vendor/llvm-project/llvmorg-14-init-11187-g222442ec2d71

96 files changed, 2018 insertions, 437 deletions

diff --git a/clang/include/clang/AST/Decl.h b/clang/include/clang/AST/Decl.h
index 85a3a8ab6970..2eacf1105c18 100644
--- a/clang/include/clang/AST/Decl.h
+++ b/clang/include/clang/AST/Decl.h
@@ -1840,7 +1840,8 @@ enum class MultiVersionKind {
   None,
   Target,
   CPUSpecific,
-  CPUDispatch
+  CPUDispatch,
+  TargetClones
 };
 
 /// Represents a function declaration or definition.
@@ -2459,6 +2460,10 @@ public:
   /// the target functionality.
   bool isTargetMultiVersion() const;
 
+  /// True if this function is a multiversioned dispatch function as a part of
+  /// the target-clones functionality.
+  bool isTargetClonesMultiVersion() const;
+
   /// \brief Get the associated-constraints of this function declaration.
   /// Currently, this will either be a vector of size 1 containing the
   /// trailing-requires-clause or an empty vector.
diff --git a/clang/include/clang/AST/DeclBase.h b/clang/include/clang/AST/DeclBase.h
index 18468c8ca1c4..2a0a19597391 100644
--- a/clang/include/clang/AST/DeclBase.h
+++ b/clang/include/clang/AST/DeclBase.h
@@ -352,7 +352,7 @@ protected:
                      DeclContext *Parent, std::size_t Extra = 0);
 
 private:
-  bool AccessDeclContextSanity() const;
+  bool AccessDeclContextCheck() const;
 
   /// Get the module ownership kind to use for a local lexical child of \p DC,
   /// which may be either a local or (rarely) an imported declaration.
@@ -472,11 +472,11 @@ public:
 
   void setAccess(AccessSpecifier AS) {
     Access = AS;
-    assert(AccessDeclContextSanity());
+    assert(AccessDeclContextCheck());
   }
 
   AccessSpecifier getAccess() const {
-    assert(AccessDeclContextSanity());
+    assert(AccessDeclContextCheck());
     return AccessSpecifier(Access);
   }
 
diff --git a/clang/include/clang/AST/Expr.h b/clang/include/clang/AST/Expr.h
index 991abef73363..2c63406fba18 100644
--- a/clang/include/clang/AST/Expr.h
+++ b/clang/include/clang/AST/Expr.h
@@ -6305,8 +6305,10 @@ public:
   bool isCmpXChg() const {
     return getOp() == AO__c11_atomic_compare_exchange_strong ||
            getOp() == AO__c11_atomic_compare_exchange_weak ||
+           getOp() == AO__hip_atomic_compare_exchange_strong ||
            getOp() == AO__opencl_atomic_compare_exchange_strong ||
            getOp() == AO__opencl_atomic_compare_exchange_weak ||
+           getOp() == AO__hip_atomic_compare_exchange_weak ||
            getOp() == AO__atomic_compare_exchange ||
            getOp() == AO__atomic_compare_exchange_n;
   }
@@ -6341,6 +6343,8 @@ public:
     auto Kind =
         (Op >= AO__opencl_atomic_load && Op <= AO__opencl_atomic_fetch_max)
             ? AtomicScopeModelKind::OpenCL
+        : (Op >= AO__hip_atomic_load && Op <= AO__hip_atomic_fetch_max)
+            ? AtomicScopeModelKind::HIP
             : AtomicScopeModelKind::None;
     return AtomicScopeModel::create(Kind);
   }
diff --git a/clang/include/clang/AST/GlobalDecl.h b/clang/include/clang/AST/GlobalDecl.h
index 8cb56fb4ae90..88abba28c991 100644
--- a/clang/include/clang/AST/GlobalDecl.h
+++ b/clang/include/clang/AST/GlobalDecl.h
@@ -18,6 +18,7 @@
 #include "clang/AST/DeclCXX.h"
 #include "clang/AST/DeclObjC.h"
 #include "clang/AST/DeclOpenMP.h"
+#include "clang/AST/DeclTemplate.h"
 #include "clang/Basic/ABI.h"
 #include "clang/Basic/LLVM.h"
 #include "llvm/ADT/DenseMapInfo.h"
@@ -129,8 +130,12 @@ public:
   }
 
   KernelReferenceKind getKernelReferenceKind() const {
-    assert(isa<FunctionDecl>(getDecl()) &&
-           cast<FunctionDecl>(getDecl())->hasAttr<CUDAGlobalAttr>() &&
+    assert(((isa<FunctionDecl>(getDecl()) &&
+             cast<FunctionDecl>(getDecl())->hasAttr<CUDAGlobalAttr>()) ||
+            (isa<FunctionTemplateDecl>(getDecl()) &&
+             cast<FunctionTemplateDecl>(getDecl())
+                 ->getTemplatedDecl()
+                 ->hasAttr<CUDAGlobalAttr>())) &&
            "Decl is not a GPU kernel!");
     return static_cast<KernelReferenceKind>(Value.getInt());
   }
diff --git a/clang/include/clang/ASTMatchers/ASTMatchers.h b/clang/include/clang/ASTMatchers/ASTMatchers.h
index d6e5b215462b..5221d05477d0 100644
--- a/clang/include/clang/ASTMatchers/ASTMatchers.h
+++ b/clang/include/clang/ASTMatchers/ASTMatchers.h
@@ -4227,8 +4227,8 @@ AST_MATCHER(VarDecl, isInitCapture) { return Node.isInitCapture(); }
 /// lambdaExpr(forEachLambdaCapture(
 ///     lambdaCapture(capturesVar(varDecl(hasType(isInteger()))))))
 /// will trigger two matches, binding for 'x' and 'y' respectively.
-AST_MATCHER_P(LambdaExpr, forEachLambdaCapture, LambdaCaptureMatcher,
-              InnerMatcher) {
+AST_MATCHER_P(LambdaExpr, forEachLambdaCapture,
+              internal::Matcher<LambdaCapture>, InnerMatcher) {
   BoundNodesTreeBuilder Result;
   bool Matched = false;
   for (const auto &Capture : Node.captures()) {
@@ -4655,7 +4655,8 @@ extern const internal::VariadicAllOfMatcher<LambdaCapture> lambdaCapture;
 /// lambdaExpr(hasAnyCapture(lambdaCapture())) and
 /// lambdaExpr(hasAnyCapture(lambdaCapture(refersToVarDecl(hasName("t")))))
 ///   both match `[=](){ return t; }`.
-AST_MATCHER_P(LambdaExpr, hasAnyCapture, LambdaCaptureMatcher, InnerMatcher) {
+AST_MATCHER_P(LambdaExpr, hasAnyCapture, internal::Matcher<LambdaCapture>,
+              InnerMatcher) {
   for (const LambdaCapture &Capture : Node.captures()) {
     clang::ast_matchers::internal::BoundNodesTreeBuilder Result(*Builder);
     if (InnerMatcher.matches(Capture, Finder, &Result)) {
diff --git a/clang/include/clang/Analysis/CFG.h b/clang/include/clang/Analysis/CFG.h
index 3b9b22e87f35..b8e453fcc235 100644
--- a/clang/include/clang/Analysis/CFG.h
+++ b/clang/include/clang/Analysis/CFG.h
@@ -515,7 +515,7 @@ public:
     /// of the most derived class while we're in the base class.
     VirtualBaseBranch,
 
