tried to scan the function twice!");
   assert(AssumeHandles.empty() && "Already have assumes when scanning!");
@@ -37,6 +140,10 @@ void AssumptionCache::scanFunction() {
 
   // Mark the scan as complete.
   Scanned = true;
+
+  // Update affected values.
+  for (auto &A : AssumeHandles)
+    updateAffectedValues(cast<CallInst>(A));
 }
 
 void AssumptionCache::registerAssumption(CallInst *CI) {
@@ -72,6 +179,8 @@ void AssumptionCache::registerAssumption(CallInst *CI) {
            "Cache contains multiple copies of a call!");
   }
 #endif
+
+  updateAffectedValues(CI);
 }
 
 AnalysisKey AssumptionAnalysis::Key;

contrib/llvm/lib/Analysis/CostModel.cpp

@@ -438,8 +438,11 @@ unsigned CostModelAnalysis::getInstructionCost(const Instruction *I) const {
       getOperandInfo(I->getOperand(0));
     TargetTransformInfo::OperandValueKind Op2VK =
       getOperandInfo(I->getOperand(1));
+    SmallVector<const Value*, 2> Operands(I->operand_values()); 
     return TTI->getArithmeticInstrCost(I->getOpcode(), I->getType(), Op1VK,
-                                       Op2VK);
+                                       Op2VK, TargetTransformInfo::OP_None, 
+                                       TargetTransformInfo::OP_None, 
+                                       Operands);
   }
   case Instruction::Select: {
     const SelectInst *SI = cast<SelectInst>(I);

contrib/llvm/lib/Analysis/IVUsers.cpp

@@ -16,8 +16,8 @@
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/Analysis/AssumptionCache.h"
 #include "llvm/Analysis/CodeMetrics.h"
+#include "llvm/Analysis/LoopAnalysisManager.h"
 #include "llvm/Analysis/LoopPass.h"
-#include "llvm/Analysis/LoopPassManager.h"
 #include "llvm/Analysis/ScalarEvolutionExpressions.h"
 #include "llvm/Analysis/ValueTracking.h"
 #include "llvm/IR/Constants.h"
@@ -36,20 +36,9 @@ using namespace llvm;
 
 AnalysisKey IVUsersAnalysis::Key;
 
-IVUsers IVUsersAnalysis::run(Loop &L, LoopAnalysisManager &AM) {
-  const auto &FAM =
-      AM.getResult<FunctionAnalysisManagerLoopProxy>(L).getManager();
-  Function *F = L.getHeader()->getParent();
-
-  return IVUsers(&L, FAM.getCachedResult<AssumptionAnalysis>(*F),
-                 FAM.getCachedResult<LoopAnalysis>(*F),
-                 FAM.getCachedResult<DominatorTreeAnalysis>(*F),
-                 FAM.getCachedResult<ScalarEvolutionAnalysis>(*F));
-}
-
-PreservedAnalyses IVUsersPrinterPass::run(Loop &L, LoopAnalysisManager &AM) {
-  AM.getResult<IVUsersAnalysis>(L).print(OS);
-  return PreservedAnalyses::all();
+IVUsers IVUsersAnalysis::run(Loop &L, LoopAnalysisManager &AM,
+                             LoopStandardAnalysisResults &AR) {
+  return IVUsers(&L, &AR.AC, &AR.LI, &AR.DT, &AR.SE);
 }
 
 char IVUsersWrapperPass::ID = 0;

contrib/llvm/lib/Analysis/InlineCost.cpp

@@ -636,30 +636,27 @@ void CallAnalyzer::updateThreshold(CallSite CS, Function &Callee) {
   else if (Caller->optForSize())
     Threshold = MinIfValid(Threshold, Params.OptSizeThreshold);
 
-  bool HotCallsite = false;
-  uint64_t TotalWeight;
-  if (PSI && CS.getInstruction()->extractProfTotalWeight(TotalWeight) &&
-      PSI->isHotCount(TotalWeight)) {
-    HotCallsite = true;
+  // Adjust the threshold based on inlinehint attribute and profile based
+  // hotness information if the caller does not have MinSize attribute.
+  if (!Caller->optForMinSize()) {
+    if (Callee.hasFnAttribute(Attribute::InlineHint))
+      Threshold = MaxIfValid(Threshold, Params.HintThreshold);
+    if (PSI) {
+      uint64_t TotalWeight;
+      if (CS.getInstruction()->extractProfTotalWeight(TotalWeight) &&
+          PSI->isHotCount(TotalWeight)) {
+        Threshold = MaxIfValid(Threshold, Params.HotCallSiteThreshold);
+      } else if (PSI->isFunctionEntryHot(&Callee)) {
+        // If callsite hotness can not be determined, we may still know
+        // that the callee is hot and treat it as a weaker hint for threshold
+        // increase.
+        Threshold = MaxIfValid(Threshold, Params.HintThreshold);
+      } else if (PSI->isFunctionEntryCold(&Callee)) {
+        Threshold = MinIfValid(Threshold, Params.ColdThreshold);
+      }
+    }
   }
 
-  // Listen to the inlinehint attribute or profile based hotness information
-  // when it would increase the threshold and the caller does not need to
-  // minimize its size.
-  bool InlineHint = Callee.hasFnAttribute(Attribute::InlineHint) ||
-                    (PSI && PSI->isFunctionEntryHot(&Callee));
-  if (InlineHint && !Caller->optForMinSize())
-    Threshold = MaxIfValid(Threshold, Params.HintThreshold);
-
-  if (HotCallsite && !Caller->optForMinSize())
-    Threshold = MaxIfValid(Threshold, Params.HotCallSiteThreshold);
-
-  bool ColdCallee = PSI && PSI->isFunctionEntryCold(&Callee);
-  // For cold callees, use the ColdThreshold knob if it is available and reduces
-  // the threshold.
-  if (ColdCallee)
-    Threshold = MinIfValid(Threshold, Params.ColdThreshold);
-
   // Finally, take the target-specific inlining threshold multiplier into
   // account.
   Threshold *= TTI.getInliningThresholdMultiplier();

contrib/llvm/lib/Analysis/InstructionSimplify.cpp

@@ -3583,7 +3583,7 @@ static Value *simplifySelectBitTest(Value *TrueVal, Value *FalseVal, Value *X,
         *Y == *C)
       return TrueWhenUnset ? TrueVal : FalseVal;
   }
-  
+
   return nullptr;
 }
 
@@ -3595,7 +3595,7 @@ static Value *simplifySelectWithFakeICmpEq(Value *CmpLHS, Value *TrueVal,
   unsigned BitWidth = TrueVal->getType()->getScalarSizeInBits();
   if (!BitWidth)
     return nullptr;
-  
+
   APInt MinSignedValue;
   Value *X;
   if (match(CmpLHS, m_Trunc(m_Value(X))) && (X == TrueVal || X == FalseVal)) {
@@ -4252,14 +4252,36 @@ static Value *SimplifyIntrinsic(Function *F, IterTy ArgBegin, IterTy ArgEnd,
                                 const Query &Q, unsigned MaxRecurse) {
   Intrinsic::ID IID = F->getIntrinsicID();
   unsigned NumOperands = std::distance(ArgBegin, ArgEnd);
-  Type *ReturnType = F->getReturnType();
+
+  // Unary Ops
+  if (NumOperands == 1) {
+    // Perform idempotent optimizations
+    if (IsIdempotent(IID)) {
+      if (IntrinsicInst *II = dyn_cast<IntrinsicInst>(*ArgBegin)) {
+        if (II->getIntrinsicID() == IID)
+          return II;
+      }
+    }
+
+    switch (IID) {
+    case Intrinsic::fabs: {
+      if (SignBitMustBeZero(*ArgBegin, Q.TLI))
+        return *ArgBegin;
+    }
+    default:
+      return nullptr;
+    }
+  }
 
   // Binary Ops
   if (NumOperands == 2) {
     Value *LHS = *ArgBegin;
     Value *RHS = *(ArgBegin + 1);
-    if (IID == Intrinsic::usub_with_overflow ||
-        IID == Intrinsic::ssub_with_overflow) {
+    Type *ReturnType = F->getReturnType();
+
+    switch (IID) {
+    case Intrinsic::usub_with_overflow:
+    case Intrinsic::ssub_with_overflow: {
       // X - X -> { 0, false }
       if (LHS == RHS)
         return Constant::getNullValue(ReturnType);
@@ -4268,17 +4290,19 @@ static Value *SimplifyIntrinsic(Function *F, IterTy ArgBegin, IterTy ArgEnd,
       // undef - X -> undef
       if (isa<UndefValue>(LHS) || isa<UndefValue>(RHS))
         return UndefValue::get(ReturnType);
-    }
 
-    if (IID == Intrinsic::uadd_with_overflow ||
-        IID == Intrinsic::sadd_with_overflow) {
+      return nullptr;
+    }
+    case Intrinsic::uadd_with_overflow:
+    case Intrinsic::sadd_with_overflow: {
       // X + undef -> undef
       if (isa<UndefValue>(RHS))
         return UndefValue::get(ReturnType);
-    }
 
-    if (IID == Intrinsic::umul_with_overflow ||
-        IID == Intrinsic::smul_with_overflow) {
+      return nullptr;
+    }
+    case Intrinsic::umul_with_overflow:
+    case Intrinsic::smul_with_overflow: {
       // X * 0 -> { 0, false }
       if (match(RHS, m_Zero()))
         return Constant::getNullValue(ReturnType);
@@ -4286,34 +4310,34 @@ static Value *SimplifyIntrinsic(Function *F, IterTy ArgBegin, IterTy ArgEnd,
       // X * undef -> { 0, false }
       if (match(RHS, m_Undef()))
         return Constant::getNullValue(ReturnType);
-    }
 
-    if (IID == Intrinsic::load_relative && isa<Constant>(LHS) &&
-        isa<Constant>(RHS))
-      return SimplifyRelativeLoad(cast<Constant>(LHS), cast<Constant>(RHS),
-                                  Q.DL);
+      return nullptr;
+    }
+    case Intrinsic::load_relative: {
+      Constant *C0 = dyn_cast<Constant>(LHS);
+      Constant *C1 = dyn_cast<Constant>(RHS);
+      if (C0 && C1)
+        return SimplifyRelativeLoad(C0, C1, Q.DL);
+      return nullptr;
+    }
+    default:
+      return nullptr;
+    }
   }
 