-    /// Number of different kinds, for validity checks. We subtract 1 so that
+    /// Number of different kinds, for assertions. We subtract 1 so that
     /// to keep receiving compiler warnings when we don't cover all enum values
     /// in a switch.
     NumKindsMinusOne = VirtualBaseBranch
diff --git a/clang/include/clang/Analysis/FlowSensitive/DataflowAnalysis.h b/clang/include/clang/Analysis/FlowSensitive/DataflowAnalysis.h
new file mode 100644
index 000000000000..a5d4a5d6ba40
--- /dev/null
+++ b/clang/include/clang/Analysis/FlowSensitive/DataflowAnalysis.h
@@ -0,0 +1,134 @@
+//===- DataflowAnalysis.h ---------------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+//  This file defines base types and functions for building dataflow analyses
+//  that run over Control-Flow Graphs (CFGs).
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_CLANG_ANALYSIS_FLOWSENSITIVE_DATAFLOWANALYSIS_H
+#define LLVM_CLANG_ANALYSIS_FLOWSENSITIVE_DATAFLOWANALYSIS_H
+
+#include <iterator>
+#include <utility>
+#include <vector>
+
+#include "clang/AST/ASTContext.h"
+#include "clang/AST/Stmt.h"
+#include "clang/Analysis/CFG.h"
+#include "clang/Analysis/FlowSensitive/DataflowEnvironment.h"
+#include "clang/Analysis/FlowSensitive/TypeErasedDataflowAnalysis.h"
+#include "llvm/ADT/Any.h"
+#include "llvm/ADT/Optional.h"
+#include "llvm/ADT/STLExtras.h"
+
+namespace clang {
+namespace dataflow {
+
+/// Base class template for dataflow analyses built on a single lattice type.
+///
+/// Requirements:
+///
+///  `Derived` must be derived from a specialization of this class template and
+///  must provide the following public members:
+///   * `LatticeT initialElement()` - returns a lattice element that models the
+///     initial state of a basic block;
+///   * `LatticeT transfer(const Stmt *, const LatticeT &, Environment &)` -
+///     applies the analysis transfer function for a given statement and lattice
+///     element.
+///
+///  `LatticeT` is a bounded join-semilattice that is used by `Derived` and must
+///  provide the following public members:
+///   * `LatticeJoinEffect join(const LatticeT &)` - joins the object and the
+///     argument by computing their least upper bound, modifies the object if
+///     necessary, and returns an effect indicating whether any changes were
+///     made to it;
+///   * `bool operator==(const LatticeT &) const` - returns true if and only if
+///     the object is equal to the argument.
+template <typename Derived, typename LatticeT>
+class DataflowAnalysis : public TypeErasedDataflowAnalysis {
+public:
+  /// Bounded join-semilattice that is used in the analysis.
+  using Lattice = LatticeT;
+
+  explicit DataflowAnalysis(ASTContext &Context) : Context(Context) {}
+
+  ASTContext &getASTContext() final { return Context; }
+
+  TypeErasedLattice typeErasedInitialElement() final {
+    return {static_cast<Derived *>(this)->initialElement()};
+  }
+
+  LatticeJoinEffect joinTypeErased(TypeErasedLattice &E1,
+                                   const TypeErasedLattice &E2) final {
+    Lattice &L1 = llvm::any_cast<Lattice &>(E1.Value);
+    const Lattice &L2 = llvm::any_cast<const Lattice &>(E2.Value);
+    return L1.join(L2);
+  }
+
+  bool isEqualTypeErased(const TypeErasedLattice &E1,
+                         const TypeErasedLattice &E2) final {
+    const Lattice &L1 = llvm::any_cast<const Lattice &>(E1.Value);
+    const Lattice &L2 = llvm::any_cast<const Lattice &>(E2.Value);
+    return L1 == L2;
+  }
+
+  TypeErasedLattice transferTypeErased(const Stmt *Stmt,
+                                       const TypeErasedLattice &E,
+                                       Environment &Env) final {
+    const Lattice &L = llvm::any_cast<const Lattice &>(E.Value);
+    return {static_cast<Derived *>(this)->transfer(Stmt, L, Env)};
+  }
+
+private:
+  ASTContext &Context;
+};
+
+// Model of the program at a given program point.
+template <typename LatticeT> struct DataflowAnalysisState {
+  // Model of a program property.
+  LatticeT Lattice;
+
+  // Model of the state of the program (store and heap).
+  Environment Env;
+};
+
+/// Performs dataflow analysis and returns a mapping from basic block IDs to
+/// dataflow analysis states that model the respective basic blocks. Indices
+/// of the returned vector correspond to basic block IDs.
+///
+/// Requirements:
+///
+///  `Cfg` must have been built with `CFG::BuildOptions::setAllAlwaysAdd()` to
+///  ensure that all sub-expressions in a basic block are evaluated.
+template <typename AnalysisT>
+std::vector<llvm::Optional<DataflowAnalysisState<typename AnalysisT::Lattice>>>
+runDataflowAnalysis(const CFG &Cfg, AnalysisT &Analysis,
+                    const Environment &InitEnv) {
+  auto TypeErasedBlockStates =
+      runTypeErasedDataflowAnalysis(Cfg, Analysis, InitEnv);
+  std::vector<
+      llvm::Optional<DataflowAnalysisState<typename AnalysisT::Lattice>>>
+      BlockStates;
+  BlockStates.reserve(TypeErasedBlockStates.size());
+  llvm::transform(std::move(TypeErasedBlockStates),
+                  std::back_inserter(BlockStates), [](auto &OptState) {
+                    return std::move(OptState).map([](auto &&State) {
+                      return DataflowAnalysisState<typename AnalysisT::Lattice>{
+                          llvm::any_cast<typename AnalysisT::Lattice>(
+                              std::move(State.Lattice.Value)),
+                          std::move(State.Env)};
+                    });
+                  });
+  return BlockStates;
+}
+
+} // namespace dataflow
+} // namespace clang
+
+#endif // LLVM_CLANG_ANALYSIS_FLOWSENSITIVE_DATAFLOWANALYSIS_H
diff --git a/clang/include/clang/Analysis/FlowSensitive/DataflowEnvironment.h b/clang/include/clang/Analysis/FlowSensitive/DataflowEnvironment.h
new file mode 100644
index 000000000000..69a5c2e47b66
--- /dev/null
+++ b/clang/include/clang/Analysis/FlowSensitive/DataflowEnvironment.h
@@ -0,0 +1,27 @@
+//===-- DataflowEnvironment.h -----------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+//  This file defines an Environment class that is used by dataflow analyses
+//  that run over Control-Flow Graphs (CFGs) to keep track of the state of the
+//  program at given program points.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_CLANG_ANALYSIS_FLOWSENSITIVE_DATAFLOWENVIRONMENT_H
+#define LLVM_CLANG_ANALYSIS_FLOWSENSITIVE_DATAFLOWENVIRONMENT_H
+
+namespace clang {
+namespace dataflow {
+
+/// Holds the state of the program (store and heap) at a given program point.
+class Environment {};
+
+} // namespace dataflow
+} // namespace clang
+
+#endif // LLVM_CLANG_ANALYSIS_FLOWSENSITIVE_DATAFLOWENVIRONMENT_H
diff --git a/clang/include/clang/Analysis/FlowSensitive/DataflowLattice.h b/clang/include/clang/Analysis/FlowSensitive/DataflowLattice.h
new file mode 100644
index 000000000000..37d2e0200410
--- /dev/null
+++ b/clang/include/clang/Analysis/FlowSensitive/DataflowLattice.h
@@ -0,0 +1,29 @@
+//===- DataflowLattice.h ----------------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+//  This file defines base types for building lattices to be used in dataflow
+//  analyses that run over Control-Flow Graphs (CFGs).
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_CLANG_ANALYSIS_FLOWSENSITIVE_DATAFLOWLATTICE_H
+#define LLVM_CLANG_ANALYSIS_FLOWSENSITIVE_DATAFLOWLATTICE_H
+
+namespace clang {
+namespace dataflow {
+
+/// Effect indicating whether a lattice join operation resulted in a new value.
+enum class LatticeJoinEffect {
+  Unchanged,
+  Changed,
+};
+
+} // namespace dataflow
+} // namespace clang
+
+#endif // LLVM_CLANG_ANALYSIS_FLOWSENSITIVE_DATAFLOWLATTICE_H
diff --git a/clang/include/clang/Analysis/FlowSensitive/TypeErasedDataflowAnalysis.h b/clang/include/clang/Analysis/FlowSensitive/TypeErasedDataflowAnalysis.h
new file mode 100644
index 000000000000..9448b911f471
--- /dev/null
+++ b/clang/include/clang/Analysis/FlowSensitive/TypeErasedDataflowAnalysis.h
@@ -0,0 +1,95 @@
+//===- TypeErasedDataflowAnalysis.h -----------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+//  This file defines type-erased base types and functions for building dataflow
+//  analyses that run over Control-Flow Graphs (CFGs).
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_CLANG_ANALYSIS_FLOWSENSITIVE_TYPEERASEDDATAFLOWANALYSIS_H
+#define LLVM_CLANG_ANALYSIS_FLOWSENSITIVE_TYPEERASEDDATAFLOWANALYSIS_H
+
+#include <vector>
+
+#include "clang/AST/ASTContext.h"
+#include "clang/AST/Stmt.h"
+#include "clang/Analysis/CFG.h"
+#include "clang/Analysis/FlowSensitive/DataflowEnvironment.h"
+#include "clang/Analysis/FlowSensitive/DataflowLattice.h"
+#include "llvm/ADT/Any.h"
+#include "llvm/ADT/Optional.h"
+
+namespace clang {
+namespace dataflow {
+
+/// Type-erased lattice element container.
+///
+/// Requirements:
+///
+///  The type of the object stored in the container must be a bounded
+///  join-semilattice.
+struct TypeErasedLattice {
+  llvm::Any Value;
+};
+
+/// Type-erased base class for dataflow analyses built on a single lattice type.
+class TypeErasedDataflowAnalysis {
+public:
+  virtual ~TypeErasedDataflowAnalysis() {}
+
+  /// Returns the `ASTContext` that is used by the analysis.
+  virtual ASTContext &getASTContext() = 0;
+
+  /// Returns a type-erased lattice element that models the initial state of a
+  /// basic block.
+  virtual TypeErasedLattice typeErasedInitialElement() = 0;
+
+  /// Joins two type-erased lattice elements by computing their least upper
+  /// bound. Places the join result in the left element and returns an effect
+  /// indicating whether any changes were made to it.
+  virtual LatticeJoinEffect joinTypeErased(TypeErasedLattice &,
+                                           const TypeErasedLattice &) = 0;
+
+  /// Returns true if and only if the two given type-erased lattice elements are
+  /// equal.
+  virtual bool isEqualTypeErased(const TypeErasedLattice &,
+                                 const TypeErasedLattice &) = 0;
+
+  /// Applies the analysis transfer function for a given statement and
+  /// type-erased lattice element.
+  virtual TypeErasedLattice transferTypeErased(const Stmt *,
+                                               const TypeErasedLattice &,
+                                               Environment &) = 0;
+};
+
+/// Type-erased model of the program at a given program point.
+struct TypeErasedDataflowAnalysisState {
+  /// Type-erased model of a program property.
+  TypeErasedLattice Lattice;
+
+  /// Model of the state of the program (store and heap).
+  Environment Env;
+};
+
+/// Performs dataflow analysis and returns a mapping from basic block IDs to
+/// dataflow analysis states that model the respective basic blocks. Indices
+/// of the returned vector correspond to basic block IDs.
+///
+/// Requirements:
+///
+///  `Cfg` must have been built with `CFG::BuildOptions::setAllAlwaysAdd()` to
+///  ensure that all sub-expressions in a basic block are evaluated.
+std::vector<llvm::Optional<TypeErasedDataflowAnalysisState>>
+runTypeErasedDataflowAnalysis(const CFG &Cfg,
+                              TypeErasedDataflowAnalysis &Analysis,
+                              const Environment &InitEnv);
+
+} // namespace dataflow
+} // namespace clang
+
+#endif // LLVM_CLANG_ANALYSIS_FLOWSENSITIVE_TYPEERASEDDATAFLOWANALYSIS_H
diff --git a/clang/include/clang/Basic/Attr.td b/clang/include/clang/Basic/Attr.td
index 39588d94cf09..fab3f3edfb83 100644
--- a/clang/include/clang/Basic/Attr.td
+++ b/clang/include/clang/Basic/Attr.td
@@ -2677,6 +2677,40 @@ def Target : InheritableAttr {
   }];
 }
 
+def TargetClones : InheritableAttr {
+  let Spellings = [GCC<"target_clones">];
+  let Args = [VariadicStringArgument<"featuresStrs">];
+  let Documentation = [TargetClonesDocs];
+  let Subjects = SubjectList<[Function], ErrorDiag>;
+  let AdditionalMembers = [{
+    StringRef getFeatureStr(unsigned Index) const {
+      return *(featuresStrs_begin() + Index);
+    }
+    // 'default' is always moved to the end, so it isn't considered
+    // when mangling the index.
+    unsigned getMangledIndex(unsigned Index) const {
+      if (getFeatureStr(Index) == "default")
+        return std::count_if(featuresStrs_begin(), featuresStrs_end(),
+                              [](StringRef S) { return S != "default"; });
+
+      return std::count_if(featuresStrs_begin(), featuresStrs_begin() + Index,
+                           [](StringRef S) { return S != "default"; });
+    }
+
+    // True if this is the first of this version to appear in the config string.
+    // This is used to make sure we don't try to emit this function multiple
+    // times.
+    bool isFirstOfVersion(unsigned Index) const {
+      StringRef FeatureStr(getFeatureStr(Index));
+      return 0 == std::count_if(
+                      featuresStrs_begin(), featuresStrs_begin() + Index,
+                      [FeatureStr](StringRef S) { return S == FeatureStr; });
+    }
+  }];
+}
+
+def : MutualExclusions<[TargetClones, Target, CPUDispatch, CPUSpecific]>;
+
 def MinVectorWidth : InheritableAttr {
   let Spellings = [Clang<"min_vector_width">];
   let Args = [UnsignedArgument<"VectorWidth">];
diff --git a/clang/include/clang/Basic/AttrDocs.td b/clang/include/clang/Basic/AttrDocs.td
index e7afb3699eb1..10cce4c2d689 100644
--- a/clang/include/clang/Basic/AttrDocs.td
+++ b/clang/include/clang/Basic/AttrDocs.td
@@ -2233,6 +2233,40 @@ Additionally, a function may not become multiversioned after its first use.
 }];
 }
 
+def TargetClonesDocs : Documentation {
+  let Category = DocCatFunction;
+  let Content = [{
+Clang supports the ``target_clones("OPTIONS")`` attribute. This attribute may be
+attached to a function declaration and causes function multiversioning, where
+multiple versions of the function will be emitted with different code
+generation options.  Additionally, these versions will be resolved at runtime
+based on the priority of their attribute options. All ``target_clone`` functions
+are considered multiversioned functions.
+
+All multiversioned functions must contain a ``default`` (fallback)
+implementation, otherwise usages of the function are considered invalid.
+Additionally, a function may not become multiversioned after its first use.
+
+The options to ``target_clones`` can either be a target-specific architecture
+(specified as ``arch=CPU``), or one of a list of subtarget features.
+
+Example "subtarget features" from the x86 backend include: "mmx", "sse", "sse4.2",
+"avx", "xop" and largely correspond to the machine specific options handled by
+the front end.
+
+The versions can either be listed as a comma-separated sequence of string
+literals or as a single string literal containing a comma-separated list of
+versions.  For compatibility with GCC, the two formats can be mixed.  For
+example, the following will emit 4 versions of the function:
+
+  .. code-block:: c++
+
+    __attribute__((target_clones("arch=atom,avx2","arch=ivybridge","default")))
+    void foo() {}
+
+}];
+}
+
 def MinVectorWidthDocs : Documentation {
   let Category = DocCatFunction;
   let Content = [{
diff --git a/clang/include/clang/Basic/Builtins.def b/clang/include/clang/Basic/Builtins.def
index b05777889e79..ad8b66aa490b 100644
--- a/clang/include/clang/Basic/Builtins.def
+++ b/clang/include/clang/Basic/Builtins.def
@@ -854,6 +854,19 @@ ATOMIC_BUILTIN(__opencl_atomic_fetch_max, "v.", "t")
 ATOMIC_BUILTIN(__atomic_fetch_min, "v.", "t")
 ATOMIC_BUILTIN(__atomic_fetch_max, "v.", "t")
 
+// HIP atomic builtins.
+ATOMIC_BUILTIN(__hip_atomic_load, "v.", "t")
+ATOMIC_BUILTIN(__hip_atomic_store, "v.", "t")
+ATOMIC_BUILTIN(__hip_atomic_compare_exchange_weak, "v.", "t")
+ATOMIC_BUILTIN(__hip_atomic_compare_exchange_strong, "v.", "t")
+ATOMIC_BUILTIN(__hip_atomic_exchange, "v.", "t")
+ATOMIC_BUILTIN(__hip_atomic_fetch_add, "v.", "t")
+ATOMIC_BUILTIN(__hip_atomic_fetch_and, "v.", "t")
+ATOMIC_BUILTIN(__hip_atomic_fetch_or, "v.", "t")
+ATOMIC_BUILTIN(__hip_atomic_fetch_xor, "v.", "t")
+ATOMIC_BUILTIN(__hip_atomic_fetch_min, "v.", "t")
+ATOMIC_BUILTIN(__hip_atomic_fetch_max, "v.", "t")
+
 #undef ATOMIC_BUILTIN
 
 // Non-overloaded atomic builtins.
diff --git a/clang/include/clang/Basic/BuiltinsPPC.def b/clang/include/clang/Basic/BuiltinsPPC.def
index cd6b2df10e52..70b0184f199f 100644
--- a/clang/include/clang/Basic/BuiltinsPPC.def
+++ b/clang/include/clang/Basic/BuiltinsPPC.def
@@ -404,6 +404,7 @@ BUILTIN(__builtin_altivec_vbpermd, "V2ULLiV2ULLiV16Uc", "")
 // P8 Crypto built-ins.
 BUILTIN(__builtin_altivec_crypto_vsbox, "V2ULLiV2ULLi", "")
 BUILTIN(__builtin_altivec_crypto_vpermxor, "V16UcV16UcV16UcV16Uc", "")
+BUILTIN(__builtin_altivec_crypto_vpermxor_be, "V16UcV16UcV16UcV16Uc", "")
 BUILTIN(__builtin_altivec_crypto_vshasigmaw, "V4UiV4UiIiIi", "")
 BUILTIN(__builtin_altivec_crypto_vshasigmad, "V2ULLiV2ULLiIiIi", "")
 BUILTIN(__builtin_altivec_crypto_vcipher, "V2ULLiV2ULLiV2ULLi", "")
@@ -424,6 +425,12 @@ BUILTIN(__builtin_altivec_vctzh, "V8UsV8Us", "")
 BUILTIN(__builtin_altivec_vctzw, "V4UiV4Ui", "")
 BUILTIN(__builtin_altivec_vctzd, "V2ULLiV2ULLi", "")
 
+// P8 BCD builtins.
+BUILTIN(__builtin_ppc_bcdadd, "V16UcV16UcV16UcIi", "")
+BUILTIN(__builtin_ppc_bcdsub, "V16UcV16UcV16UcIi", "")
+BUILTIN(__builtin_ppc_bcdadd_p, "iiV16UcV16Uc", "")
+BUILTIN(__builtin_ppc_bcdsub_p, "iiV16UcV16Uc", "")
+
 BUILTIN(__builtin_altivec_vclzlsbb, "SiV16Uc", "")
 BUILTIN(__builtin_altivec_vctzlsbb, "SiV16Uc", "")
 BUILTIN(__builtin_altivec_vprtybw, "V4UiV4Ui", "")
diff --git a/clang/include/clang/Basic/DiagnosticDriverKinds.td b/clang/include/clang/Basic/DiagnosticDriverKinds.td
index ff8c36910e13..2f50918b527b 100644
--- a/clang/include/clang/Basic/DiagnosticDriverKinds.td
+++ b/clang/include/clang/Basic/DiagnosticDriverKinds.td
@@ -189,6 +189,12 @@ def err_drv_invalid_mtp : Error<
   "invalid thread pointer reading mode '%0'">;
 def err_drv_missing_arg_mtp : Error<
   "missing argument to '%0'">;
+def warn_drv_missing_plugin_name : Warning<
+  "missing plugin name in %0">,
+  InGroup<InvalidCommandLineArgument>;
+def warn_drv_missing_plugin_arg : Warning<
+  "missing plugin argument for plugin %0 in %1">,
+  InGroup<InvalidCommandLineArgument>;
 def err_drv_invalid_libcxx_deployment : Error<
   "invalid deployment target for -stdlib=libc++ (requires %0 or later)">;
 def err_drv_invalid_argument_to_option : Error<
@@ -394,6 +400,8 @@ def warn_ignoring_verify_debuginfo_preserve_export : Warning<
   InGroup<UnusedCommandLineArgument>;
 def err_invalid_branch_protection: Error <
   "invalid branch protection option '%0' in '%1'">;
+def warn_unsupported_branch_protection: Warning <
+  "invalid branch protection option '%0' in '%1'">, InGroup<BranchProtection>;
 def err_invalid_sls_hardening : Error<
   "invalid sls hardening option '%0' in '%1'">;
 def err_sls_hardening_arm_not_supported : Error<
diff --git a/clang/include/clang/Basic/DiagnosticGroups.td b/clang/include/clang/Basic/DiagnosticGroups.td
index 85d373845c81..90df3a424406 100644
--- a/clang/include/clang/Basic/DiagnosticGroups.td
+++ b/clang/include/clang/Basic/DiagnosticGroups.td
@@ -56,7 +56,9 @@ def CoroutineMissingUnhandledException :
   DiagGroup<"coroutine-missing-unhandled-exception">;
 def DeprecatedExperimentalCoroutine :
   DiagGroup<"deprecated-experimental-coroutine">;
-def Coroutine : DiagGroup<"coroutine", [CoroutineMissingUnhandledException, DeprecatedExperimentalCoroutine]>;
+def DeprecatedCoroutine :
+  DiagGroup<"deprecated-coroutine", [DeprecatedExperimentalCoroutine]>;
+def Coroutine : DiagGroup<"coroutine", [CoroutineMissingUnhandledException, DeprecatedCoroutine]>;
 def ObjCBoolConstantConversion : DiagGroup<"objc-bool-constant-conversion">;
 def ConstantConversion : DiagGroup<"constant-conversion",
                                    [BitFieldConstantConversion,
@@ -1273,9 +1275,14 @@ def : DiagGroup<"spirv-compat", [SpirCompat]>; // Alias.
 // Warning for the GlobalISel options.
 def GlobalISel : DiagGroup<"global-isel">;
 
+// A warning group for the GNU extension to allow mixed specifier types for
+// target-clones multiversioning.
+def TargetClonesMixedSpecifiers : DiagGroup<"target-clones-mixed-specifiers">;
+
 // A warning group specifically for warnings related to function
 // multiversioning.
-def FunctionMultiVersioning : DiagGroup<"function-multiversion">;
+def FunctionMultiVersioning
+    : DiagGroup<"function-multiversion", [TargetClonesMixedSpecifiers]>;
 
 def NoDeref : DiagGroup<"noderef">;
 
@@ -1331,3 +1338,6 @@ def PedanticMacros : DiagGroup<"pedantic-macros",
                      BuiltinMacroRedefined,
                      RestrictExpansionMacro,
                      FinalMacro]>;
+
+def BranchProtection : DiagGroup<"branch-protection">;
+
diff --git a/clang/include/clang/Basic/DiagnosticIDs.h b/clang/include/clang/Basic/DiagnosticIDs.h
index aef86516707c..375930c14848 100644
--- a/clang/include/clang/Basic/DiagnosticIDs.h
+++ b/clang/include/clang/Basic/DiagnosticIDs.h
@@ -30,7 +30,7 @@ namespace clang {
     // Size of each of the diagnostic categories.
     enum {
       DIAG_SIZE_COMMON        =  300,
-      DIAG_SIZE_DRIVER        =  250,
+      DIAG_SIZE_DRIVER        =  300,
       DIAG_SIZE_FRONTEND      =  150,
       DIAG_SIZE_SERIALIZATION =  120,
       DIAG_SIZE_LEX           =  400,
diff --git a/clang/include/clang/Basic/DiagnosticParseKinds.td b/clang/include/clang/Basic/DiagnosticParseKinds.td
index 1bc2e8b0c7ef..92e877074ad3 100644
--- a/clang/include/clang/Basic/DiagnosticParseKinds.td
+++ b/clang/include/clang/Basic/DiagnosticParseKinds.td
@@ -1548,6 +1548,9 @@ def note_meant_to_use_typename : Note<
 let CategoryName = "Coroutines Issue" in {
 def err_for_co_await_not_range_for : Error<
   "'co_await' modifier can only be applied to range-based for loop">;
+def warn_deprecated_for_co_await : Warning<
+  "'for co_await' belongs to CoroutineTS instead of C++20, which is deprecated">,
+  InGroup<DeprecatedCoroutine>;
 }
 
 let CategoryName = "Concepts Issue" in {
diff --git a/clang/include/clang/Basic/DiagnosticSemaKinds.td b/clang/include/clang/Basic/DiagnosticSemaKinds.td
index dc67f86f25ca..fb5bd53f7432 100644
--- a/clang/include/clang/Basic/DiagnosticSemaKinds.td
+++ b/clang/include/clang/Basic/DiagnosticSemaKinds.td
@@ -2979,9 +2979,13 @@ def err_attribute_requires_opencl_version : Error<
   "attribute %0 is supported in the OpenCL version %1%select{| onwards}2">;
 def err_invalid_branch_protection_spec : Error<
   "invalid or misplaced branch protection specification '%0'">;
+def warn_unsupported_branch_protection_spec : Warning<
+  "unsupported branch protection specification '%0'">, InGroup<BranchProtection>;
+
 def warn_unsupported_target_attribute
     : Warning<"%select{unsupported|duplicate|unknown}0%select{| architecture|"
-              " tune CPU}1 '%2' in the 'target' attribute string; 'target' "
+              " tune CPU}1 '%2' in the '%select{target|target_clones}3' "
+              "attribute string; '%select{target|target_clones}3' "
               "attribute ignored">,
       InGroup<IgnoredAttributes>;
 def err_attribute_unsupported
@@ -9864,6 +9868,8 @@ def warn_duplicate_attribute_exact : Warning<
 def warn_duplicate_attribute : Warning<
   "attribute %0 is already applied with different arguments">,
   InGroup<IgnoredAttributes>;
+def err_disallowed_duplicate_attribute : Error<
+  "attribute %0 cannot appear more than once on a declaration">;
 
 def warn_sync_fetch_and_nand_semantics_change : Warning<
   "the semantics of this intrinsic changed with GCC "
@@ -11254,9 +11260,11 @@ def err_multiversion_duplicate : Error<
   "multiversioned function redeclarations require identical target attributes">;
 def err_multiversion_noproto : Error<
   "multiversioned function must have a prototype">;
-def err_multiversion_disallowed_other_attr : Error<
-  "attribute '%select{target|cpu_specific|cpu_dispatch}0' multiversioning cannot be combined"
-  " with attribute %1">;
+def err_multiversion_disallowed_other_attr
+    : Error<"attribute "
+            "'%select{|target|cpu_specific|cpu_dispatch|target_clones}0' "
+            "multiversioning cannot be combined"
+            " with attribute %1">;
 def err_multiversion_mismatched_attrs
     : Error<"attributes on multiversioned functions must all match, attribute "
             "%0 %select{is missing|has different arguments}1">;
@@ -11264,11 +11272,14 @@ def err_multiversion_diff : Error<
   "multiversioned function declaration has a different %select{calling convention"
   "|return type|constexpr specification|inline specification|linkage|"
   "language linkage}0">;
-def err_multiversion_doesnt_support : Error<
-  "attribute '%select{target|cpu_specific|cpu_dispatch}0' multiversioned functions do not "
-  "yet support %select{function templates|virtual functions|"
-  "deduced return types|constructors|destructors|deleted functions|"
-  "defaulted functions|constexpr functions|consteval function}1">;
+def err_multiversion_doesnt_support
+    : Error<"attribute "
+            "'%select{|target|cpu_specific|cpu_dispatch|target_clones}0' "
+            "multiversioned functions do not "
+            "yet support %select{function templates|virtual functions|"
+            "deduced return types|constructors|destructors|deleted functions|"
+            "defaulted functions|constexpr functions|consteval "
+            "function|lambdas}1">;
 def err_multiversion_not_allowed_on_main : Error<
   "'main' cannot be a multiversioned function">;
 def err_multiversion_not_supported : Error<
@@ -11285,6 +11296,19 @@ def warn_multiversion_duplicate_entries : Warning<
 def warn_dispatch_body_ignored : Warning<
   "body of cpu_dispatch function will be ignored">,
   InGroup<FunctionMultiVersioning>;
+def err_target_clone_must_have_default
+    : Error<"'target_clones' multiversioning requires a default target">;
+def err_target_clone_doesnt_match
+    : Error<"'target_clones' attribute does not match previous declaration">;
+def warn_target_clone_mixed_values
+    : ExtWarn<
+          "mixing 'target_clones' specifier mechanisms is permitted for GCC "
+          "compatibility; use a comma separated sequence of string literals, "
+          "or a string literal containing a comma-separated list of versions">,
+      InGroup<TargetClonesMixedSpecifiers>;
+def warn_target_clone_duplicate_options
+    : Warning<"version list contains duplicate entries">,
+      InGroup<FunctionMultiVersioning>;
 
 // three-way comparison operator diagnostics
 def err_implied_comparison_category_type_not_found : Error<
diff --git a/clang/include/clang/Basic/SyncScope.h b/clang/include/clang/Basic/SyncScope.h
index ce8fb9cbed13..34703310af2b 100644
--- a/clang/include/clang/Basic/SyncScope.h
+++ b/clang/include/clang/Basic/SyncScope.h
@@ -40,6 +40,11 @@ namespace clang {
 ///   Update getAsString.
 ///
 enum class SyncScope {
+  HIPSingleThread,
+  HIPWavefront,
+  HIPWorkgroup,
+  HIPAgent,
+  HIPSystem,
   OpenCLWorkGroup,
   OpenCLDevice,
   OpenCLAllSVMDevices,
@@ -49,6 +54,16 @@ enum class SyncScope {
 
 inline llvm::StringRef getAsString(SyncScope S) {
   switch (S) {
+  case SyncScope::HIPSingleThread:
+    return "hip_singlethread";
+  case SyncScope::HIPWavefront:
+    return "hip_wavefront";
+  case SyncScope::HIPWorkgroup:
+    return "hip_workgroup";
+  case SyncScope::HIPAgent:
+    return "hip_agent";
+  case SyncScope::HIPSystem:
+    return "hip_system";
   case SyncScope::OpenCLWorkGroup:
     return "opencl_workgroup";
   case SyncScope::OpenCLDevice:
@@ -62,7 +77,7 @@ inline llvm::StringRef getAsString(SyncScope S) {
 }
 
 /// Defines the kind of atomic scope models.
-enum class AtomicScopeModelKind { None, OpenCL };
+enum class AtomicScopeModelKind { None, OpenCL, HIP };
 
 /// Defines the interface for synch scope model.
 class AtomicScopeModel {
@@ -138,6 +153,58 @@ public:
   }
 };
 