   // Simplify calls to llvm.masked.load.*
-  if (IID == Intrinsic::masked_load) {
+  switch (IID) {
+  case Intrinsic::masked_load: {
     Value *MaskArg = ArgBegin[2];
     Value *PassthruArg = ArgBegin[3];
     // If the mask is all zeros or undef, the "passthru" argument is the result.
     if (maskIsAllZeroOrUndef(MaskArg))
       return PassthruArg;
-  }
-
-  // Perform idempotent optimizations
-  if (!IsIdempotent(IID))
     return nullptr;
-
-  // Unary Ops
-  if (NumOperands == 1)
-    if (IntrinsicInst *II = dyn_cast<IntrinsicInst>(*ArgBegin))
-      if (II->getIntrinsicID() == IID)
-        return II;
-
-  return nullptr;
+  }
+  default:
+    return nullptr;
+  }
 }
 
 template <typename IterTy>

contrib/llvm/lib/Analysis/LazyValueInfo.cpp

@@ -925,7 +925,7 @@ void LazyValueInfoImpl::intersectAssumeOrGuardBlockValueConstantRange(
   if (!BBI)
     return;
 
-  for (auto &AssumeVH : AC->assumptions()) {
+  for (auto &AssumeVH : AC->assumptionsFor(Val)) {
     if (!AssumeVH)
       continue;
     auto *I = cast<CallInst>(AssumeVH);

contrib/llvm/lib/Analysis/LoopAccessAnalysis.cpp

@@ -12,22 +12,22 @@
 //
 //===----------------------------------------------------------------------===//
 
+#include "llvm/Analysis/LoopAccessAnalysis.h"
 #include "llvm/ADT/APInt.h"
 #include "llvm/ADT/DenseMap.h"
 #include "llvm/ADT/DepthFirstIterator.h"
 #include "llvm/ADT/EquivalenceClasses.h"
-#include "llvm/ADT/iterator_range.h"
 #include "llvm/ADT/PointerIntPair.h"
+#include "llvm/ADT/STLExtras.h"
 #include "llvm/ADT/SetVector.h"
 #include "llvm/ADT/SmallPtrSet.h"
 #include "llvm/ADT/SmallSet.h"
 #include "llvm/ADT/SmallVector.h"
-#include "llvm/ADT/STLExtras.h"
+#include "llvm/ADT/iterator_range.h"
 #include "llvm/Analysis/AliasAnalysis.h"
 #include "llvm/Analysis/AliasSetTracker.h"
-#include "llvm/Analysis/LoopAccessAnalysis.h"
+#include "llvm/Analysis/LoopAnalysisManager.h"
 #include "llvm/Analysis/LoopInfo.h"
-#include "llvm/Analysis/LoopPassManager.h"
 #include "llvm/Analysis/MemoryLocation.h"
 #include "llvm/Analysis/OptimizationDiagnosticInfo.h"
 #include "llvm/Analysis/ScalarEvolution.h"
@@ -44,10 +44,10 @@
 #include "llvm/IR/DiagnosticInfo.h"
 #include "llvm/IR/Dominators.h"
 #include "llvm/IR/Function.h"
+#include "llvm/IR/IRBuilder.h"
 #include "llvm/IR/InstrTypes.h"
 #include "llvm/IR/Instruction.h"
 #include "llvm/IR/Instructions.h"
-#include "llvm/IR/IRBuilder.h"
 #include "llvm/IR/Operator.h"
 #include "llvm/IR/PassManager.h"
 #include "llvm/IR/Type.h"
@@ -2120,35 +2120,9 @@ INITIALIZE_PASS_END(LoopAccessLegacyAnalysis, LAA_NAME, laa_name, false, true)
 
 AnalysisKey LoopAccessAnalysis::Key;
 
-LoopAccessInfo LoopAccessAnalysis::run(Loop &L, LoopAnalysisManager &AM) {
-  const FunctionAnalysisManager &FAM =
-      AM.getResult<FunctionAnalysisManagerLoopProxy>(L).getManager();
-  Function &F = *L.getHeader()->getParent();
-  auto *SE = FAM.getCachedResult<ScalarEvolutionAnalysis>(F);
-  auto *TLI = FAM.getCachedResult<TargetLibraryAnalysis>(F);
-  auto *AA = FAM.getCachedResult<AAManager>(F);
-  auto *DT = FAM.getCachedResult<DominatorTreeAnalysis>(F);
-  auto *LI = FAM.getCachedResult<LoopAnalysis>(F);
-  if (!SE)
-    report_fatal_error(
-        "ScalarEvolution must have been cached at a higher level");
-  if (!AA)
-    report_fatal_error("AliasAnalysis must have been cached at a higher level");
-  if (!DT)
-    report_fatal_error("DominatorTree must have been cached at a higher level");
-  if (!LI)
-    report_fatal_error("LoopInfo must have been cached at a higher level");
-  return LoopAccessInfo(&L, SE, TLI, AA, DT, LI);
-}
-
-PreservedAnalyses LoopAccessInfoPrinterPass::run(Loop &L,
-                                                 LoopAnalysisManager &AM) {
-  Function &F = *L.getHeader()->getParent();
-  auto &LAI = AM.getResult<LoopAccessAnalysis>(L);
-  OS << "Loop access info in function '" << F.getName() << "':\n";
-  OS.indent(2) << L.getHeader()->getName() << ":\n";
-  LAI.print(OS, 4);
-  return PreservedAnalyses::all();
+LoopAccessInfo LoopAccessAnalysis::run(Loop &L, LoopAnalysisManager &AM,
+                                       LoopStandardAnalysisResults &AR) {
+  return LoopAccessInfo(&L, &AR.SE, &AR.TLI, &AR.AA, &AR.DT, &AR.LI);
 }
 
 namespace llvm {

contrib/llvm/lib/Analysis/LoopAnalysisManager.cpp

@@ -0,0 +1,160 @@
+//===- LoopAnalysisManager.cpp - Loop analysis management -----------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+#include "llvm/Analysis/LoopAnalysisManager.h"
+#include "llvm/Analysis/BasicAliasAnalysis.h"
+#include "llvm/Analysis/GlobalsModRef.h"
+#include "llvm/Analysis/LoopInfo.h"
+#include "llvm/Analysis/ScalarEvolution.h"
+#include "llvm/Analysis/ScalarEvolutionAliasAnalysis.h"
+#include "llvm/IR/Dominators.h"
+
+using namespace llvm;
+
+// Explicit template instantiations and specialization defininitions for core
+// template typedefs.
+namespace llvm {
+template class AllAnalysesOn<Loop>;
+template class AnalysisManager<Loop, LoopStandardAnalysisResults &>;
+template class InnerAnalysisManagerProxy<LoopAnalysisManager, Function>;
+template class OuterAnalysisManagerProxy<FunctionAnalysisManager, Loop,
+                                         LoopStandardAnalysisResults &>;
+
+bool LoopAnalysisManagerFunctionProxy::Result::invalidate(
+    Function &F, const PreservedAnalyses &PA,
+    FunctionAnalysisManager::Invalidator &Inv) {
+  // First compute the sequence of IR units covered by this proxy. We will want
+  // to visit this in postorder, but because this is a tree structure we can do
+  // this by building a preorder sequence and walking it in reverse.
+  SmallVector<Loop *, 4> PreOrderLoops, PreOrderWorklist;
+  // Note that we want to walk the roots in reverse order because we will end
+  // up reversing the preorder sequence. However, it happens that the loop nest
+  // roots are in reverse order within the LoopInfo object. So we just walk
+  // forward here.
+  // FIXME: If we change the order of LoopInfo we will want to add a reverse
+  // here.
+  for (Loop *RootL : *LI) {
+    assert(PreOrderWorklist.empty() &&
+           "Must start with an empty preorder walk worklist.");
+    PreOrderWorklist.push_back(RootL);
+    do {
+      Loop *L = PreOrderWorklist.pop_back_val();
+      PreOrderWorklist.append(L->begin(), L->end());
+      PreOrderLoops.push_back(L);
+    } while (!PreOrderWorklist.empty());
+  }
+
+  // If this proxy or the loop info is going to be invalidated, we also need
+  // to clear all the keys coming from that analysis. We also completely blow
+  // away the loop analyses if any of the standard analyses provided by the
+  // loop pass manager go away so that loop analyses can freely use these
+  // without worrying about declaring dependencies on them etc.
+  // FIXME: It isn't clear if this is the right tradeoff. We could instead make
+  // loop analyses declare any dependencies on these and use the more general
+  // invalidation logic below to act on that.
+  auto PAC = PA.getChecker<LoopAnalysisManagerFunctionProxy>();
+  if (!(PAC.preserved() || PAC.preservedSet<AllAnalysesOn<Function>>()) ||
+      Inv.invalidate<AAManager>(F, PA) ||
+      Inv.invalidate<AssumptionAnalysis>(F, PA) ||
+      Inv.invalidate<DominatorTreeAnalysis>(F, PA) ||
+      Inv.invalidate<LoopAnalysis>(F, PA) ||
+      Inv.invalidate<ScalarEvolutionAnalysis>(F, PA)) {
+    // Note that the LoopInfo may be stale at this point, however the loop
+    // objects themselves remain the only viable keys that could be in the
+    // analysis manager's cache. So we just walk the keys and forcibly clear
+    // those results. Note that the order doesn't matter here as this will just
+    // directly destroy the results without calling methods on them.
+    for (Loop *L : PreOrderLoops)
+      InnerAM->clear(*L);
+
+    // We also need to null out the inner AM so that when the object gets
+    // destroyed as invalid we don't try to clear the inner AM again. At that
+    // point we won't be able to reliably walk the loops for this function and
+    // only clear results associated with those loops the way we do here.
+    // FIXME: Making InnerAM null at this point isn't very nice. Most analyses
+    // try to remain valid during invalidation. Maybe we should add an
+    // `IsClean` flag?
+    InnerAM = nullptr;
+
+    // Now return true to indicate this *is* invalid and a fresh proxy result
+    // needs to be built. This is especially important given the null InnerAM.
+    return true;
+  }
+
+  // Directly check if the relevant set is preserved so we can short circuit
+  // invalidating loops.
+  bool AreLoopAnalysesPreserved =
+      PA.allAnalysesInSetPreserved<AllAnalysesOn<Loop>>();
+
+  // Since we have a valid LoopInfo we can actually leave the cached results in
+  // the analysis manager associated with the Loop keys, but we need to
+  // propagate any necessary invalidation logic into them. We'd like to
+  // invalidate things in roughly the same order as they were put into the
+  // cache and so we walk the preorder list in reverse to form a valid
+  // postorder.
+  for (Loop *L : reverse(PreOrderLoops)) {
+    Optional<PreservedAnalyses> InnerPA;
+
+    // Check to see whether the preserved set needs to be adjusted based on
+    // function-level analysis invalidation triggering deferred invalidation
+    // for this loop.
+    if (auto *OuterProxy =
+            InnerAM->getCachedResult<FunctionAnalysisManagerLoopProxy>(*L))
+      for (const auto &OuterInvalidationPair :
+           OuterProxy->getOuterInvalidations()) {
+        AnalysisKey *OuterAnalysisID = OuterInvalidationPair.first;
+        const auto &InnerAnalysisIDs = OuterInvalidationPair.second;
+        if (Inv.invalidate(OuterAnalysisID, F, PA)) {
+          if (!InnerPA)
+            InnerPA = PA;
+          for (AnalysisKey *InnerAnalysisID : InnerAnalysisIDs)
+            InnerPA->abandon(InnerAnalysisID);
+        }
+      }
+
+    // Check if we needed a custom PA set. If so we'll need to run the inner
+    // invalidation.
+    if (InnerPA) {
+      InnerAM->invalidate(*L, *InnerPA);
+      continue;
+    }
+
+    // Otherwise we only need to do invalidation if the original PA set didn't
+    // preserve all Loop analyses.
+    if (!AreLoopAnalysesPreserved)
+      InnerAM->invalidate(*L, PA);
+  }
+
+  // Return false to indicate that this result is still a valid proxy.
+  return false;
+}
+
+template <>
+LoopAnalysisManagerFunctionProxy::Result
+LoopAnalysisManagerFunctionProxy::run(Function &F,
+                                      FunctionAnalysisManager &AM) {
+  return Result(*InnerAM, AM.getResult<LoopAnalysis>(F));
+}
+}
+
+PreservedAnalyses llvm::getLoopPassPreservedAnalyses() {
+  PreservedAnalyses PA;
+  PA.preserve<AssumptionAnalysis>();
+  PA.preserve<DominatorTreeAnalysis>();
+  PA.preserve<LoopAnalysis>();
+  PA.preserve<LoopAnalysisManagerFunctionProxy>();
+  PA.preserve<ScalarEvolutionAnalysis>();
+  // TODO: What we really want to do here is preserve an AA category, but that
+  // concept doesn't exist yet.
+  PA.preserve<AAManager>();
+  PA.preserve<BasicAA>();
+  PA.preserve<GlobalsAA>();
+  PA.preserve<SCEVAA>();
+  return PA;
+}