+/// Defines the synch scope model for HIP.
+class AtomicScopeHIPModel : public AtomicScopeModel {
+public:
+  /// The enum values match the pre-defined macros
+  /// __HIP_MEMORY_SCOPE_*, which are used to define memory_scope_*
+  /// enums in hip-c.h.
+  enum ID {
+    SingleThread = 1,
+    Wavefront = 2,
+    Workgroup = 3,
+    Agent = 4,
+    System = 5,
+    Last = System
+  };
+
+  AtomicScopeHIPModel() {}
+
+  SyncScope map(unsigned S) const override {
+    switch (static_cast<ID>(S)) {
+    case SingleThread:
+      return SyncScope::HIPSingleThread;
+    case Wavefront:
+      return SyncScope::HIPWavefront;
+    case Workgroup:
+      return SyncScope::HIPWorkgroup;
+    case Agent:
+      return SyncScope::HIPAgent;
+    case System:
+      return SyncScope::HIPSystem;
+    }
+    llvm_unreachable("Invalid language synch scope value");
+  }
+
+  bool isValid(unsigned S) const override {
+    return S >= static_cast<unsigned>(SingleThread) &&
+           S <= static_cast<unsigned>(Last);
+  }
+
+  ArrayRef<unsigned> getRuntimeValues() const override {
+    static_assert(Last == System, "Does not include all synch scopes");
+    static const unsigned Scopes[] = {
+        static_cast<unsigned>(SingleThread), static_cast<unsigned>(Wavefront),
+        static_cast<unsigned>(Workgroup), static_cast<unsigned>(Agent),
+        static_cast<unsigned>(System)};
+    return llvm::makeArrayRef(Scopes);
+  }
+
+  unsigned getFallBackValue() const override {
+    return static_cast<unsigned>(System);
+  }
+};
+
 inline std::unique_ptr<AtomicScopeModel>
 AtomicScopeModel::create(AtomicScopeModelKind K) {
   switch (K) {
@@ -145,9 +212,11 @@ AtomicScopeModel::create(AtomicScopeModelKind K) {
     return std::unique_ptr<AtomicScopeModel>{};
   case AtomicScopeModelKind::OpenCL:
     return std::make_unique<AtomicScopeOpenCLModel>();
+  case AtomicScopeModelKind::HIP:
+    return std::make_unique<AtomicScopeHIPModel>();
   }
   llvm_unreachable("Invalid atomic scope model kind");
 }
-}
+} // namespace clang
 
 #endif
diff --git a/clang/include/clang/Driver/Options.td b/clang/include/clang/Driver/Options.td
index 9bde64cf49fd..4e6dd2050344 100644
--- a/clang/include/clang/Driver/Options.td
+++ b/clang/include/clang/Driver/Options.td
@@ -2504,6 +2504,9 @@ defm rwpi : BoolFOption<"rwpi",
   NegFlag<SetFalse>>;
 def fplugin_EQ : Joined<["-"], "fplugin=">, Group<f_Group>, Flags<[NoXarchOption]>, MetaVarName<"<dsopath>">,
   HelpText<"Load the named plugin (dynamic shared object)">;
+def fplugin_arg : Joined<["-"], "fplugin-arg-">,
+  MetaVarName<"<name>-<arg>">,
+  HelpText<"Pass <arg> to plugin <name>">;
 def fpass_plugin_EQ : Joined<["-"], "fpass-plugin=">,
   Group<f_Group>, Flags<[CC1Option]>, MetaVarName<"<dsopath>">,
   HelpText<"Load pass plugin from a dynamic shared object file (only with new pass manager).">,
@@ -2786,10 +2789,11 @@ def fvisibility_ms_compat : Flag<["-"], "fvisibility-ms-compat">, Group<f_Group>
 def fvisibility_global_new_delete_hidden : Flag<["-"], "fvisibility-global-new-delete-hidden">, Group<f_Group>,
   HelpText<"Give global C++ operator new and delete declarations hidden visibility">, Flags<[CC1Option]>,
   MarshallingInfoFlag<LangOpts<"GlobalAllocationFunctionVisibilityHidden">>;
-def fnew_infallible : Flag<["-"], "fnew-infallible">, Group<f_Group>,
-  HelpText<"Treats throwing global C++ operator new as always returning valid memory "
-  "(annotates with __attribute__((returns_nonnull)) and throw()). This is detectable in source.">,
-  Flags<[CC1Option]>, MarshallingInfoFlag<LangOpts<"NewInfallible">>;
+defm new_infallible : BoolFOption<"new-infallible",
+  LangOpts<"NewInfallible">, DefaultFalse,
+  PosFlag<SetTrue, [], "Enable">, NegFlag<SetFalse, [], "Disable">,
+  BothFlags<[CC1Option], " treating throwing global C++ operator new as always returning valid memory "
+  "(annotates with __attribute__((returns_nonnull)) and throw()). This is detectable in source.">>;
 defm whole_program_vtables : BoolFOption<"whole-program-vtables",
   CodeGenOpts<"WholeProgramVTables">, DefaultFalse,
   PosFlag<SetTrue, [CC1Option], "Enables whole-program vtable optimization. Requires -flto">,
@@ -4519,7 +4523,7 @@ def frecord_marker_EQ : Joined<["-"], "frecord-marker=">, Group<gfortran_Group>;
 defm aggressive_function_elimination : BooleanFFlag<"aggressive-function-elimination">, Group<gfortran_Group>;
 defm align_commons : BooleanFFlag<"align-commons">, Group<gfortran_Group>;
 defm all_intrinsics : BooleanFFlag<"all-intrinsics">, Group<gfortran_Group>;
-defm automatic : BooleanFFlag<"automatic">, Group<gfortran_Group>;
+def fautomatic : Flag<["-"], "fautomatic">; // -fno-automatic is significant
 defm backtrace : BooleanFFlag<"backtrace">, Group<gfortran_Group>;
 defm bounds_check : BooleanFFlag<"bounds-check">, Group<gfortran_Group>;
 defm check_array_temporaries : BooleanFFlag<"check-array-temporaries">, Group<gfortran_Group>;
@@ -4616,6 +4620,9 @@ defm backslash : OptInFC1FFlag<"backslash", "Specify that backslash in string in
 defm xor_operator : OptInFC1FFlag<"xor-operator", "Enable .XOR. as a synonym of .NEQV.">;
 defm logical_abbreviations : OptInFC1FFlag<"logical-abbreviations", "Enable logical abbreviations">;
 defm implicit_none : OptInFC1FFlag<"implicit-none", "No implicit typing allowed unless overridden by IMPLICIT statements">;
+
+def fno_automatic : Flag<["-"], "fno-automatic">, Group<f_Group>,
+  HelpText<"Implies the SAVE attribute for non-automatic local objects in subprograms unless RECURSIVE">;
 }
 
 def J : JoinedOrSeparate<["-"], "J">,
@@ -5059,9 +5066,10 @@ def msmall_data_limit : Separate<["-"], "msmall-data-limit">,
 def funwind_tables_EQ : Joined<["-"], "funwind-tables=">,
   HelpText<"Generate unwinding tables for all functions">,
   MarshallingInfoInt<CodeGenOpts<"UnwindTables">>;
-def mconstructor_aliases : Flag<["-"], "mconstructor-aliases">,
-  HelpText<"Emit complete constructors and destructors as aliases when possible">,
-  MarshallingInfoFlag<CodeGenOpts<"CXXCtorDtorAliases">>;
+defm constructor_aliases : BoolOption<"m", "constructor-aliases",
+  CodeGenOpts<"CXXCtorDtorAliases">, DefaultFalse,
+  PosFlag<SetTrue, [], "Enable">, NegFlag<SetFalse, [], "Disable">,
+  BothFlags<[CC1Option], " emitting complete constructors and destructors as aliases when possible">>;
 def mlink_bitcode_file : Separate<["-"], "mlink-bitcode-file">,
   HelpText<"Link the given bitcode file before performing optimizations.">;
 def mlink_builtin_bitcode : Separate<["-"], "mlink-builtin-bitcode">,
@@ -5174,10 +5182,6 @@ defm debug_pass_manager : BoolOption<"f", "debug-pass-manager",
   CodeGenOpts<"DebugPassManager">, DefaultFalse,
   PosFlag<SetTrue, [], "Prints debug information for the new pass manager">,
   NegFlag<SetFalse, [], "Disables debug printing for the new pass manager">>;
-def fexperimental_debug_variable_locations : Flag<["-"],
-    "fexperimental-debug-variable-locations">,
-    HelpText<"Use experimental new value-tracking variable locations">,
-    MarshallingInfoFlag<CodeGenOpts<"ValueTrackingVariableLocations">>;
 def fverify_debuginfo_preserve
     : Flag<["-"], "fverify-debuginfo-preserve">,
       HelpText<"Enable Debug Info Metadata preservation testing in "
diff --git a/clang/include/clang/Frontend/PrecompiledPreamble.h b/clang/include/clang/Frontend/PrecompiledPreamble.h
index bb7fd97fe5df..dacbffef0b12 100644
--- a/clang/include/clang/Frontend/PrecompiledPreamble.h
+++ b/clang/include/clang/Frontend/PrecompiledPreamble.h
@@ -274,7 +274,7 @@ class PreambleCallbacks {
 public:
   virtual ~PreambleCallbacks() = default;
 
-  /// Called before FrontendAction::BeginSourceFile.
+  /// Called before FrontendAction::Execute.
   /// Can be used to store references to various CompilerInstance fields
   /// (e.g. SourceManager) that may be interesting to the consumers of other
   /// callbacks.
@@ -291,7 +291,7 @@ public:
   /// used instead, but having only this method allows a simpler API.
   virtual void HandleTopLevelDecl(DeclGroupRef DG);
   /// Creates wrapper class for PPCallbacks so we can also process information
-  /// about includes that are inside of a preamble
+  /// about includes that are inside of a preamble. Called after BeforeExecute.
   virtual std::unique_ptr<PPCallbacks> createPPCallbacks();
   /// The returned CommentHandler will be added to the preprocessor if not null.
   virtual CommentHandler *getCommentHandler();
diff --git a/clang/include/clang/Sema/Sema.h b/clang/include/clang/Sema/Sema.h
index 43ce5d983217..1a82a9498d1d 100644
--- a/clang/include/clang/Sema/Sema.h
+++ b/clang/include/clang/Sema/Sema.h
@@ -1296,6 +1296,11 @@ public:
       EK_Decltype, EK_TemplateArgument, EK_Other
     } ExprContext;
 
+    // A context can be nested in both a discarded statement context and
+    // an immediate function context, so they need to be tracked independently.
+    bool InDiscardedStatement;
+    bool InImmediateFunctionContext;
+
     ExpressionEvaluationContextRecord(ExpressionEvaluationContext Context,
                                       unsigned NumCleanupObjects,
                                       CleanupInfo ParentCleanup,
@@ -1303,7 +1308,8 @@ public:
                                       ExpressionKind ExprContext)
         : Context(Context), ParentCleanup(ParentCleanup),
           NumCleanupObjects(NumCleanupObjects), NumTypos(0),
-          ManglingContextDecl(ManglingContextDecl), ExprContext(ExprContext) {}
+          ManglingContextDecl(ManglingContextDecl), ExprContext(ExprContext),
+          InDiscardedStatement(false), InImmediateFunctionContext(false) {}
 
     bool isUnevaluated() const {
       return Context == ExpressionEvaluationContext::Unevaluated ||
@@ -1317,7 +1323,13 @@ public:
     }
 
     bool isImmediateFunctionContext() const {
-      return Context == ExpressionEvaluationContext::ImmediateFunctionContext;
+      return Context == ExpressionEvaluationContext::ImmediateFunctionContext ||
+             InImmediateFunctionContext;
+    }
+
+    bool isDiscardedStatementContext() const {
+      return Context == ExpressionEvaluationContext::DiscardedStatement ||
+             InDiscardedStatement;
     }
   };
 
@@ -4351,6 +4363,10 @@ public:
   llvm::Error isValidSectionSpecifier(StringRef Str);
   bool checkSectionName(SourceLocation LiteralLoc, StringRef Str);
   bool checkTargetAttr(SourceLocation LiteralLoc, StringRef Str);
+  bool checkTargetClonesAttrString(SourceLocation LiteralLoc, StringRef Str,
+                                   const StringLiteral *Literal,
+                                   bool &HasDefault, bool &HasCommas,
+                                   SmallVectorImpl<StringRef> &Strings);
   bool checkMSInheritanceAttrOnDefinition(
       CXXRecordDecl *RD, SourceRange Range, bool BestCase,
       MSInheritanceModel SemanticSpelling);
@@ -9150,14 +9166,7 @@ public:
   bool isImmediateFunctionContext() const {
     assert(!ExprEvalContexts.empty() &&
            "Must be in an expression evaluation context");
-    for (const ExpressionEvaluationContextRecord &context :
-         llvm::reverse(ExprEvalContexts)) {
-      if (context.isImmediateFunctionContext())
-        return true;
-      if (context.isUnevaluated())
-        return false;
-    }
-    return false;
+    return ExprEvalContexts.back().isImmediateFunctionContext();
   }
 
   /// RAII class used to determine whether SFINAE has
diff --git a/clang/include/clang/Tooling/DependencyScanning/DependencyScanningFilesystem.h b/clang/include/clang/Tooling/DependencyScanning/DependencyScanningFilesystem.h
index c52da3305f7c..af02fa2e7e87 100644
--- a/clang/include/clang/Tooling/DependencyScanning/DependencyScanningFilesystem.h
+++ b/clang/include/clang/Tooling/DependencyScanning/DependencyScanningFilesystem.h
@@ -195,15 +195,23 @@ public:
   llvm::ErrorOr<std::unique_ptr<llvm::vfs::File>>
   openFileForRead(const Twine &Path) override;
 