contrib/llvm/lib/Analysis/LoopInfo.cpp

@@ -689,18 +689,13 @@ PreservedAnalyses LoopPrinterPass::run(Function &F,
   return PreservedAnalyses::all();
 }
 
-PrintLoopPass::PrintLoopPass() : OS(dbgs()) {}
-PrintLoopPass::PrintLoopPass(raw_ostream &OS, const std::string &Banner)
-    : OS(OS), Banner(Banner) {}
-
-PreservedAnalyses PrintLoopPass::run(Loop &L, AnalysisManager<Loop> &) {
+void llvm::printLoop(Loop &L, raw_ostream &OS, const std::string &Banner) {
   OS << Banner;
   for (auto *Block : L.blocks())
     if (Block)
       Block->print(OS);
     else
       OS << "Printing <null> block";
-  return PreservedAnalyses::all();
 }
 
 //===----------------------------------------------------------------------===//

contrib/llvm/lib/Analysis/LoopPass.cpp

@@ -14,7 +14,7 @@
 //===----------------------------------------------------------------------===//
 
 #include "llvm/Analysis/LoopPass.h"
-#include "llvm/Analysis/LoopPassManager.h"
+#include "llvm/Analysis/LoopAnalysisManager.h"
 #include "llvm/IR/Dominators.h"
 #include "llvm/IR/IRPrintingPasses.h"
 #include "llvm/IR/LLVMContext.h"
@@ -32,13 +32,14 @@ namespace {
 /// PrintLoopPass - Print a Function corresponding to a Loop.
 ///
 class PrintLoopPassWrapper : public LoopPass {
-  PrintLoopPass P;
+  raw_ostream &OS;
+  std::string Banner;
 
 public:
   static char ID;
-  PrintLoopPassWrapper() : LoopPass(ID) {}
+  PrintLoopPassWrapper() : LoopPass(ID), OS(dbgs()) {}
   PrintLoopPassWrapper(raw_ostream &OS, const std::string &Banner)
-      : LoopPass(ID), P(OS, Banner) {}
+      : LoopPass(ID), OS(OS), Banner(Banner) {}
 
   void getAnalysisUsage(AnalysisUsage &AU) const override {
     AU.setPreservesAll();
@@ -49,8 +50,7 @@ class PrintLoopPassWrapper : public LoopPass {
                        [](BasicBlock *BB) { return BB; });
     if (BBI != L->blocks().end() &&
         isFunctionInPrintList((*BBI)->getParent()->getName())) {
-      LoopAnalysisManager DummyLAM;
-      P.run(*L, DummyLAM);
+      printLoop(*L, OS, Banner);
     }
     return false;
   }

contrib/llvm/lib/Analysis/LoopPassManager.cpp

@@ -1,59 +0,0 @@
-//===- LoopPassManager.cpp - Loop pass management -------------------------===//
-//
-//                     The LLVM Compiler Infrastructure
-//
-// This file is distributed under the University of Illinois Open Source
-// License. See LICENSE.TXT for details.
-//
-//===----------------------------------------------------------------------===//
-
-#include "llvm/Analysis/LoopPassManager.h"
-#include "llvm/Analysis/BasicAliasAnalysis.h"
-#include "llvm/Analysis/GlobalsModRef.h"
-#include "llvm/Analysis/LoopInfo.h"
-#include "llvm/Analysis/ScalarEvolution.h"
-#include "llvm/Analysis/ScalarEvolutionAliasAnalysis.h"
-#include "llvm/IR/Dominators.h"
-
-using namespace llvm;
-
-// Explicit template instantiations and specialization defininitions for core
-// template typedefs.
-namespace llvm {
-template class PassManager<Loop>;
-template class AnalysisManager<Loop>;
-template class InnerAnalysisManagerProxy<LoopAnalysisManager, Function>;
-template class OuterAnalysisManagerProxy<FunctionAnalysisManager, Loop>;
-
-template <>
-bool LoopAnalysisManagerFunctionProxy::Result::invalidate(
-    Function &F, const PreservedAnalyses &PA,
-    FunctionAnalysisManager::Invalidator &Inv) {
-  // If this proxy isn't marked as preserved, the set of Function objects in
-  // the module may have changed. We therefore can't call
-  // InnerAM->invalidate(), because any pointers to Functions it has may be
-  // stale.
-  auto PAC = PA.getChecker<LoopAnalysisManagerFunctionProxy>();
-  if (!PAC.preserved() && !PAC.preservedSet<AllAnalysesOn<Loop>>())
-    InnerAM->clear();
-
-  // FIXME: Proper suppor for invalidation isn't yet implemented for the LPM.
-
-  // Return false to indicate that this result is still a valid proxy.
-  return false;
-}
-}
-
-PreservedAnalyses llvm::getLoopPassPreservedAnalyses() {
-  PreservedAnalyses PA;
-  PA.preserve<DominatorTreeAnalysis>();
-  PA.preserve<LoopAnalysis>();
-  PA.preserve<ScalarEvolutionAnalysis>();
-  // TODO: What we really want to do here is preserve an AA category, but that
-  // concept doesn't exist yet.
-  PA.preserve<AAManager>();
-  PA.preserve<BasicAA>();
-  PA.preserve<GlobalsAA>();
-  PA.preserve<SCEVAA>();
-  return PA;
-}

contrib/llvm/lib/Analysis/MemoryDependenceAnalysis.cpp

@@ -323,17 +323,28 @@ MemDepResult MemoryDependenceResults::getPointerDependencyFrom(
     const MemoryLocation &MemLoc, bool isLoad, BasicBlock::iterator ScanIt,
     BasicBlock *BB, Instruction *QueryInst, unsigned *Limit) {
 
+  MemDepResult InvariantGroupDependency = MemDepResult::getUnknown();
   if (QueryInst != nullptr) {
     if (auto *LI = dyn_cast<LoadInst>(QueryInst)) {
-      MemDepResult invariantGroupDependency =
-          getInvariantGroupPointerDependency(LI, BB);
+      InvariantGroupDependency = getInvariantGroupPointerDependency(LI, BB);
 
-      if (invariantGroupDependency.isDef())
-        return invariantGroupDependency;
+      if (InvariantGroupDependency.isDef())
+        return InvariantGroupDependency;
     }
   }
-  return getSimplePointerDependencyFrom(MemLoc, isLoad, ScanIt, BB, QueryInst,
-                                        Limit);
+  MemDepResult SimpleDep = getSimplePointerDependencyFrom(
+      MemLoc, isLoad, ScanIt, BB, QueryInst, Limit);
+  if (SimpleDep.isDef())
+    return SimpleDep;
+  // Non-local invariant group dependency indicates there is non local Def
+  // (it only returns nonLocal if it finds nonLocal def), which is better than
+  // local clobber and everything else.
+  if (InvariantGroupDependency.isNonLocal())
+    return InvariantGroupDependency;
+
+  assert(InvariantGroupDependency.isUnknown() &&
+         "InvariantGroupDependency should be only unknown at this point");
+  return SimpleDep;
 }
 
 MemDepResult
@@ -358,6 +369,20 @@ MemoryDependenceResults::getInvariantGroupPointerDependency(LoadInst *LI,
   // Queue to process all pointers that are equivalent to load operand.
   SmallVector<const Value *, 8> LoadOperandsQueue;
   LoadOperandsQueue.push_back(LoadOperand);
+
+  Instruction *ClosestDependency = nullptr;
+  // Order of instructions in uses list is unpredictible. In order to always
+  // get the same result, we will look for the closest dominance.
+  auto GetClosestDependency = [this](Instruction *Best, Instruction *Other) {
+    assert(Other && "Must call it with not null instruction");
+    if (Best == nullptr || DT.dominates(Best, Other))
+      return Other;
+    return Best;
+  };
+
+
+  // FIXME: This loop is O(N^2) because dominates can be O(n) and in worst case
+  // we will see all the instructions. This should be fixed in MSSA.
   while (!LoadOperandsQueue.empty()) {
     const Value *Ptr = LoadOperandsQueue.pop_back_val();
     assert(Ptr && !isa<GlobalValue>(Ptr) &&
@@ -388,12 +413,24 @@ MemoryDependenceResults::getInvariantGroupPointerDependency(LoadInst *LI,
       // If we hit load/store with the same invariant.group metadata (and the
       // same pointer operand) we can assume that value pointed by pointer
       // operand didn't change.
-      if ((isa<LoadInst>(U) || isa<StoreInst>(U)) && U->getParent() == BB &&
+      if ((isa<LoadInst>(U) || isa<StoreInst>(U)) &&
           U->getMetadata(LLVMContext::MD_invariant_group) == InvariantGroupMD)
-        return MemDepResult::getDef(U);
+        ClosestDependency = GetClosestDependency(ClosestDependency, U);
     }
   }
-  return MemDepResult::getUnknown();
+
+  if (!ClosestDependency)
+    return MemDepResult::getUnknown();
+  if (ClosestDependency->getParent() == BB)
+    return MemDepResult::getDef(ClosestDependency);
+  // Def(U) can't be returned here because it is non-local. If local
+  // dependency won't be found then return nonLocal counting that the
+  // user will call getNonLocalPointerDependency, which will return cached
+  // result.
+  NonLocalDefsCache.try_emplace(
+      LI, NonLocalDepResult(ClosestDependency->getParent(),
+                            MemDepResult::getDef(ClosestDependency), nullptr));
+  return MemDepResult::getNonLocal();
 }
 
 MemDepResult MemoryDependenceResults::getSimplePointerDependencyFrom(
@@ -877,7 +914,17 @@ void MemoryDependenceResults::getNonLocalPointerDependency(
   assert(Loc.Ptr->getType()->isPointerTy() &&
          "Can't get pointer deps of a non-pointer!");
   Result.clear();
-
+  {
+    // Check if there is cached Def with invariant.group. FIXME: cache might be
+    // invalid if cached instruction would be removed between call to
+    // getPointerDependencyFrom and this function.
+    auto NonLocalDefIt = NonLocalDefsCache.find(QueryInst);
+    if (NonLocalDefIt != NonLocalDefsCache.end()) {
+      Result.push_back(std::move(NonLocalDefIt->second));
+      NonLocalDefsCache.erase(NonLocalDefIt);
+      return;
+    }
+  }
   // This routine does not expect to deal with volatile instructions.
   // Doing so would require piping through the QueryInst all the way through.
   // TODO: volatiles can't be elided, but they can be reordered with other

contrib/llvm/lib/Analysis/ScalarEvolution.cpp

@@ -7032,20 +7032,21 @@ static const SCEV *SolveLinEquationWithOverflow(const APInt &A, const APInt &B,
   // 3. Compute I: the multiplicative inverse of (A / D) in arithmetic
   // modulo (N / D).
   //
-  // (N / D) may need BW+1 bits in its representation.  Hence, we'll use this
-  // bit width during computations.
+  // If D == 1, (N / D) == N == 2^BW, so we need one extra bit to represent
+  // (N / D) in general. The inverse itself always fits into BW bits, though,
+  // so we immediately truncate it.
   APInt AD = A.lshr(Mult2).zext(BW + 1);  // AD = A / D
   APInt Mod(BW + 1, 0);
   Mod.setBit(BW - Mult2);  // Mod = N / D
-  APInt I = AD.multiplicativeInverse(Mod);
+  APInt I = AD.multiplicativeInverse(Mod).trunc(BW);
 
   // 4. Compute the minimum unsigned root of the equation:
   // I * (B / D) mod (N / D)
-  APInt Result = (I * B.lshr(Mult2).zext(BW + 1)).urem(Mod);
+  // To simplify the computation, we factor out the divide by D:
+  // (I * B mod N) / D
+  APInt Result = (I * B).lshr(Mult2);
 
-  // The result is guaranteed to be less than 2^BW so we may truncate it to BW
-  // bits.
-  return SE.getConstant(Result.trunc(BW));
+  return SE.getConstant(Result);
 }
 
 /// Find the roots of the quadratic equation for the given quadratic chrec
@@ -7206,17 +7207,25 @@ ScalarEvolution::howFarToZero(const SCEV *V, const Loop *L, bool ControlsExit,
   // 1*N = -Start; -1*N = Start (mod 2^BW), so:
   //   N = Distance (as unsigned)
   if (StepC->getValue()->equalsInt(1) || StepC->getValue()->isAllOnesValue()) {
-    ConstantRange CR = getUnsignedRange(Start);
-    const SCEV *MaxBECount;
-    if (!CountDown && CR.getUnsignedMin().isMinValue())
-      // When counting up, the worst starting value is 1, not 0.
-      MaxBECount = CR.getUnsignedMax().isMinValue()
-        ? getConstant(APInt::getMinValue(CR.getBitWidth()))
-        : getConstant(APInt::getMaxValue(CR.getBitWidth()));
-    else
-      MaxBECount = getConstant(CountDown ? CR.getUnsignedMax()
-                                         : -CR.getUnsignedMin());
-    return ExitLimit(Distance, MaxBECount, false, Predicates);
+    APInt MaxBECount = getUnsignedRange(Distance).getUnsignedMax();
+
+    // When a loop like "for (int i = 0; i != n; ++i) { /* body */ }" is rotated,
+    // we end up with a loop whose backedge-taken count is n - 1.  Detect this
+    // case, and see if we can improve the bound.
+    //
+    // Explicitly handling this here is necessary because getUnsignedRange
+    // isn't context-sensitive; it doesn't know that we only care about the
+    // range inside the loop.
+    const SCEV *Zero = getZero(Distance->getType());
+    const SCEV *One = getOne(Distance->getType());
+    const SCEV *DistancePlusOne = getAddExpr(Distance, One);
+    if (isLoopEntryGuardedByCond(L, ICmpInst::ICMP_NE, DistancePlusOne, Zero)) {
+      // If Distance + 1 doesn't overflow, we can compute the maximum distance
+      // as "unsigned_max(Distance + 1) - 1".
+      ConstantRange CR = getUnsignedRange(DistancePlusOne);
+      MaxBECount = APIntOps::umin(MaxBECount, CR.getUnsignedMax() - 1);
+    }
+    return ExitLimit(Distance, getConstant(MaxBECount), false, Predicates);
   }
 
   // As a special case, handle the instance where Step is a positive power of

contrib/llvm/lib/Analysis/TargetTransformInfo.cpp

@@ -277,9 +277,10 @@ unsigned TargetTransformInfo::getMaxInterleaveFactor(unsigned VF) const {
 int TargetTransformInfo::getArithmeticInstrCost(
     unsigned Opcode, Type *Ty, OperandValueKind Opd1Info,
     OperandValueKind Opd2Info, OperandValueProperties Opd1PropInfo,
-    OperandValueProperties Opd2PropInfo) const {
+    OperandValueProperties Opd2PropInfo,
+    ArrayRef<const Value *> Args) const {
   int Cost = TTIImpl->getArithmeticInstrCost(Opcode, Ty, Opd1Info, Opd2Info,
-                                             Opd1PropInfo, Opd2PropInfo);
+                                             Opd1PropInfo, Opd2PropInfo, Args);
   assert(Cost >= 0 && "TTI should not produce negative costs!");
   return Cost;
 }

contrib/llvm/lib/Analysis/ValueTracking.cpp

@@ -526,7 +526,10 @@ static void computeKnownBitsFromAssume(const Value *V, APInt &KnownZero,
 
   unsigned BitWidth = KnownZero.getBitWidth();
 
-  for (auto &AssumeVH : Q.AC->assumptions()) {
+  // Note that the patterns below need to be kept in sync with the code
+  // in AssumptionCache::updateAffectedValues.
+
+  for (auto &AssumeVH : Q.AC->assumptionsFor(V)) {
     if (!AssumeVH)
       continue;
     CallInst *I = cast<CallInst>(AssumeVH);
@@ -2580,51 +2583,70 @@ bool llvm::CannotBeNegativeZero(const Value *V, const TargetLibraryInfo *TLI,
   return false;
 }
 
-bool llvm::CannotBeOrderedLessThanZero(const Value *V,
-                                       const TargetLibraryInfo *TLI,
-                                       unsigned Depth) {
-  if (const ConstantFP *CFP = dyn_cast<ConstantFP>(V))
-    return !CFP->getValueAPF().isNegative() || CFP->getValueAPF().isZero();
+/// If \p SignBitOnly is true, test for a known 0 sign bit rather than a
+/// standard ordered compare. e.g. make -0.0 olt 0.0 be true because of the sign
+/// bit despite comparing equal.
+static bool cannotBeOrderedLessThanZeroImpl(const Value *V,
+                                            const TargetLibraryInfo *TLI,
+                                            bool SignBitOnly,
+                                            unsigned Depth) {
+  if (const ConstantFP *CFP = dyn_cast<ConstantFP>(V)) {
+    return !CFP->getValueAPF().isNegative() ||
+           (!SignBitOnly && CFP->getValueAPF().isZero());
+  }
 
   if (Depth == MaxDepth)
-    return false;  // Limit search depth.
+    return false; // Limit search depth.
 