-  void clearIgnoredFiles() { IgnoredFiles.clear(); }
-  void ignoreFile(StringRef Filename);
+  /// Disable minimization of the given file.
+  void disableMinimization(StringRef Filename);
+  /// Enable minimization of all files.
+  void enableMinimizationOfAllFiles() { NotToBeMinimized.clear(); }
 
 private:
-  bool shouldIgnoreFile(StringRef Filename);
+  /// Check whether the file should be minimized.
+  bool shouldMinimize(StringRef Filename);
 
   llvm::ErrorOr<const CachedFileSystemEntry *>
   getOrCreateFileSystemEntry(const StringRef Filename);
 
+  /// Create a cached file system entry based on the initial status result.
+  CachedFileSystemEntry
+  createFileSystemEntry(llvm::ErrorOr<llvm::vfs::Status> &&MaybeStatus,
+                        StringRef Filename, bool ShouldMinimize);
+
   /// The global cache shared between worker threads.
   DependencyScanningFilesystemSharedCache &SharedCache;
   /// The local cache is used by the worker thread to cache file system queries
@@ -214,7 +222,7 @@ private:
   /// currently active preprocessor.
   ExcludedPreprocessorDirectiveSkipMapping *PPSkipMappings;
   /// The set of files that should not be minimized.
-  llvm::StringSet<> IgnoredFiles;
+  llvm::StringSet<> NotToBeMinimized;
 };
 
 } // end namespace dependencies
diff --git a/clang/lib/AST/ASTContext.cpp b/clang/lib/AST/ASTContext.cpp
index 294cc20f76c5..2d85d72e5b8a 100644
--- a/clang/lib/AST/ASTContext.cpp
+++ b/clang/lib/AST/ASTContext.cpp
@@ -11800,6 +11800,15 @@ void ASTContext::getFunctionFeatureMap(llvm::StringMap<bool> &FeatureMap,
                     Target->getTargetOpts().FeaturesAsWritten.begin(),
                     Target->getTargetOpts().FeaturesAsWritten.end());
     Target->initFeatureMap(FeatureMap, getDiagnostics(), TargetCPU, Features);
+  } else if (const auto *TC = FD->getAttr<TargetClonesAttr>()) {
+    std::vector<std::string> Features;
+    StringRef VersionStr = TC->getFeatureStr(GD.getMultiVersionIndex());
+    if (VersionStr.startswith("arch="))
+      TargetCPU = VersionStr.drop_front(sizeof("arch=") - 1);
+    else if (VersionStr != "default")
+      Features.push_back((StringRef{"+"} + VersionStr).str());
+
+    Target->initFeatureMap(FeatureMap, getDiagnostics(), TargetCPU, Features);
   } else {
     FeatureMap = Target->getTargetOpts().FeatureMap;
   }
diff --git a/clang/lib/AST/ASTStructuralEquivalence.cpp b/clang/lib/AST/ASTStructuralEquivalence.cpp
index e85feb779190..7fd24e2aa9ad 100644
--- a/clang/lib/AST/ASTStructuralEquivalence.cpp
+++ b/clang/lib/AST/ASTStructuralEquivalence.cpp
@@ -1347,6 +1347,42 @@ IsStructurallyEquivalentLambdas(StructuralEquivalenceContext &Context,
   return true;
 }
 
+/// Determine if context of a class is equivalent.
+static bool IsRecordContextStructurallyEquivalent(RecordDecl *D1,
+                                                  RecordDecl *D2) {
+  // The context should be completely equal, including anonymous and inline
+  // namespaces.
+  // We compare objects as part of full translation units, not subtrees of
+  // translation units.
+  DeclContext *DC1 = D1->getDeclContext()->getNonTransparentContext();
+  DeclContext *DC2 = D2->getDeclContext()->getNonTransparentContext();
+  while (true) {
+    // Special case: We allow a struct defined in a function to be equivalent
+    // with a similar struct defined outside of a function.
+    if ((DC1->isFunctionOrMethod() && DC2->isTranslationUnit()) ||
+        (DC2->isFunctionOrMethod() && DC1->isTranslationUnit()))
+      return true;
+
+    if (DC1->getDeclKind() != DC2->getDeclKind())
+      return false;
+    if (DC1->isTranslationUnit())
+      break;
+    if (DC1->isInlineNamespace() != DC2->isInlineNamespace())
+      return false;
+    if (const auto *ND1 = dyn_cast<NamedDecl>(DC1)) {
+      const auto *ND2 = cast<NamedDecl>(DC2);
+      if (!DC1->isInlineNamespace() &&
+          !IsStructurallyEquivalent(ND1->getIdentifier(), ND2->getIdentifier()))
+        return false;
+    }
+
+    DC1 = DC1->getParent()->getNonTransparentContext();
+    DC2 = DC2->getParent()->getNonTransparentContext();
+  }
+
+  return true;
+}
+
 /// Determine structural equivalence of two records.
 static bool IsStructurallyEquivalent(StructuralEquivalenceContext &Context,
                                      RecordDecl *D1, RecordDecl *D2) {
@@ -1386,6 +1422,12 @@ static bool IsStructurallyEquivalent(StructuralEquivalenceContext &Context,
     }
   }
 
+  // If the records occur in different context (namespace), these should be
+  // different. This is specially important if the definition of one or both
+  // records is missing.
+  if (!IsRecordContextStructurallyEquivalent(D1, D2))
+    return false;
+
   // If both declarations are class template specializations, we know
   // the ODR applies, so check the template and template arguments.
   const auto *Spec1 = dyn_cast<ClassTemplateSpecializationDecl>(D1);
diff --git a/clang/lib/AST/Decl.cpp b/clang/lib/AST/Decl.cpp
index 5ea091edcf4c..68dfef248f65 100644
--- a/clang/lib/AST/Decl.cpp
+++ b/clang/lib/AST/Decl.cpp
@@ -3271,6 +3271,8 @@ MultiVersionKind FunctionDecl::getMultiVersionKind() const {
     return MultiVersionKind::CPUDispatch;
   if (hasAttr<CPUSpecificAttr>())
     return MultiVersionKind::CPUSpecific;
+  if (hasAttr<TargetClonesAttr>())
+    return MultiVersionKind::TargetClones;
   return MultiVersionKind::None;
 }
 
@@ -3286,6 +3288,10 @@ bool FunctionDecl::isTargetMultiVersion() const {
   return isMultiVersion() && hasAttr<TargetAttr>();
 }
 
+bool FunctionDecl::isTargetClonesMultiVersion() const {
+  return isMultiVersion() && hasAttr<TargetClonesAttr>();
+}
+
 void
 FunctionDecl::setPreviousDeclaration(FunctionDecl *PrevDecl) {
   redeclarable_base::setPreviousDecl(PrevDecl);
diff --git a/clang/lib/AST/DeclBase.cpp b/clang/lib/AST/DeclBase.cpp
index 4044404f74ef..064012ba865c 100644
--- a/clang/lib/AST/DeclBase.cpp
+++ b/clang/lib/AST/DeclBase.cpp
@@ -964,7 +964,7 @@ SourceLocation Decl::getBodyRBrace() const {
   return {};
 }
 
-bool Decl::AccessDeclContextSanity() const {
+bool Decl::AccessDeclContextCheck() const {
 #ifndef NDEBUG
   // Suppress this check if any of the following hold:
   // 1. this is the translation unit (and thus has no parent)
@@ -1212,7 +1212,7 @@ bool DeclContext::Encloses(const DeclContext *DC) const {
     return getPrimaryContext()->Encloses(DC);
 
   for (; DC; DC = DC->getParent())
-    if (DC->getPrimaryContext() == this)
+    if (!isa<LinkageSpecDecl>(DC) && DC->getPrimaryContext() == this)
       return true;
   return false;
 }
diff --git a/clang/lib/AST/Expr.cpp b/clang/lib/AST/Expr.cpp
index 7bd3dce43f4d..d3cb2ff3734c 100644
--- a/clang/lib/AST/Expr.cpp
+++ b/clang/lib/AST/Expr.cpp
@@ -4681,6 +4681,7 @@ unsigned AtomicExpr::getNumSubExprs(AtomicOp Op) {
     return 2;
 
   case AO__opencl_atomic_load:
+  case AO__hip_atomic_load:
   case AO__c11_atomic_store:
   case AO__c11_atomic_exchange:
   case AO__atomic_load:
@@ -4713,7 +4714,15 @@ unsigned AtomicExpr::getNumSubExprs(AtomicOp Op) {
   case AO__atomic_fetch_max:
     return 3;
 
+  case AO__hip_atomic_exchange:
+  case AO__hip_atomic_fetch_add:
+  case AO__hip_atomic_fetch_and:
+  case AO__hip_atomic_fetch_or:
+  case AO__hip_atomic_fetch_xor:
+  case AO__hip_atomic_fetch_min:
+  case AO__hip_atomic_fetch_max:
   case AO__opencl_atomic_store:
+  case AO__hip_atomic_store:
   case AO__opencl_atomic_exchange:
   case AO__opencl_atomic_fetch_add:
   case AO__opencl_atomic_fetch_sub:
@@ -4728,9 +4737,10 @@ unsigned AtomicExpr::getNumSubExprs(AtomicOp Op) {
   case AO__c11_atomic_compare_exchange_strong:
   case AO__c11_atomic_compare_exchange_weak:
     return 5;
-
+  case AO__hip_atomic_compare_exchange_strong:
   case AO__opencl_atomic_compare_exchange_strong:
   case AO__opencl_atomic_compare_exchange_weak:
+  case AO__hip_atomic_compare_exchange_weak:
   case AO__atomic_compare_exchange:
   case AO__atomic_compare_exchange_n:
     return 6;
diff --git a/clang/lib/AST/ExprConstant.cpp b/clang/lib/AST/ExprConstant.cpp
index fe96db9ca918..99babd58b027 100644
--- a/clang/lib/AST/ExprConstant.cpp
+++ b/clang/lib/AST/ExprConstant.cpp
@@ -7483,7 +7483,7 @@ public:
     const Expr *Source = E->getSourceExpr();
     if (!Source)
       return Error(E);
-    if (Source == E) { // sanity checking.
+    if (Source == E) {
       assert(0 && "OpaqueValueExpr recursively refers to itself");
       return Error(E);
     }
diff --git a/clang/lib/AST/MicrosoftMangle.cpp b/clang/lib/AST/MicrosoftMangle.cpp
index 163d4e95386e..79a448a2435c 100644
--- a/clang/lib/AST/MicrosoftMangle.cpp
+++ b/clang/lib/AST/MicrosoftMangle.cpp
@@ -21,6 +21,7 @@
 #include "clang/AST/DeclTemplate.h"
 #include "clang/AST/Expr.h"
 #include "clang/AST/ExprCXX.h"
+#include "clang/AST/GlobalDecl.h"
 #include "clang/AST/Mangle.h"
 #include "clang/AST/VTableBuilder.h"
 #include "clang/Basic/ABI.h"
@@ -39,6 +40,18 @@ using namespace clang;
 
 namespace {
 
+// Get GlobalDecl of DeclContext of local entities.
+static GlobalDecl getGlobalDeclAsDeclContext(const DeclContext *DC) {
+  GlobalDecl GD;
+  if (auto *CD = dyn_cast<CXXConstructorDecl>(DC))
+    GD = GlobalDecl(CD, Ctor_Complete);
+  else if (auto *DD = dyn_cast<CXXDestructorDecl>(DC))
+    GD = GlobalDecl(DD, Dtor_Complete);
+  else
+    GD = GlobalDecl(cast<FunctionDecl>(DC));
+  return GD;
+}
+
 struct msvc_hashing_ostream : public llvm::raw_svector_ostream {
   raw_ostream &OS;
   llvm::SmallString<64> Buffer;
@@ -345,9 +358,9 @@ public:
 
   raw_ostream &getStream() const { return Out; }
 
-  void mangle(const NamedDecl *D, StringRef Prefix = "?");
-  void mangleName(const NamedDecl *ND);
-  void mangleFunctionEncoding(const FunctionDecl *FD, bool ShouldMangle);
+  void mangle(GlobalDecl GD, StringRef Prefix = "?");
+  void mangleName(GlobalDecl GD);
+  void mangleFunctionEncoding(GlobalDecl GD, bool ShouldMangle);
   void mangleVariableEncoding(const VarDecl *VD);
   void mangleMemberDataPointer(const CXXRecordDecl *RD, const ValueDecl *VD,
                                StringRef Prefix = "$");
@@ -370,7 +383,7 @@ public:
                           const FunctionDecl *D = nullptr,
                           bool ForceThisQuals = false,
                           bool MangleExceptionSpec = true);
-  void mangleNestedName(const NamedDecl *ND);
+  void mangleNestedName(GlobalDecl GD);
 
 private:
   bool isStructorDecl(const NamedDecl *ND) const {
@@ -384,10 +397,10 @@ private:
                                   AddrSpace == LangAS::ptr32_uptr));
   }
 