   const Operator *I = dyn_cast<Operator>(V);
-  if (!I) return false;
+  if (!I)
+    return false;
 
   switch (I->getOpcode()) {
-  default: break;
+  default:
+    break;
   // Unsigned integers are always nonnegative.
   case Instruction::UIToFP:
     return true;
   case Instruction::FMul:
     // x*x is always non-negative or a NaN.
-    if (I->getOperand(0) == I->getOperand(1))
+    if (I->getOperand(0) == I->getOperand(1) &&
+        (!SignBitOnly || cast<FPMathOperator>(I)->hasNoNaNs()))
       return true;
+
     LLVM_FALLTHROUGH;
   case Instruction::FAdd:
   case Instruction::FDiv:
   case Instruction::FRem:
-    return CannotBeOrderedLessThanZero(I->getOperand(0), TLI, Depth + 1) &&
-           CannotBeOrderedLessThanZero(I->getOperand(1), TLI, Depth + 1);
+    return cannotBeOrderedLessThanZeroImpl(I->getOperand(0), TLI, SignBitOnly,
+                                           Depth + 1) &&
+           cannotBeOrderedLessThanZeroImpl(I->getOperand(1), TLI, SignBitOnly,
+                                           Depth + 1);
   case Instruction::Select:
-    return CannotBeOrderedLessThanZero(I->getOperand(1), TLI, Depth + 1) &&
-           CannotBeOrderedLessThanZero(I->getOperand(2), TLI, Depth + 1);
+    return cannotBeOrderedLessThanZeroImpl(I->getOperand(1), TLI, SignBitOnly,
+                                           Depth + 1) &&
+           cannotBeOrderedLessThanZeroImpl(I->getOperand(2), TLI, SignBitOnly,
+                                           Depth + 1);
   case Instruction::FPExt:
   case Instruction::FPTrunc:
     // Widening/narrowing never change sign.
-    return CannotBeOrderedLessThanZero(I->getOperand(0), TLI, Depth + 1);
+    return cannotBeOrderedLessThanZeroImpl(I->getOperand(0), TLI, SignBitOnly,
+                                           Depth + 1);
   case Instruction::Call:
     Intrinsic::ID IID = getIntrinsicForCallSite(cast<CallInst>(I), TLI);
     switch (IID) {
     default:
       break;
     case Intrinsic::maxnum:
-      return CannotBeOrderedLessThanZero(I->getOperand(0), TLI, Depth + 1) ||
-             CannotBeOrderedLessThanZero(I->getOperand(1), TLI, Depth + 1);
+      return cannotBeOrderedLessThanZeroImpl(I->getOperand(0), TLI, SignBitOnly,
+                                             Depth + 1) ||
+             cannotBeOrderedLessThanZeroImpl(I->getOperand(1), TLI, SignBitOnly,
+                                             Depth + 1);
     case Intrinsic::minnum:
-      return CannotBeOrderedLessThanZero(I->getOperand(0), TLI, Depth + 1) &&
-             CannotBeOrderedLessThanZero(I->getOperand(1), TLI, Depth + 1);
+      return cannotBeOrderedLessThanZeroImpl(I->getOperand(0), TLI, SignBitOnly,
+                                             Depth + 1) &&
+             cannotBeOrderedLessThanZeroImpl(I->getOperand(1), TLI, SignBitOnly,
+                                             Depth + 1);
     case Intrinsic::exp:
     case Intrinsic::exp2:
     case Intrinsic::fabs:
@@ -2636,18 +2658,30 @@ bool llvm::CannotBeOrderedLessThanZero(const Value *V,
         if (CI->getBitWidth() <= 64 && CI->getSExtValue() % 2u == 0)
           return true;
       }
-      return CannotBeOrderedLessThanZero(I->getOperand(0), TLI, Depth + 1);
+      return cannotBeOrderedLessThanZeroImpl(I->getOperand(0), TLI, SignBitOnly,
+                                             Depth + 1);
     case Intrinsic::fma:
     case Intrinsic::fmuladd:
       // x*x+y is non-negative if y is non-negative.
       return I->getOperand(0) == I->getOperand(1) &&
-             CannotBeOrderedLessThanZero(I->getOperand(2), TLI, Depth + 1);
+             (!SignBitOnly || cast<FPMathOperator>(I)->hasNoNaNs()) &&
+             cannotBeOrderedLessThanZeroImpl(I->getOperand(2), TLI, SignBitOnly,
+                                             Depth + 1);
     }
     break;
   }
   return false;
 }
 
+bool llvm::CannotBeOrderedLessThanZero(const Value *V,
+                                       const TargetLibraryInfo *TLI) {
+  return cannotBeOrderedLessThanZeroImpl(V, TLI, false, 0);
+}
+
+bool llvm::SignBitMustBeZero(const Value *V, const TargetLibraryInfo *TLI) {
+  return cannotBeOrderedLessThanZeroImpl(V, TLI, true, 0);
+}
+
 /// If the specified value can be set by repeating the same byte in memory,
 /// return the i8 value that it is represented with.  This is
 /// true for all i8 values obviously, but is also true for i32 0, i32 -1,

contrib/llvm/lib/CodeGen/AsmPrinter/CodeViewDebug.cpp

@@ -13,11 +13,13 @@
 
 #include "CodeViewDebug.h"
 #include "llvm/ADT/TinyPtrVector.h"
+#include "llvm/DebugInfo/CodeView/CVTypeDumper.h"
 #include "llvm/DebugInfo/CodeView/CVTypeVisitor.h"
 #include "llvm/DebugInfo/CodeView/CodeView.h"
 #include "llvm/DebugInfo/CodeView/Line.h"
 #include "llvm/DebugInfo/CodeView/SymbolRecord.h"
-#include "llvm/DebugInfo/CodeView/TypeDumper.h"
+#include "llvm/DebugInfo/CodeView/TypeDatabase.h"
+#include "llvm/DebugInfo/CodeView/TypeDumpVisitor.h"
 #include "llvm/DebugInfo/CodeView/TypeIndex.h"
 #include "llvm/DebugInfo/CodeView/TypeRecord.h"
 #include "llvm/DebugInfo/CodeView/TypeVisitorCallbacks.h"
@@ -467,7 +469,8 @@ void CodeViewDebug::emitTypeInformation() {
     CommentPrefix += ' ';
   }
 
-  CVTypeDumper CVTD(nullptr, /*PrintRecordBytes=*/false);
+  TypeDatabase TypeDB;
+  CVTypeDumper CVTD(TypeDB);
   TypeTable.ForEachRecord([&](TypeIndex Index, ArrayRef<uint8_t> Record) {
     if (OS.isVerboseAsm()) {
       // Emit a block comment describing the type record for readability.
@@ -475,8 +478,8 @@ void CodeViewDebug::emitTypeInformation() {
       raw_svector_ostream CommentOS(CommentBlock);
       ScopedPrinter SP(CommentOS);
       SP.setPrefix(CommentPrefix);
-      CVTD.setPrinter(&SP);
-      Error E = CVTD.dump(Record);
+      TypeDumpVisitor TDV(TypeDB, &SP, false);
+      Error E = CVTD.dump(Record, TDV);
       if (E) {
         logAllUnhandledErrors(std::move(E), errs(), "error: ");
         llvm_unreachable("produced malformed type record");

contrib/llvm/lib/CodeGen/AsmPrinter/DIE.cpp

@@ -79,6 +79,13 @@ void DIEAbbrev::Emit(const AsmPrinter *AP) const {
     // Emit form type.
     AP->EmitULEB128(AttrData.getForm(),
                     dwarf::FormEncodingString(AttrData.getForm()).data());
+
+    // Emit value for DW_FORM_implicit_const.
+    if (AttrData.getForm() == dwarf::DW_FORM_implicit_const) {
+      assert(AP->getDwarfVersion() >= 5 &&
+            "DW_FORM_implicit_const is supported starting from DWARFv5");
+      AP->EmitSLEB128(AttrData.getValue());
+    }
   }
 
   // Mark end of abbreviation.
@@ -160,7 +167,11 @@ DIE *DIE::getParent() const {
 DIEAbbrev DIE::generateAbbrev() const {
   DIEAbbrev Abbrev(Tag, hasChildren());
   for (const DIEValue &V : values())
-    Abbrev.AddAttribute(V.getAttribute(), V.getForm());
+    if (V.getForm() == dwarf::DW_FORM_implicit_const)
+      Abbrev.AddImplicitConstAttribute(V.getAttribute(),
+                                       V.getDIEInteger().getValue());
+    else
+      Abbrev.AddAttribute(V.getAttribute(), V.getForm());
   return Abbrev;
 }
 
@@ -342,6 +353,8 @@ void DIEValue::dump() const {
 ///
 void DIEInteger::EmitValue(const AsmPrinter *Asm, dwarf::Form Form) const {
   switch (Form) {
+  case dwarf::DW_FORM_implicit_const:
+    LLVM_FALLTHROUGH;
   case dwarf::DW_FORM_flag_present:
     // Emit something to keep the lines and comments in sync.
     // FIXME: Is there a better way to do this?
@@ -406,6 +419,7 @@ void DIEInteger::EmitValue(const AsmPrinter *Asm, dwarf::Form Form) const {
 ///
 unsigned DIEInteger::SizeOf(const AsmPrinter *AP, dwarf::Form Form) const {
   switch (Form) {
+  case dwarf::DW_FORM_implicit_const: LLVM_FALLTHROUGH;
   case dwarf::DW_FORM_flag_present: return 0;
   case dwarf::DW_FORM_flag:  LLVM_FALLTHROUGH;
   case dwarf::DW_FORM_ref1:  LLVM_FALLTHROUGH;

contrib/llvm/lib/CodeGen/AsmPrinter/DwarfUnit.cpp

@@ -200,6 +200,8 @@ void DwarfUnit::addUInt(DIEValueList &Die, dwarf::Attribute Attribute,
                         Optional<dwarf::Form> Form, uint64_t Integer) {
   if (!Form)
     Form = DIEInteger::BestForm(false, Integer);
+  assert(Form != dwarf::DW_FORM_implicit_const &&
+         "DW_FORM_implicit_const is used only for signed integers");
   Die.addValue(DIEValueAllocator, Attribute, *Form, DIEInteger(Integer));
 }
 

contrib/llvm/lib/CodeGen/GlobalISel/RegBankSelect.cpp

@@ -223,6 +223,7 @@ RegisterBankInfo::InstructionMapping &RegBankSelect::findBestMapping(
   for (RegisterBankInfo::InstructionMapping &CurMapping : PossibleMappings) {
     MappingCost CurCost = computeMapping(MI, CurMapping, LocalRepairPts, &Cost);
     if (CurCost < Cost) {
+      DEBUG(dbgs() << "New best: " << CurCost << '\n');
       Cost = CurCost;
       BestMapping = &CurMapping;
       RepairPts.clear();
@@ -377,8 +378,10 @@ RegBankSelect::MappingCost RegBankSelect::computeMapping(
   DEBUG(dbgs() << "Evaluating mapping cost for: " << MI);
   DEBUG(dbgs() << "With: " << InstrMapping << '\n');
   RepairPts.clear();
-  if (BestCost && Cost > *BestCost)
+  if (BestCost && Cost > *BestCost) {
+    DEBUG(dbgs() << "Mapping is too expensive from the start\n");
     return Cost;
+  }
 