-  void mangleUnqualifiedName(const NamedDecl *ND) {
-    mangleUnqualifiedName(ND, ND->getDeclName());
+  void mangleUnqualifiedName(GlobalDecl GD) {
+    mangleUnqualifiedName(GD, cast<NamedDecl>(GD.getDecl())->getDeclName());
   }
-  void mangleUnqualifiedName(const NamedDecl *ND, DeclarationName Name);
+  void mangleUnqualifiedName(GlobalDecl GD, DeclarationName Name);
   void mangleSourceName(StringRef Name);
   void mangleOperatorName(OverloadedOperatorKind OO, SourceLocation Loc);
   void mangleCXXDtorType(CXXDtorType T);
@@ -396,9 +409,9 @@ private:
   void manglePointerCVQualifiers(Qualifiers Quals);
   void manglePointerExtQualifiers(Qualifiers Quals, QualType PointeeType);
 
-  void mangleUnscopedTemplateName(const TemplateDecl *ND);
+  void mangleUnscopedTemplateName(GlobalDecl GD);
   void
-  mangleTemplateInstantiationName(const TemplateDecl *TD,
+  mangleTemplateInstantiationName(GlobalDecl GD,
                                   const TemplateArgumentList &TemplateArgs);
   void mangleObjCMethodName(const ObjCMethodDecl *MD);
 
@@ -533,7 +546,8 @@ MicrosoftMangleContextImpl::shouldMangleStringLiteral(const StringLiteral *SL) {
   return true;
 }
 
-void MicrosoftCXXNameMangler::mangle(const NamedDecl *D, StringRef Prefix) {
+void MicrosoftCXXNameMangler::mangle(GlobalDecl GD, StringRef Prefix) {
+  const NamedDecl *D = cast<NamedDecl>(GD.getDecl());
   // MSVC doesn't mangle C++ names the same way it mangles extern "C" names.
   // Therefore it's really important that we don't decorate the
   // name with leading underscores or leading/trailing at signs. So, by
@@ -542,9 +556,9 @@ void MicrosoftCXXNameMangler::mangle(const NamedDecl *D, StringRef Prefix) {
 
   // <mangled-name> ::= ? <name> <type-encoding>
   Out << Prefix;
-  mangleName(D);
+  mangleName(GD);
   if (const FunctionDecl *FD = dyn_cast<FunctionDecl>(D))
-    mangleFunctionEncoding(FD, Context.shouldMangleDeclName(FD));
+    mangleFunctionEncoding(GD, Context.shouldMangleDeclName(FD));
   else if (const VarDecl *VD = dyn_cast<VarDecl>(D))
     mangleVariableEncoding(VD);
   else if (isa<MSGuidDecl>(D))
@@ -558,8 +572,9 @@ void MicrosoftCXXNameMangler::mangle(const NamedDecl *D, StringRef Prefix) {
     llvm_unreachable("Tried to mangle unexpected NamedDecl!");
 }
 
-void MicrosoftCXXNameMangler::mangleFunctionEncoding(const FunctionDecl *FD,
+void MicrosoftCXXNameMangler::mangleFunctionEncoding(GlobalDecl GD,
                                                      bool ShouldMangle) {
+  const FunctionDecl *FD = cast<FunctionDecl>(GD.getDecl());
   // <type-encoding> ::= <function-class> <function-type>
 
   // Since MSVC operates on the type as written and not the canonical type, it
@@ -770,13 +785,13 @@ void MicrosoftCXXNameMangler::mangleVirtualMemPtrThunk(
   mangleCallingConvention(MD->getType()->castAs<FunctionProtoType>());
 }
 
-void MicrosoftCXXNameMangler::mangleName(const NamedDecl *ND) {
+void MicrosoftCXXNameMangler::mangleName(GlobalDecl GD) {
   // <name> ::= <unscoped-name> {[<named-scope>]+ | [<nested-name>]}? @
 
   // Always start with the unqualified name.
-  mangleUnqualifiedName(ND);
+  mangleUnqualifiedName(GD);
 
-  mangleNestedName(ND);
+  mangleNestedName(GD);
 
   // Terminate the whole name with an '@'.
   Out << '@';
@@ -844,13 +859,14 @@ void MicrosoftCXXNameMangler::mangleBits(llvm::APInt Value) {
   }
 }
 
-static const TemplateDecl *
-isTemplate(const NamedDecl *ND, const TemplateArgumentList *&TemplateArgs) {
+static GlobalDecl isTemplate(GlobalDecl GD,
+                             const TemplateArgumentList *&TemplateArgs) {
+  const NamedDecl *ND = cast<NamedDecl>(GD.getDecl());
   // Check if we have a function template.
   if (const FunctionDecl *FD = dyn_cast<FunctionDecl>(ND)) {
     if (const TemplateDecl *TD = FD->getPrimaryTemplate()) {
       TemplateArgs = FD->getTemplateSpecializationArgs();
-      return TD;
+      return GD.getWithDecl(TD);
     }
   }
 
@@ -858,21 +874,22 @@ isTemplate(const NamedDecl *ND, const TemplateArgumentList *&TemplateArgs) {
   if (const ClassTemplateSpecializationDecl *Spec =
           dyn_cast<ClassTemplateSpecializationDecl>(ND)) {
     TemplateArgs = &Spec->getTemplateArgs();
-    return Spec->getSpecializedTemplate();
+    return GD.getWithDecl(Spec->getSpecializedTemplate());
   }
 
   // Check if we have a variable template.
   if (const VarTemplateSpecializationDecl *Spec =
           dyn_cast<VarTemplateSpecializationDecl>(ND)) {
     TemplateArgs = &Spec->getTemplateArgs();
-    return Spec->getSpecializedTemplate();
+    return GD.getWithDecl(Spec->getSpecializedTemplate());
   }
 
-  return nullptr;
+  return GlobalDecl();
 }
 
-void MicrosoftCXXNameMangler::mangleUnqualifiedName(const NamedDecl *ND,
+void MicrosoftCXXNameMangler::mangleUnqualifiedName(GlobalDecl GD,
                                                     DeclarationName Name) {
+  const NamedDecl *ND = cast<NamedDecl>(GD.getDecl());
   //  <unqualified-name> ::= <operator-name>
   //                     ::= <ctor-dtor-name>
   //                     ::= <source-name>
@@ -880,11 +897,11 @@ void MicrosoftCXXNameMangler::mangleUnqualifiedName(const NamedDecl *ND,
 
   // Check if we have a template.
   const TemplateArgumentList *TemplateArgs = nullptr;
-  if (const TemplateDecl *TD = isTemplate(ND, TemplateArgs)) {
+  if (GlobalDecl TD = isTemplate(GD, TemplateArgs)) {
     // Function templates aren't considered for name back referencing.  This
     // makes sense since function templates aren't likely to occur multiple
     // times in a symbol.
-    if (isa<FunctionTemplateDecl>(TD)) {
+    if (isa<FunctionTemplateDecl>(TD.getDecl())) {
       mangleTemplateInstantiationName(TD, *TemplateArgs);
       Out << '@';
       return;
@@ -945,7 +962,19 @@ void MicrosoftCXXNameMangler::mangleUnqualifiedName(const NamedDecl *ND,
   switch (Name.getNameKind()) {
     case DeclarationName::Identifier: {
       if (const IdentifierInfo *II = Name.getAsIdentifierInfo()) {
-        mangleSourceName(II->getName());
+        bool IsDeviceStub =
+            ND &&
+            ((isa<FunctionDecl>(ND) && ND->hasAttr<CUDAGlobalAttr>()) ||
+             (isa<FunctionTemplateDecl>(ND) &&
+              cast<FunctionTemplateDecl>(ND)
+                  ->getTemplatedDecl()
+                  ->hasAttr<CUDAGlobalAttr>())) &&
+            GD.getKernelReferenceKind() == KernelReferenceKind::Stub;
+        if (IsDeviceStub)
+          mangleSourceName(
+              (llvm::Twine("__device_stub__") + II->getName()).str());
+        else
+          mangleSourceName(II->getName());
         break;
       }
 
@@ -1146,7 +1175,8 @@ void MicrosoftCXXNameMangler::mangleUnqualifiedName(const NamedDecl *ND,
 
 // <postfix> ::= <unqualified-name> [<postfix>]
 //           ::= <substitution> [<postfix>]
-void MicrosoftCXXNameMangler::mangleNestedName(const NamedDecl *ND) {
+void MicrosoftCXXNameMangler::mangleNestedName(GlobalDecl GD) {
+  const NamedDecl *ND = cast<NamedDecl>(GD.getDecl());
   const DeclContext *DC = getEffectiveDeclContext(ND);
   while (!DC->isTranslationUnit()) {
     if (isa<TagDecl>(ND) || isa<VarDecl>(ND)) {
@@ -1229,7 +1259,7 @@ void MicrosoftCXXNameMangler::mangleNestedName(const NamedDecl *ND) {
     } else if (isa<NamedDecl>(DC)) {
       ND = cast<NamedDecl>(DC);
       if (const FunctionDecl *FD = dyn_cast<FunctionDecl>(ND)) {
-        mangle(FD, "?");
+        mangle(getGlobalDeclAsDeclContext(FD), "?");
         break;
       } else {
         mangleUnqualifiedName(ND);
@@ -1418,7 +1448,7 @@ void MicrosoftCXXNameMangler::mangleObjCMethodName(const ObjCMethodDecl *MD) {
 }
 
 void MicrosoftCXXNameMangler::mangleTemplateInstantiationName(
-    const TemplateDecl *TD, const TemplateArgumentList &TemplateArgs) {
+    GlobalDecl GD, const TemplateArgumentList &TemplateArgs) {
   // <template-name> ::= <unscoped-template-name> <template-args>
   //                 ::= <substitution>
   // Always start with the unqualified name.
@@ -1433,8 +1463,8 @@ void MicrosoftCXXNameMangler::mangleTemplateInstantiationName(
   TemplateArgBackReferences.swap(OuterTemplateArgsContext);
   PassObjectSizeArgs.swap(OuterPassObjectSizeArgs);
 
-  mangleUnscopedTemplateName(TD);
-  mangleTemplateArgs(TD, TemplateArgs);
+  mangleUnscopedTemplateName(GD);
+  mangleTemplateArgs(cast<TemplateDecl>(GD.getDecl()), TemplateArgs);
 
   // Restore the previous back reference contexts.
   NameBackReferences.swap(OuterTemplateContext);
@@ -1443,11 +1473,10 @@ void MicrosoftCXXNameMangler::mangleTemplateInstantiationName(
   PassObjectSizeArgs.swap(OuterPassObjectSizeArgs);
 }
 
-void
-MicrosoftCXXNameMangler::mangleUnscopedTemplateName(const TemplateDecl *TD) {
+void MicrosoftCXXNameMangler::mangleUnscopedTemplateName(GlobalDecl GD) {
   // <unscoped-template-name> ::= ?$ <unqualified-name>
   Out << "?$";
-  mangleUnqualifiedName(TD);
+  mangleUnqualifiedName(GD);
 }
 
 void MicrosoftCXXNameMangler::mangleIntegerLiteral(
@@ -3323,17 +3352,17 @@ void MicrosoftMangleContextImpl::mangleCXXName(GlobalDecl GD,
   if (auto *CD = dyn_cast<CXXConstructorDecl>(D)) {
     auto Type = GD.getCtorType();
     MicrosoftCXXNameMangler mangler(*this, MHO, CD, Type);
-    return mangler.mangle(D);
+    return mangler.mangle(GD);
   }
 
   if (auto *DD = dyn_cast<CXXDestructorDecl>(D)) {
     auto Type = GD.getDtorType();
     MicrosoftCXXNameMangler mangler(*this, MHO, DD, Type);
-    return mangler.mangle(D);
+    return mangler.mangle(GD);
   }
 
   MicrosoftCXXNameMangler Mangler(*this, MHO);
-  return Mangler.mangle(D);
+  return Mangler.mangle(GD);
 }
 
 void MicrosoftCXXNameMangler::mangleType(const ExtIntType *T, Qualifiers,
diff --git a/clang/lib/AST/StmtPrinter.cpp b/clang/lib/AST/StmtPrinter.cpp
index fc267d7006a1..b65a38d1e566 100644
--- a/clang/lib/AST/StmtPrinter.cpp
+++ b/clang/lib/AST/StmtPrinter.cpp
@@ -1691,7 +1691,8 @@ void StmtPrinter::VisitAtomicExpr(AtomicExpr *Node) {
   PrintExpr(Node->getPtr());
   if (Node->getOp() != AtomicExpr::AO__c11_atomic_load &&
       Node->getOp() != AtomicExpr::AO__atomic_load_n &&
-      Node->getOp() != AtomicExpr::AO__opencl_atomic_load) {
+      Node->getOp() != AtomicExpr::AO__opencl_atomic_load &&
+      Node->getOp() != AtomicExpr::AO__hip_atomic_load) {
     OS << ", ";
     PrintExpr(Node->getVal1());
   }
diff --git a/clang/lib/Analysis/FlowSensitive/TypeErasedDataflowAnalysis.cpp b/clang/lib/Analysis/FlowSensitive/TypeErasedDataflowAnalysis.cpp
new file mode 100644
index 000000000000..bb7eb9971068
--- /dev/null
+++ b/clang/lib/Analysis/FlowSensitive/TypeErasedDataflowAnalysis.cpp
@@ -0,0 +1,35 @@
+//===- TypeErasedDataflowAnalysis.cpp -------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+//  This file defines type-erased base types and functions for building dataflow
+//  analyses that run over Control-Flow Graphs (CFGs).
+//
+//===----------------------------------------------------------------------===//
+
+#include <vector>
+
+#include "clang/Analysis/CFG.h"
+#include "clang/Analysis/FlowSensitive/DataflowEnvironment.h"
+#include "clang/Analysis/FlowSensitive/TypeErasedDataflowAnalysis.h"
+#include "llvm/ADT/Optional.h"
+
+using namespace clang;
+using namespace dataflow;
+
+std::vector<llvm::Optional<TypeErasedDataflowAnalysisState>>
+runTypeErasedDataflowAnalysis(const CFG &Cfg,
+                              TypeErasedDataflowAnalysis &Analysis,
+                              const Environment &InitEnv) {
+  // FIXME: Consider enforcing that `Cfg` meets the requirements that
+  // are specified in the header. This could be done by remembering
+  // what options were used to build `Cfg` and asserting on them here.
+
+  // FIXME: Implement work list-based algorithm to compute the fixed
+  // point of `Analysis::transform` for every basic block in `Cfg`.
+  return {};
+}
diff --git a/clang/lib/Basic/Targets/AArch64.cpp b/clang/lib/Basic/Targets/AArch64.cpp
index f75b8ffcb53d..4d403ae1809d 100644
--- a/clang/lib/Basic/Targets/AArch64.cpp
+++ b/clang/lib/Basic/Targets/AArch64.cpp
@@ -140,8 +140,8 @@ bool AArch64TargetInfo::setABI(const std::string &Name) {
 bool AArch64TargetInfo::validateBranchProtection(StringRef Spec,
                                                  BranchProtectionInfo &BPI,
                                                  StringRef &Err) const {
-  llvm::AArch64::ParsedBranchProtection PBP;
-  if (!llvm::AArch64::parseBranchProtection(Spec, PBP, Err))
+  llvm::ARM::ParsedBranchProtection PBP;
+  if (!llvm::ARM::parseBranchProtection(Spec, PBP, Err))
     return false;
 
   BPI.SignReturnAddr =
diff --git a/clang/lib/Basic/Targets/ARM.cpp b/clang/lib/Basic/Targets/ARM.cpp
index fc6b01c87fd2..f330780300f2 100644
--- a/clang/lib/Basic/Targets/ARM.cpp
+++ b/clang/lib/Basic/Targets/ARM.cpp
@@ -367,6 +367,28 @@ bool ARMTargetInfo::setABI(const std::string &Name) {
   return false;
 }
 
+bool ARMTargetInfo::validateBranchProtection(StringRef Spec,
+                                             BranchProtectionInfo &BPI,
+                                             StringRef &Err) const {
+  llvm::ARM::ParsedBranchProtection PBP;
+  if (!llvm::ARM::parseBranchProtection(Spec, PBP, Err))
+    return false;
+
+  BPI.SignReturnAddr =
+      llvm::StringSwitch<LangOptions::SignReturnAddressScopeKind>(PBP.Scope)
+          .Case("non-leaf", LangOptions::SignReturnAddressScopeKind::NonLeaf)
+          .Case("all", LangOptions::SignReturnAddressScopeKind::All)
+          .Default(LangOptions::SignReturnAddressScopeKind::None);
+
+  // Don't care for the sign key, beyond issuing a warning.
+  if (PBP.Key == "b_key")
+    Err = "b-key";
+  BPI.SignKey = LangOptions::SignReturnAddressKeyKind::AKey;
+
+  BPI.BranchTargetEnforcement = PBP.BranchTargetEnforcement;
+  return true;
+}
+
 // FIXME: This should be based on Arch attributes, not CPU names.
 bool ARMTargetInfo::initFeatureMap(
     llvm::StringMap<bool> &Features, DiagnosticsEngine &Diags, StringRef CPU,
@@ -874,6 +896,16 @@ void ARMTargetInfo::getTargetDefines(const LangOptions &Opts,
     Builder.defineMacro("__ARM_BF16_FORMAT_ALTERNATIVE", "1");
   }
 
+  if (Opts.BranchTargetEnforcement)
+    Builder.defineMacro("__ARM_FEATURE_BTI_DEFAULT", "1");
+
+  if (Opts.hasSignReturnAddress()) {
+    unsigned Value = Opts.isSignReturnAddressWithAKey() ? 1 : 2;
+    if (Opts.isSignReturnAddressScopeAll())
+      Value |= 1 << 2;
+    Builder.defineMacro("__ARM_FEATURE_PAC_DEFAULT", Twine(Value));
+  }
+
   switch (ArchKind) {
   default:
     break;
diff --git a/clang/lib/Basic/Targets/ARM.h b/clang/lib/Basic/Targets/ARM.h
index d54a049042d6..7d0011d134ea 100644
--- a/clang/lib/Basic/Targets/ARM.h
+++ b/clang/lib/Basic/Targets/ARM.h
@@ -123,6 +123,9 @@ public:
   StringRef getABI() const override;
   bool setABI(const std::string &Name) override;
 
+  bool validateBranchProtection(StringRef, BranchProtectionInfo &,
+                                StringRef &) const override;
+
   // FIXME: This should be based on Arch attributes, not CPU names.
   bool
   initFeatureMap(llvm::StringMap<bool> &Features, DiagnosticsEngine &Diags,
diff --git a/clang/lib/Basic/Targets/OSTargets.cpp b/clang/lib/Basic/Targets/OSTargets.cpp
index 7cd4a5190120..53748bf067cd 100644
--- a/clang/lib/Basic/Targets/OSTargets.cpp
+++ b/clang/lib/Basic/Targets/OSTargets.cpp
@@ -181,8 +181,10 @@ static void addVisualCDefines(const LangOptions &Opts, MacroBuilder &Builder) {
       Builder.defineMacro("_HAS_CHAR16_T_LANGUAGE_SUPPORT", Twine(1));
 
     if (Opts.isCompatibleWithMSVC(LangOptions::MSVC2015)) {
-      if (Opts.CPlusPlus20)
-        Builder.defineMacro("_MSVC_LANG", "201705L");
+      if (Opts.CPlusPlus2b)
+        Builder.defineMacro("_MSVC_LANG", "202004L");
+      else if (Opts.CPlusPlus20)
+        Builder.defineMacro("_MSVC_LANG", "202002L");
       else if (Opts.CPlusPlus17)
         Builder.defineMacro("_MSVC_LANG", "201703L");
       else if (Opts.CPlusPlus14)
@@ -201,6 +203,14 @@ static void addVisualCDefines(const LangOptions &Opts, MacroBuilder &Builder) {
   }
 
   Builder.defineMacro("_INTEGRAL_MAX_BITS", "64");
+
+  // Starting with VS 2022 17.1, MSVC predefines the below macro to inform
+  // users of the execution character set defined at compile time.
+  // The value given is the Windows Code Page Identifier:
+  // https://docs.microsoft.com/en-us/windows/win32/intl/code-page-identifiers
+  //
+  // Clang currently only supports UTF-8, so we'll use 65001
+  Builder.defineMacro("_MSVC_EXECUTION_CHARACTER_SET", "65001");
 }
 
 void addWindowsDefines(const llvm::Triple &Triple, const LangOptions &Opts,
diff --git a/clang/lib/Basic/Targets/SPIR.h b/clang/lib/Basic/Targets/SPIR.h
index 704b1843dfed..8cf18b6c20f1 100644
--- a/clang/lib/Basic/Targets/SPIR.h
+++ b/clang/lib/Basic/Targets/SPIR.h
@@ -56,9 +56,14 @@ static const unsigned SPIRDefIsGenMap[] = {
     0, // opencl_generic
     0, // opencl_global_device
     0, // opencl_global_host
-    0, // cuda_device
-    0, // cuda_constant
-    0, // cuda_shared
+    // cuda_* address space mapping is intended for HIPSPV (HIP to SPIR-V
+    // translation). This mapping is enabled when the language mode is HIP.
+    1, // cuda_device
+    // cuda_constant pointer can be casted to default/"flat" pointer, but in
+    // SPIR-V casts between constant and generic pointers are not allowed. For
+    // this reason cuda_constant is mapped to SPIR-V CrossWorkgroup.
+    1, // cuda_constant
+    3, // cuda_shared
     1, // sycl_global
     5, // sycl_global_device
     6, // sycl_global_host
@@ -74,6 +79,8 @@ class LLVM_LIBRARY_VISIBILITY BaseSPIRTargetInfo : public TargetInfo {
 protected:
   BaseSPIRTargetInfo(const llvm::Triple &Triple, const TargetOptions &)
       : TargetInfo(Triple) {
+    assert((Triple.isSPIR() || Triple.isSPIRV()) &&
+           "Invalid architecture for SPIR or SPIR-V.");
     assert(getTriple().getOS() == llvm::Triple::UnknownOS &&
            "SPIR(-V) target must use unknown OS");
     assert(getTriple().getEnvironment() == llvm::Triple::UnknownEnvironment &&
@@ -137,11 +144,16 @@ public:
     // FIXME: SYCL specification considers unannotated pointers and references
     // to be pointing to the generic address space. See section 5.9.3 of
     // SYCL 2020 specification.
-    // Currently, there is no way of representing SYCL's default address space
-    // language semantic along with the semantics of embedded C's default
-    // address space in the same address space map. Hence the map needs to be
-    // reset to allow mapping to the desired value of 'Default' entry for SYCL.
-    setAddressSpaceMap(/*DefaultIsGeneric=*/Opts.SYCLIsDevice);
+    // Currently, there is no way of representing SYCL's and HIP's default
+    // address space language semantic along with the semantics of embedded C's
+    // default address space in the same address space map. Hence the map needs
+    // to be reset to allow mapping to the desired value of 'Default' entry for
+    // SYCL and HIP.
+    setAddressSpaceMap(
+        /*DefaultIsGeneric=*/Opts.SYCLIsDevice ||
+        // The address mapping from HIP language for device code is only defined
+        // for SPIR-V.
+        (getTriple().isSPIRV() && Opts.HIP && Opts.CUDAIsDevice));
   }
 
   void setSupportedOpenCLOpts() override {
@@ -159,6 +171,7 @@ class LLVM_LIBRARY_VISIBILITY SPIRTargetInfo : public BaseSPIRTargetInfo {
 public:
   SPIRTargetInfo(const llvm::Triple &Triple, const TargetOptions &Opts)
       : BaseSPIRTargetInfo(Triple, Opts) {
+    assert(Triple.isSPIR() && "Invalid architecture for SPIR.");
     assert(getTriple().getOS() == llvm::Triple::UnknownOS &&
            "SPIR target must use unknown OS");
     assert(getTriple().getEnvironment() == llvm::Triple::UnknownEnvironment &&
@@ -177,6 +190,8 @@ class LLVM_LIBRARY_VISIBILITY SPIR32TargetInfo : public SPIRTargetInfo {
 public:
   SPIR32TargetInfo(const llvm::Triple &Triple, const TargetOptions &Opts)
       : SPIRTargetInfo(Triple, Opts) {
+    assert(Triple.getArch() == llvm::Triple::spir &&
+           "Invalid architecture for 32-bit SPIR.");
     PointerWidth = PointerAlign = 32;
     SizeType = TargetInfo::UnsignedInt;
     PtrDiffType = IntPtrType = TargetInfo::SignedInt;
@@ -192,6 +207,8 @@ class LLVM_LIBRARY_VISIBILITY SPIR64TargetInfo : public SPIRTargetInfo {
 public:
   SPIR64TargetInfo(const llvm::Triple &Triple, const TargetOptions &Opts)
       : SPIRTargetInfo(Triple, Opts) {
+    assert(Triple.getArch() == llvm::Triple::spir64 &&
+           "Invalid architecture for 64-bit SPIR.");
     PointerWidth = PointerAlign = 64;
     SizeType = TargetInfo::UnsignedLong;
     PtrDiffType = IntPtrType = TargetInfo::SignedLong;
@@ -207,6 +224,7 @@ class LLVM_LIBRARY_VISIBILITY SPIRVTargetInfo : public BaseSPIRTargetInfo {
 public:
   SPIRVTargetInfo(const llvm::Triple &Triple, const TargetOptions &Opts)
       : BaseSPIRTargetInfo(Triple, Opts) {
+    assert(Triple.isSPIRV() && "Invalid architecture for SPIR-V.");
     assert(getTriple().getOS() == llvm::Triple::UnknownOS &&
            "SPIR-V target must use unknown OS");
     assert(getTriple().getEnvironment() == llvm::Triple::UnknownEnvironment &&
@@ -225,6 +243,8 @@ class LLVM_LIBRARY_VISIBILITY SPIRV32TargetInfo : public SPIRVTargetInfo {
 public:
   SPIRV32TargetInfo(const llvm::Triple &Triple, const TargetOptions &Opts)
       : SPIRVTargetInfo(Triple, Opts) {
+    assert(Triple.getArch() == llvm::Triple::spirv32 &&
+           "Invalid architecture for 32-bit SPIR-V.");
     PointerWidth = PointerAlign = 32;
     SizeType = TargetInfo::UnsignedInt;
     PtrDiffType = IntPtrType = TargetInfo::SignedInt;
@@ -240,6 +260,8 @@ class LLVM_LIBRARY_VISIBILITY SPIRV64TargetInfo : public SPIRVTargetInfo {
 public:
   SPIRV64TargetInfo(const llvm::Triple &Triple, const TargetOptions &Opts)
       : SPIRVTargetInfo(Triple, Opts) {
+    assert(Triple.getArch() == llvm::Triple::spirv64 &&
+           "Invalid architecture for 64-bit SPIR-V.");
     PointerWidth = PointerAlign = 64;
     SizeType = TargetInfo::UnsignedLong;
     PtrDiffType = IntPtrType = TargetInfo::SignedLong;
diff --git a/clang/lib/Basic/Targets/X86.cpp b/clang/lib/Basic/Targets/X86.cpp
index 5e3686893719..5c4bd364b06a 100644
--- a/clang/lib/Basic/Targets/X86.cpp
+++ b/clang/lib/Basic/Targets/X86.cpp
@@ -239,9 +239,9 @@ bool X86TargetInfo::handleTargetFeatures(std::vector<std::string> &Features,
       HasAVX512ER = true;
     } else if (Feature == "+avx512fp16") {
       HasAVX512FP16 = true;
+      HasFloat16 = true;
     } else if (Feature == "+avx512pf") {
       HasAVX512PF = true;
-      HasLegalHalfType = true;
     } else if (Feature == "+avx512dq") {
       HasAVX512DQ = true;
     } else if (Feature == "+avx512bitalg") {
@@ -369,8 +369,6 @@ bool X86TargetInfo::handleTargetFeatures(std::vector<std::string> &Features,
                          .Default(NoXOP);
     XOPLevel = std::max(XOPLevel, XLevel);
   }
-  // Turn on _float16 for x86 (feature sse2)
-  HasFloat16 = SSELevel >= SSE2;
 