   // Moreover, to realize this mapping, the register bank of each operand must
   // match this mapping. In other words, we may need to locally reassign the
@@ -392,17 +395,17 @@ RegBankSelect::MappingCost RegBankSelect::computeMapping(
     unsigned Reg = MO.getReg();
     if (!Reg)
       continue;
-    DEBUG(dbgs() << "Opd" << OpIdx);
+    DEBUG(dbgs() << "Opd" << OpIdx << '\n');
     const RegisterBankInfo::ValueMapping &ValMapping =
         InstrMapping.getOperandMapping(OpIdx);
     // If Reg is already properly mapped, this is free.
     bool Assign;
     if (assignmentMatch(Reg, ValMapping, Assign)) {
-      DEBUG(dbgs() << " is free (match).\n");
+      DEBUG(dbgs() << "=> is free (match).\n");
       continue;
     }
     if (Assign) {
-      DEBUG(dbgs() << " is free (simple assignment).\n");
+      DEBUG(dbgs() << "=> is free (simple assignment).\n");
       RepairPts.emplace_back(RepairingPlacement(MI, OpIdx, *TRI, *this,
                                                 RepairingPlacement::Reassign));
       continue;
@@ -420,8 +423,10 @@ RegBankSelect::MappingCost RegBankSelect::computeMapping(
       tryAvoidingSplit(RepairPt, MO, ValMapping);
 
     // Check that the materialization of the repairing is possible.
-    if (!RepairPt.canMaterialize())
+    if (!RepairPt.canMaterialize()) {
+      DEBUG(dbgs() << "Mapping involves impossible repairing\n");
       return MappingCost::ImpossibleCost();
+    }
 
     // Account for the split cost and repair cost.
     // Unless the cost is already saturated or we do not care about the cost.
@@ -476,8 +481,10 @@ RegBankSelect::MappingCost RegBankSelect::computeMapping(
 
       // Stop looking into what it takes to repair, this is already
       // too expensive.
-      if (BestCost && Cost > *BestCost)
+      if (BestCost && Cost > *BestCost) {
+        DEBUG(dbgs() << "Mapping is too expensive, stop processing\n");
         return Cost;
+      }
 
       // No need to accumulate more cost information.
       // We need to still gather the repairing information though.
@@ -485,6 +492,7 @@ RegBankSelect::MappingCost RegBankSelect::computeMapping(
         break;
     }
   }
+  DEBUG(dbgs() << "Total cost is: " << Cost << "\n");
   return Cost;
 }
 
@@ -550,7 +558,7 @@ bool RegBankSelect::assignInstr(MachineInstr &MI) {
   // Make sure the mapping is valid for MI.
   assert(BestMapping.verify(MI) && "Invalid instruction mapping");
 
-  DEBUG(dbgs() << "Mapping: " << BestMapping << '\n');
+  DEBUG(dbgs() << "Best Mapping: " << BestMapping << '\n');
 
   // After this call, MI may not be valid anymore.
   // Do not use it.
@@ -959,3 +967,20 @@ bool RegBankSelect::MappingCost::operator==(const MappingCost &Cost) const {
   return LocalCost == Cost.LocalCost && NonLocalCost == Cost.NonLocalCost &&
          LocalFreq == Cost.LocalFreq;
 }
+
+void RegBankSelect::MappingCost::dump() const {
+  print(dbgs());
+  dbgs() << '\n';
+}
+
+void RegBankSelect::MappingCost::print(raw_ostream &OS) const {
+  if (*this == ImpossibleCost()) {
+    OS << "impossible";
+    return;
+  }
+  if (isSaturated()) {
+    OS << "saturated";
+    return;
+  }
+  OS << LocalFreq << " * " << LocalCost << " + " << NonLocalCost;
+}

contrib/llvm/lib/CodeGen/GlobalISel/RegisterBank.cpp

@@ -19,12 +19,15 @@ using namespace llvm;
 
 const unsigned RegisterBank::InvalidID = UINT_MAX;
 
-RegisterBank::RegisterBank() : ID(InvalidID), Name(nullptr), Size(0) {}
+RegisterBank::RegisterBank(unsigned ID, const char *Name, unsigned Size,
+                           const uint32_t *CoveredClasses)
+    : ID(ID), Name(Name), Size(Size) {
+  ContainedRegClasses.resize(200);
+  ContainedRegClasses.setBitsInMask(CoveredClasses);
+}
 
 bool RegisterBank::verify(const TargetRegisterInfo &TRI) const {
   assert(isValid() && "Invalid register bank");
-  assert(ContainedRegClasses.size() == TRI.getNumRegClasses() &&
-         "TRI does not match the initialization process?");
   for (unsigned RCId = 0, End = TRI.getNumRegClasses(); RCId != End; ++RCId) {
     const TargetRegisterClass &RC = *TRI.getRegClass(RCId);
 

contrib/llvm/lib/CodeGen/GlobalISel/RegisterBankInfo.cpp

@@ -56,8 +56,10 @@ RegisterBankInfo::RegisterBankInfo(RegisterBank **RegBanks,
                                    unsigned NumRegBanks)
     : RegBanks(RegBanks), NumRegBanks(NumRegBanks) {
 #ifndef NDEBUG
-  for (unsigned Idx = 0, End = getNumRegBanks(); Idx != End; ++Idx)
+  for (unsigned Idx = 0, End = getNumRegBanks(); Idx != End; ++Idx) {
     assert(RegBanks[Idx] != nullptr && "Invalid RegisterBank");
+    assert(RegBanks[Idx]->isValid() && "RegisterBank should be valid");
+  }
 #endif // NDEBUG
 }
 
@@ -74,116 +76,13 @@ bool RegisterBankInfo::verify(const TargetRegisterInfo &TRI) const {
     const RegisterBank &RegBank = getRegBank(Idx);
     assert(Idx == RegBank.getID() &&
            "ID does not match the index in the array");
-    dbgs() << "Verify " << RegBank << '\n';
+    DEBUG(dbgs() << "Verify " << RegBank << '\n');
     assert(RegBank.verify(TRI) && "RegBank is invalid");
   }
 #endif // NDEBUG
   return true;
 }
 
-void RegisterBankInfo::createRegisterBank(unsigned ID, const char *Name) {
-  DEBUG(dbgs() << "Create register bank: " << ID << " with name \"" << Name
-               << "\"\n");
-  RegisterBank &RegBank = getRegBank(ID);
-  assert(RegBank.getID() == RegisterBank::InvalidID &&
-         "A register bank should be created only once");
-  RegBank.ID = ID;
-  RegBank.Name = Name;
-}
-
-void RegisterBankInfo::addRegBankCoverage(unsigned ID, unsigned RCId,
-                                          const TargetRegisterInfo &TRI) {
-  RegisterBank &RB = getRegBank(ID);
-  unsigned NbOfRegClasses = TRI.getNumRegClasses();
-
-  DEBUG(dbgs() << "Add coverage for: " << RB << '\n');
-
-  // Check if RB is underconstruction.
-  if (!RB.isValid())
-    RB.ContainedRegClasses.resize(NbOfRegClasses);
-  else if (RB.covers(*TRI.getRegClass(RCId)))
-    // If RB already covers this register class, there is nothing
-    // to do.
-    return;
-
-  BitVector &Covered = RB.ContainedRegClasses;
-  SmallVector<unsigned, 8> WorkList;
-
-  WorkList.push_back(RCId);
-  Covered.set(RCId);
-
-  unsigned &MaxSize = RB.Size;
-  do {
-    unsigned RCId = WorkList.pop_back_val();
-
-    const TargetRegisterClass &CurRC = *TRI.getRegClass(RCId);
-
-    DEBUG(dbgs() << "Examine: " << TRI.getRegClassName(&CurRC)
-                 << "(Size*8: " << (CurRC.getSize() * 8) << ")\n");
-
-    // Remember the biggest size in bits.
-    MaxSize = std::max(MaxSize, CurRC.getSize() * 8);
-
-    // Walk through all sub register classes and push them into the worklist.
-    bool First = true;
-    for (BitMaskClassIterator It(CurRC.getSubClassMask(), TRI); It.isValid();
-         ++It) {
-      unsigned SubRCId = It.getID();
-      if (!Covered.test(SubRCId)) {
-        if (First)
-          DEBUG(dbgs() << "  Enqueue sub-class: ");
-        DEBUG(dbgs() << TRI.getRegClassName(TRI.getRegClass(SubRCId)) << ", ");
-        WorkList.push_back(SubRCId);
-        // Remember that we saw the sub class.
-        Covered.set(SubRCId);
-        First = false;
-      }
-    }
-    if (!First)
-      DEBUG(dbgs() << '\n');
-
-    // Push also all the register classes that can be accessed via a
-    // subreg index, i.e., its subreg-class (which is different than
-    // its subclass).
-    //
-    // Note: It would probably be faster to go the other way around
-    // and have this method add only super classes, since this
-    // information is available in a more efficient way. However, it
-    // feels less natural for the client of this APIs plus we will
-    // TableGen the whole bitset at some point, so compile time for
-    // the initialization is not very important.
-    First = true;
-    for (unsigned SubRCId = 0; SubRCId < NbOfRegClasses; ++SubRCId) {
-      if (Covered.test(SubRCId))
-        continue;
-      bool Pushed = false;
-      const TargetRegisterClass *SubRC = TRI.getRegClass(SubRCId);
-      for (SuperRegClassIterator SuperRCIt(SubRC, &TRI); SuperRCIt.isValid();
-           ++SuperRCIt) {
-        if (Pushed)
-          break;
-        for (BitMaskClassIterator It(SuperRCIt.getMask(), TRI); It.isValid();
-             ++It) {
-          unsigned SuperRCId = It.getID();
-          if (SuperRCId == RCId) {
-            if (First)
-              DEBUG(dbgs() << "  Enqueue subreg-class: ");
-            DEBUG(dbgs() << TRI.getRegClassName(SubRC) << ", ");
-            WorkList.push_back(SubRCId);
-            // Remember that we saw the sub class.
-            Covered.set(SubRCId);
-            Pushed = true;
-            First = false;
-            break;
-          }
-        }
-      }
-    }
-    if (!First)
-      DEBUG(dbgs() << '\n');
-  } while (!WorkList.empty());
-}
-
 const RegisterBank *
 RegisterBankInfo::getRegBank(unsigned Reg, const MachineRegisterInfo &MRI,
                              const TargetRegisterInfo &TRI) const {

contrib/llvm/lib/CodeGen/MachineInstr.cpp

@@ -1840,7 +1840,8 @@ void MachineInstr::print(raw_ostream &OS, ModuleSlotTracker &MST,
         OS << "!\"" << DIV->getName() << '\"';
       else
         MO.print(OS, MST, TRI);
-    } else if (TRI && (isInsertSubreg() || isRegSequence()) && MO.isImm()) {
+    } else if (TRI && (isInsertSubreg() || isRegSequence() ||
+                       (isSubregToReg() && i == 3)) && MO.isImm()) {
       OS << TRI->getSubRegIndexName(MO.getImm());
     } else if (i == AsmDescOp && MO.isImm()) {
       // Pretty print the inline asm operand descriptor.

contrib/llvm/lib/CodeGen/PeepholeOptimizer.cpp

@@ -1715,7 +1715,8 @@ ValueTrackerResult ValueTracker::getNextSourceFromBitcast() {
   // Bitcasts with more than one def are not supported.
   if (Def->getDesc().getNumDefs() != 1)
     return ValueTrackerResult();
-  if (Def->getOperand(DefIdx).getSubReg() != DefSubReg)
+  const MachineOperand DefOp = Def->getOperand(DefIdx);
+  if (DefOp.getSubReg() != DefSubReg)
     // If we look for a different subreg, it means we want a subreg of the src.
     // Bails as we do not support composing subregs yet.
     return ValueTrackerResult();
@@ -1735,6 +1736,14 @@ ValueTrackerResult ValueTracker::getNextSourceFromBitcast() {
       return ValueTrackerResult();
     SrcIdx = OpIdx;
   }
+
+  // Stop when any user of the bitcast is a SUBREG_TO_REG, replacing with a COPY
+  // will break the assumed guarantees for the upper bits.
+  for (const MachineInstr &UseMI : MRI.use_nodbg_instructions(DefOp.getReg())) {
+    if (UseMI.isSubregToReg())
+      return ValueTrackerResult();
+  }
+
   const MachineOperand &Src = Def->getOperand(SrcIdx);
   return ValueTrackerResult(Src.getReg(), Src.getSubReg());
 }

contrib/llvm/lib/CodeGen/ScheduleDAG.cpp

@@ -310,19 +310,19 @@ void SUnit::biasCriticalPath() {
 }
 
 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
-static void dumpSUIdentifier(const ScheduleDAG &DAG, const SUnit &SU) {
-  if (&SU == &DAG.ExitSU)
-    dbgs() << "ExitSU";
-  else if (&SU == &DAG.EntrySU)
-    dbgs() << "EntrySU";
+void SUnit::print(raw_ostream &OS, const ScheduleDAG *DAG) const {
+  if (this == &DAG->ExitSU)
+    OS << "ExitSU";
+  else if (this == &DAG->EntrySU)
+    OS << "EntrySU";
   else
-    dbgs() << "SU(" << SU.NodeNum << ")";
+    OS << "SU(" << NodeNum << ")";
 }
 
 /// SUnit - Scheduling unit. It's an wrapper around either a single SDNode or
 /// a group of nodes flagged together.
 void SUnit::dump(const ScheduleDAG *G) const {
-  dumpSUIdentifier(*G, *this);
+  print(dbgs(), G);
   dbgs() << ": ";
   G->dumpNode(this);
 }
@@ -352,7 +352,7 @@ void SUnit::dumpAll(const ScheduleDAG *G) const {
       case SDep::Output: dbgs() << "out  "; break;
       case SDep::Order:  dbgs() << "ord  "; break;
       }
-      dumpSUIdentifier(*G, *I->getSUnit());
+      I->getSUnit()->print(dbgs(), G);
       if (I->isArtificial())
         dbgs() << " *";
       dbgs() << ": Latency=" << I->getLatency();
@@ -372,7 +372,7 @@ void SUnit::dumpAll(const ScheduleDAG *G) const {
       case SDep::Output: dbgs() << "out  "; break;
       case SDep::Order:  dbgs() << "ord  "; break;
       }
-      dumpSUIdentifier(*G, *I->getSUnit());
+      I->getSUnit()->print(dbgs(), G);
       if (I->isArtificial())
         dbgs() << " *";
       dbgs() << ": Latency=" << I->getLatency();

contrib/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp

@@ -5361,8 +5361,9 @@ SDValue DAGCombiner::visitSELECT(SDNode *N) {
     // fold (select false, X, Y) -> Y
     return !N0C->isNullValue() ? N1 : N2;
   }
-  // fold (select C, 1, X) -> (or C, X)
-  if (VT == MVT::i1 && isOneConstant(N1))
+  // fold (select X, X, Y) -> (or X, Y)
+  // fold (select X, 1, Y) -> (or C, Y)
+  if (VT == VT0 && VT == MVT::i1 && (N0 == N1 || isOneConstant(N1)))
     return DAG.getNode(ISD::OR, SDLoc(N), VT, N0, N2);
 
   if (SDValue V = foldSelectOfConstants(N))
@@ -5380,16 +5381,9 @@ SDValue DAGCombiner::visitSELECT(SDNode *N) {
     AddToWorklist(NOTNode.getNode());
     return DAG.getNode(ISD::OR, SDLoc(N), VT, NOTNode, N1);
   }
-  // fold (select C, X, 0) -> (and C, X)
-  if (VT == MVT::i1 && isNullConstant(N2))
-    return DAG.getNode(ISD::AND, SDLoc(N), VT, N0, N1);
-  // fold (select X, X, Y) -> (or X, Y)
-  // fold (select X, 1, Y) -> (or X, Y)
-  if (VT == MVT::i1 && (N0 == N1 || isOneConstant(N1)))
-    return DAG.getNode(ISD::OR, SDLoc(N), VT, N0, N2);
   // fold (select X, Y, X) -> (and X, Y)
   // fold (select X, Y, 0) -> (and X, Y)
-  if (VT == MVT::i1 && (N0 == N2 || isNullConstant(N2)))
+  if (VT == VT0 && VT == MVT::i1 && (N0 == N2 || isNullConstant(N2)))
     return DAG.getNode(ISD::AND, SDLoc(N), VT, N0, N1);
 
   // If we can fold this based on the true/false value, do so.
@@ -5470,7 +5464,6 @@ SDValue DAGCombiner::visitSELECT(SDNode *N) {
   }
 
   // select (xor Cond, 1), X, Y -> select Cond, Y, X
-  // select (xor Cond, 0), X, Y -> selext Cond, X, Y
   if (VT0 == MVT::i1) {
     if (N0->getOpcode() == ISD::XOR) {
       if (auto *C = dyn_cast<ConstantSDNode>(N0->getOperand(1))) {
@@ -5478,9 +5471,6 @@ SDValue DAGCombiner::visitSELECT(SDNode *N) {
         if (C->isOne())
           return DAG.getNode(ISD::SELECT, SDLoc(N), N1.getValueType(),
                              Cond0, N2, N1);
-        else
-          return DAG.getNode(ISD::SELECT, SDLoc(N), N1.getValueType(),
-                             Cond0, N1, N2);
       }
     }
   }
@@ -8136,7 +8126,8 @@ SDValue DAGCombiner::visitFADDForFMACombine(SDNode *N) {
   if ((AllowFusion || HasFMAD)  && Aggressive) {
     // fold (fadd (fma x, y, (fmul u, v)), z) -> (fma x, y (fma u, v, z))
     if (N0.getOpcode() == PreferredFusedOpcode &&
-        N0.getOperand(2).getOpcode() == ISD::FMUL) {
+        N0.getOperand(2).getOpcode() == ISD::FMUL &&
+        N0->hasOneUse() && N0.getOperand(2)->hasOneUse()) {
       return DAG.getNode(PreferredFusedOpcode, SL, VT,
                          N0.getOperand(0), N0.getOperand(1),
                          DAG.getNode(PreferredFusedOpcode, SL, VT,
@@ -8147,7 +8138,8 @@ SDValue DAGCombiner::visitFADDForFMACombine(SDNode *N) {
 
     // fold (fadd x, (fma y, z, (fmul u, v)) -> (fma y, z (fma u, v, x))
     if (N1->getOpcode() == PreferredFusedOpcode &&
-        N1.getOperand(2).getOpcode() == ISD::FMUL) {
+        N1.getOperand(2).getOpcode() == ISD::FMUL &&
+        N1->hasOneUse() && N1.getOperand(2)->hasOneUse()) {
       return DAG.getNode(PreferredFusedOpcode, SL, VT,
                          N1.getOperand(0), N1.getOperand(1),
                          DAG.getNode(PreferredFusedOpcode, SL, VT,
@@ -8379,7 +8371,8 @@ SDValue DAGCombiner::visitFSUBForFMACombine(SDNode *N) {
     // fold (fsub (fma x, y, (fmul u, v)), z)
     //   -> (fma x, y (fma u, v, (fneg z)))
     if (N0.getOpcode() == PreferredFusedOpcode &&
-        N0.getOperand(2).getOpcode() == ISD::FMUL) {
+        N0.getOperand(2).getOpcode() == ISD::FMUL &&
+        N0->hasOneUse() && N0.getOperand(2)->hasOneUse()) {
       return DAG.getNode(PreferredFusedOpcode, SL, VT,
                          N0.getOperand(0), N0.getOperand(1),
                          DAG.getNode(PreferredFusedOpcode, SL, VT,

contrib/llvm/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp

@@ -330,8 +330,6 @@ SDValue SelectionDAGLegalize::PerformInsertVectorEltInMemory(SDValue Vec,
   // supported by the target.
   EVT VT    = Tmp1.getValueType();
   EVT EltVT = VT.getVectorElementType();
-  EVT IdxVT = Tmp3.getValueType();
-  EVT PtrVT = TLI.getPointerTy(DAG.getDataLayout());
   SDValue StackPtr = DAG.CreateStackTemporary(VT);
 
   int SPFI = cast<FrameIndexSDNode>(StackPtr.getNode())->getIndex();
@@ -341,13 +339,8 @@ SDValue SelectionDAGLegalize::PerformInsertVectorEltInMemory(SDValue Vec,
       DAG.getEntryNode(), dl, Tmp1, StackPtr,
       MachinePointerInfo::getFixedStack(DAG.getMachineFunction(), SPFI));
 
-  // Truncate or zero extend offset to target pointer type.
-  Tmp3 = DAG.getZExtOrTrunc(Tmp3, dl, PtrVT);
-  // Add the offset to the index.
-  unsigned EltSize = EltVT.getSizeInBits()/8;
-  Tmp3 = DAG.getNode(ISD::MUL, dl, IdxVT, Tmp3,
-                     DAG.getConstant(EltSize, dl, IdxVT));
-  SDValue StackPtr2 = DAG.getNode(ISD::ADD, dl, IdxVT, Tmp3, StackPtr);
+  SDValue StackPtr2 = TLI.getVectorElementPointer(DAG, StackPtr, VT, Tmp3);
+
   // Store the scalar value.
   Ch = DAG.getTruncStore(Ch, dl, Tmp2, StackPtr2, MachinePointerInfo(), EltVT);
   // Load the updated vector.
@@ -1209,20 +1202,16 @@ SDValue SelectionDAGLegalize::ExpandExtractFromVectorThroughStack(SDValue Op) {
     }
   }
 
+  EVT VecVT = Vec.getValueType();
+
   if (!Ch.getNode()) {
     // Store the value to a temporary stack slot, then LOAD the returned part.
-    StackPtr = DAG.CreateStackTemporary(Vec.getValueType());
+    StackPtr = DAG.CreateStackTemporary(VecVT);
     Ch = DAG.getStore(DAG.getEntryNode(), dl, Vec, StackPtr,
                       MachinePointerInfo());
   }
 
-  // Add the offset to the index.
-  unsigned EltSize = Vec.getScalarValueSizeInBits() / 8;
-  Idx = DAG.getNode(ISD::MUL, dl, Idx.getValueType(), Idx,
-                    DAG.getConstant(EltSize, SDLoc(Vec), Idx.getValueType()));
-
-  Idx = DAG.getZExtOrTrunc(Idx, dl, TLI.getPointerTy(DAG.getDataLayout()));
-  StackPtr = DAG.getNode(ISD::ADD, dl, Idx.getValueType(), Idx, StackPtr);
+  StackPtr = TLI.getVectorElementPointer(DAG, StackPtr, VecVT, Idx);
 
   SDValue NewLoad;
 
@@ -1232,7 +1221,7 @@ SDValue SelectionDAGLegalize::ExpandExtractFromVectorThroughStack(SDValue Op) {
   else
     NewLoad = DAG.getExtLoad(ISD::EXTLOAD, dl, Op.getValueType(), Ch, StackPtr,
                              MachinePointerInfo(),
-                             Vec.getValueType().getVectorElementType());
+                             VecVT.getVectorElementType());
 
   // Replace the chain going out of the store, by the one out of the load.
   DAG.ReplaceAllUsesOfValueWith(Ch, SDValue(NewLoad.getNode(), 1));
@@ -1256,8 +1245,8 @@ SDValue SelectionDAGLegalize::ExpandInsertToVectorThroughStack(SDValue Op) {
   SDLoc dl(Op);
 
   // Store the value to a temporary stack slot, then LOAD the returned part.
-
-  SDValue StackPtr = DAG.CreateStackTemporary(Vec.getValueType());
+  EVT VecVT = Vec.getValueType();
+  SDValue StackPtr = DAG.CreateStackTemporary(VecVT);
   int FI = cast<FrameIndexSDNode>(StackPtr.getNode())->getIndex();
   MachinePointerInfo PtrInfo =
       MachinePointerInfo::getFixedStack(DAG.getMachineFunction(), FI);
@@ -1266,16 +1255,7 @@ SDValue SelectionDAGLegalize::ExpandInsertToVectorThroughStack(SDValue Op) {
   SDValue Ch = DAG.getStore(DAG.getEntryNode(), dl, Vec, StackPtr, PtrInfo);
 
   // Then store the inserted part.
-
-  // Add the offset to the index.
-  unsigned EltSize = Vec.getScalarValueSizeInBits() / 8;
-
-  Idx = DAG.getNode(ISD::MUL, dl, Idx.getValueType(), Idx,
-                    DAG.getConstant(EltSize, SDLoc(Vec), Idx.getValueType()));
-  Idx = DAG.getZExtOrTrunc(Idx, dl, TLI.getPointerTy(DAG.getDataLayout()));
-
-  SDValue SubStackPtr = DAG.getNode(ISD::ADD, dl, Idx.getValueType(), Idx,
-                                    StackPtr);
+  SDValue SubStackPtr = TLI.getVectorElementPointer(DAG, StackPtr, VecVT, Idx);
 
   // Store the subvector.
   Ch = DAG.getStore(Ch, dl, Part, SubStackPtr, MachinePointerInfo());

contrib/llvm/lib/CodeGen/SelectionDAG/LegalizeIntegerTypes.cpp

@@ -57,8 +57,6 @@ void DAGTypeLegalizer::PromoteIntegerResult(SDNode *N, unsigned ResNo) {
   case ISD::BSWAP:       Res = PromoteIntRes_BSWAP(N); break;
   case ISD::BUILD_PAIR:  Res = PromoteIntRes_BUILD_PAIR(N); break;
   case ISD::Constant:    Res = PromoteIntRes_Constant(N); break;
-  case ISD::CONVERT_RNDSAT:
-                         Res = PromoteIntRes_CONVERT_RNDSAT(N); break;
   case ISD::CTLZ_ZERO_UNDEF:
   case ISD::CTLZ:        Res = PromoteIntRes_CTLZ(N); break;
   case ISD::CTPOP:       Res = PromoteIntRes_CTPOP(N); break;
@@ -354,18 +352,6 @@ SDValue DAGTypeLegalizer::PromoteIntRes_Constant(SDNode *N) {
   return Result;
 }
 
-SDValue DAGTypeLegalizer::PromoteIntRes_CONVERT_RNDSAT(SDNode *N) {
-  ISD::CvtCode CvtCode = cast<CvtRndSatSDNode>(N)->getCvtCode();
-  assert ((CvtCode == ISD::CVT_SS || CvtCode == ISD::CVT_SU ||
-           CvtCode == ISD::CVT_US || CvtCode == ISD::CVT_UU ||
-           CvtCode == ISD::CVT_SF || CvtCode == ISD::CVT_UF) &&
-          "can only promote integers");
-  EVT OutVT = TLI.getTypeToTransformTo(*DAG.getContext(), N->getValueType(0));
-  return DAG.getConvertRndSat(OutVT, SDLoc(N), N->getOperand(0),
-                              N->getOperand(1), N->getOperand(2),
-                              N->getOperand(3), N->getOperand(4), CvtCode);
-}
-
 SDValue DAGTypeLegalizer::PromoteIntRes_CTLZ(SDNode *N) {
   // Zero extend to the promoted type and do the count there.
   SDValue Op = ZExtPromotedInteger(N->getOperand(0));
@@ -512,7 +498,7 @@ SDValue DAGTypeLegalizer::PromoteIntRes_MGATHER(MaskedGatherSDNode *N) {
                    N->getIndex()};
   SDValue Res = DAG.getMaskedGather(DAG.getVTList(NVT, MVT::Other),
                                     N->getMemoryVT(), dl, Ops,
-                                    N->getMemOperand()); 
+                                    N->getMemOperand());
   // Legalize the chain result - switch anything that used the old chain to
   // use the new one.
   ReplaceValueWith(SDValue(N, 1), Res.getValue(1));
@@ -887,8 +873,6 @@ bool DAGTypeLegalizer::PromoteIntegerOperand(SDNode *N, unsigned OpNo) {
   case ISD::BUILD_VECTOR: Res = PromoteIntOp_BUILD_VECTOR(N); break;
   case ISD::CONCAT_VECTORS: Res = PromoteIntOp_CONCAT_VECTORS(N); break;
   case ISD::EXTRACT_VECTOR_ELT: Res = PromoteIntOp_EXTRACT_VECTOR_ELT(N); break;
-  case ISD::CONVERT_RNDSAT:
-                          Res = PromoteIntOp_CONVERT_RNDSAT(N); break;
   case ISD::INSERT_VECTOR_ELT:
                           Res = PromoteIntOp_INSERT_VECTOR_ELT(N, OpNo);break;
   case ISD::SCALAR_TO_VECTOR:
@@ -1068,18 +1052,6 @@ SDValue DAGTypeLegalizer::PromoteIntOp_BUILD_VECTOR(SDNode *N) {
   return SDValue(DAG.UpdateNodeOperands(N, NewOps), 0);
 }
 
-SDValue DAGTypeLegalizer::PromoteIntOp_CONVERT_RNDSAT(SDNode *N) {
-  ISD::CvtCode CvtCode = cast<CvtRndSatSDNode>(N)->getCvtCode();
-  assert ((CvtCode == ISD::CVT_SS || CvtCode == ISD::CVT_SU ||
-           CvtCode == ISD::CVT_US || CvtCode == ISD::CVT_UU ||
-           CvtCode == ISD::CVT_FS || CvtCode == ISD::CVT_FU) &&
-           "can only promote integer arguments");
-  SDValue InOp = GetPromotedInteger(N->getOperand(0));
-  return DAG.getConvertRndSat(N->getValueType(0), SDLoc(N), InOp,
-                              N->getOperand(1), N->getOperand(2),
-                              N->getOperand(3), N->getOperand(4), CvtCode);
-}
-
 SDValue DAGTypeLegalizer::PromoteIntOp_INSERT_VECTOR_ELT(SDNode *N,
                                                          unsigned OpNo) {
   if (OpNo == 1) {

contrib/llvm/lib/CodeGen/SelectionDAG/LegalizeTypes.cpp

@@ -1021,22 +1021,6 @@ void DAGTypeLegalizer::GetPairElements(SDValue Pair,
                    DAG.getIntPtrConstant(1, dl));
 }
 
-SDValue DAGTypeLegalizer::GetVectorElementPointer(SDValue VecPtr, EVT EltVT,
-                                                  SDValue Index) {
-  SDLoc dl(Index);
-  // Make sure the index type is big enough to compute in.
-  Index = DAG.getZExtOrTrunc(Index, dl, TLI.getPointerTy(DAG.getDataLayout()));
-
-  // Calculate the element offset and add it to the pointer.
-  unsigned EltSize = EltVT.getSizeInBits() / 8; // FIXME: should be ABI size.
-  assert(EltSize * 8 == EltVT.getSizeInBits() &&
-         "Converting bits to bytes lost precision");
-
-  Index = DAG.getNode(ISD::MUL, dl, Index.getValueType(), Index,
-                      DAG.getConstant(EltSize, dl, Index.getValueType()));
-  return DAG.getNode(ISD::ADD, dl, Index.getValueType(), Index, VecPtr);
-}
-
 /// Build an integer with low bits Lo and high bits Hi.
 SDValue DAGTypeLegalizer::JoinIntegers(SDValue Lo, SDValue Hi) {
   // Arbitrarily use dlHi for result SDLoc