   // LLVM doesn't have a separate switch for fpmath, so only accept it if it
   // matches the selected sse level.
@@ -384,12 +382,10 @@ bool X86TargetInfo::handleTargetFeatures(std::vector<std::string> &Features,
   SimdDefaultAlign =
       hasFeature("avx512f") ? 512 : hasFeature("avx") ? 256 : 128;
 
-  if (!HasX87) {
-    if (LongDoubleFormat == &llvm::APFloat::x87DoubleExtended())
-      HasLongDouble = false;
-    if (getTriple().getArch() == llvm::Triple::x86)
-      HasFPReturn = false;
-  }
+  // FIXME: We should allow long double type on 32-bits to match with GCC.
+  // This requires backend to be able to lower f80 without x87 first.
+  if (!HasX87 && LongDoubleFormat == &llvm::APFloat::x87DoubleExtended())
+    HasLongDouble = false;
 
   return true;
 }
diff --git a/clang/lib/CodeGen/BackendUtil.cpp b/clang/lib/CodeGen/BackendUtil.cpp
index 648c7b3df8ed..510f3911939c 100644
--- a/clang/lib/CodeGen/BackendUtil.cpp
+++ b/clang/lib/CodeGen/BackendUtil.cpp
@@ -1034,8 +1034,9 @@ void EmitAssemblyHelper::EmitAssemblyWithLegacyPassManager(
         if (!ThinLinkOS)
           return;
       }
-      TheModule->addModuleFlag(Module::Error, "EnableSplitLTOUnit",
-                               CodeGenOpts.EnableSplitLTOUnit);
+      if (!TheModule->getModuleFlag("EnableSplitLTOUnit"))
+        TheModule->addModuleFlag(Module::Error, "EnableSplitLTOUnit",
+                                 CodeGenOpts.EnableSplitLTOUnit);
       PerModulePasses.add(createWriteThinLTOBitcodePass(
           *OS, ThinLinkOS ? &ThinLinkOS->os() : nullptr));
     } else {
@@ -1049,8 +1050,9 @@ void EmitAssemblyHelper::EmitAssemblyWithLegacyPassManager(
       if (EmitLTOSummary) {
         if (!TheModule->getModuleFlag("ThinLTO"))
           TheModule->addModuleFlag(Module::Error, "ThinLTO", uint32_t(0));
-        TheModule->addModuleFlag(Module::Error, "EnableSplitLTOUnit",
-                                 uint32_t(1));
+        if (!TheModule->getModuleFlag("EnableSplitLTOUnit"))
+          TheModule->addModuleFlag(Module::Error, "EnableSplitLTOUnit",
+                                   uint32_t(1));
       }
 
       PerModulePasses.add(createBitcodeWriterPass(
@@ -1451,8 +1453,9 @@ void EmitAssemblyHelper::RunOptimizationPipeline(
         if (!ThinLinkOS)
           return;
       }
-      TheModule->addModuleFlag(Module::Error, "EnableSplitLTOUnit",
-                               CodeGenOpts.EnableSplitLTOUnit);
+      if (!TheModule->getModuleFlag("EnableSplitLTOUnit"))
+        TheModule->addModuleFlag(Module::Error, "EnableSplitLTOUnit",
+                                 CodeGenOpts.EnableSplitLTOUnit);
       MPM.addPass(ThinLTOBitcodeWriterPass(*OS, ThinLinkOS ? &ThinLinkOS->os()
                                                            : nullptr));
     } else {
@@ -1465,8 +1468,9 @@ void EmitAssemblyHelper::RunOptimizationPipeline(
       if (EmitLTOSummary) {
         if (!TheModule->getModuleFlag("ThinLTO"))
           TheModule->addModuleFlag(Module::Error, "ThinLTO", uint32_t(0));
-        TheModule->addModuleFlag(Module::Error, "EnableSplitLTOUnit",
-                                 uint32_t(1));
+        if (!TheModule->getModuleFlag("EnableSplitLTOUnit"))
+          TheModule->addModuleFlag(Module::Error, "EnableSplitLTOUnit",
+                                   uint32_t(1));
       }
       MPM.addPass(
           BitcodeWriterPass(*OS, CodeGenOpts.EmitLLVMUseLists, EmitLTOSummary));
diff --git a/clang/lib/CodeGen/CGAtomic.cpp b/clang/lib/CodeGen/CGAtomic.cpp
index 326ca8d50533..b68e6328acdf 100644
--- a/clang/lib/CodeGen/CGAtomic.cpp
+++ b/clang/lib/CodeGen/CGAtomic.cpp
@@ -524,12 +524,14 @@ static void EmitAtomicOp(CodeGenFunction &CGF, AtomicExpr *E, Address Dest,
     llvm_unreachable("Already handled!");
 
   case AtomicExpr::AO__c11_atomic_compare_exchange_strong:
+  case AtomicExpr::AO__hip_atomic_compare_exchange_strong:
   case AtomicExpr::AO__opencl_atomic_compare_exchange_strong:
     emitAtomicCmpXchgFailureSet(CGF, E, false, Dest, Ptr, Val1, Val2,
                                 FailureOrder, Size, Order, Scope);
     return;
   case AtomicExpr::AO__c11_atomic_compare_exchange_weak:
   case AtomicExpr::AO__opencl_atomic_compare_exchange_weak:
+  case AtomicExpr::AO__hip_atomic_compare_exchange_weak:
     emitAtomicCmpXchgFailureSet(CGF, E, true, Dest, Ptr, Val1, Val2,
                                 FailureOrder, Size, Order, Scope);
     return;
@@ -565,6 +567,7 @@ static void EmitAtomicOp(CodeGenFunction &CGF, AtomicExpr *E, Address Dest,
   }
   case AtomicExpr::AO__c11_atomic_load:
   case AtomicExpr::AO__opencl_atomic_load:
+  case AtomicExpr::AO__hip_atomic_load:
   case AtomicExpr::AO__atomic_load_n:
   case AtomicExpr::AO__atomic_load: {
     llvm::LoadInst *Load = CGF.Builder.CreateLoad(Ptr);
@@ -576,6 +579,7 @@ static void EmitAtomicOp(CodeGenFunction &CGF, AtomicExpr *E, Address Dest,
 
   case AtomicExpr::AO__c11_atomic_store:
   case AtomicExpr::AO__opencl_atomic_store:
+  case AtomicExpr::AO__hip_atomic_store:
   case AtomicExpr::AO__atomic_store:
   case AtomicExpr::AO__atomic_store_n: {
     llvm::Value *LoadVal1 = CGF.Builder.CreateLoad(Val1);
@@ -586,6 +590,7 @@ static void EmitAtomicOp(CodeGenFunction &CGF, AtomicExpr *E, Address Dest,
   }
 
   case AtomicExpr::AO__c11_atomic_exchange:
+  case AtomicExpr::AO__hip_atomic_exchange:
   case AtomicExpr::AO__opencl_atomic_exchange:
   case AtomicExpr::AO__atomic_exchange_n:
   case AtomicExpr::AO__atomic_exchange:
@@ -597,6 +602,7 @@ static void EmitAtomicOp(CodeGenFunction &CGF, AtomicExpr *E, Address Dest,
                                                  : llvm::Instruction::Add;
     LLVM_FALLTHROUGH;
   case AtomicExpr::AO__c11_atomic_fetch_add:
+  case AtomicExpr::AO__hip_atomic_fetch_add:
   case AtomicExpr::AO__opencl_atomic_fetch_add:
   case AtomicExpr::AO__atomic_fetch_add:
     Op = E->getValueType()->isFloatingType() ? llvm::AtomicRMWInst::FAdd
@@ -618,6 +624,7 @@ static void EmitAtomicOp(CodeGenFunction &CGF, AtomicExpr *E, Address Dest,
     PostOpMinMax = true;
     LLVM_FALLTHROUGH;
   case AtomicExpr::AO__c11_atomic_fetch_min:
+  case AtomicExpr::AO__hip_atomic_fetch_min:
   case AtomicExpr::AO__opencl_atomic_fetch_min:
   case AtomicExpr::AO__atomic_fetch_min:
     Op = E->getValueType()->isSignedIntegerType() ? llvm::AtomicRMWInst::Min
@@ -628,6 +635,7 @@ static void EmitAtomicOp(CodeGenFunction &CGF, AtomicExpr *E, Address Dest,
     PostOpMinMax = true;
     LLVM_FALLTHROUGH;
   case AtomicExpr::AO__c11_atomic_fetch_max:
+  case AtomicExpr::AO__hip_atomic_fetch_max:
   case AtomicExpr::AO__opencl_atomic_fetch_max:
   case AtomicExpr::AO__atomic_fetch_max:
     Op = E->getValueType()->isSignedIntegerType() ? llvm::AtomicRMWInst::Max
@@ -638,6 +646,7 @@ static void EmitAtomicOp(CodeGenFunction &CGF, AtomicExpr *E, Address Dest,
     PostOp = llvm::Instruction::And;
     LLVM_FALLTHROUGH;
   case AtomicExpr::AO__c11_atomic_fetch_and:
+  case AtomicExpr::AO__hip_atomic_fetch_and:
   case AtomicExpr::AO__opencl_atomic_fetch_and:
   case AtomicExpr::AO__atomic_fetch_and:
     Op = llvm::AtomicRMWInst::And;
@@ -647,6 +656,7 @@ static void EmitAtomicOp(CodeGenFunction &CGF, AtomicExpr *E, Address Dest,
     PostOp = llvm::Instruction::Or;
     LLVM_FALLTHROUGH;
   case AtomicExpr::AO__c11_atomic_fetch_or:
+  case AtomicExpr::AO__hip_atomic_fetch_or:
   case AtomicExpr::AO__opencl_atomic_fetch_or:
   case AtomicExpr::AO__atomic_fetch_or:
     Op = llvm::AtomicRMWInst::Or;
@@ -656,6 +666,7 @@ static void EmitAtomicOp(CodeGenFunction &CGF, AtomicExpr *E, Address Dest,
     PostOp = llvm::Instruction::Xor;
     LLVM_FALLTHROUGH;
   case AtomicExpr::AO__c11_atomic_fetch_xor:
+  case AtomicExpr::AO__hip_atomic_fetch_xor:
   case AtomicExpr::AO__opencl_atomic_fetch_xor:
   case AtomicExpr::AO__atomic_fetch_xor:
     Op = llvm::AtomicRMWInst::Xor;
@@ -838,6 +849,7 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
 
   case AtomicExpr::AO__c11_atomic_load:
   case AtomicExpr::AO__opencl_atomic_load:
+  case AtomicExpr::AO__hip_atomic_load:
   case AtomicExpr::AO__atomic_load_n:
     break;
 
@@ -857,7 +869,9 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
   case AtomicExpr::AO__c11_atomic_compare_exchange_strong:
   case AtomicExpr::AO__c11_atomic_compare_exchange_weak:
   case AtomicExpr::AO__opencl_atomic_compare_exchange_strong:
+  case AtomicExpr::AO__hip_atomic_compare_exchange_strong:
   case AtomicExpr::AO__opencl_atomic_compare_exchange_weak:
+  case AtomicExpr::AO__hip_atomic_compare_exchange_weak:
   case AtomicExpr::AO__atomic_compare_exchange_n:
   case AtomicExpr::AO__atomic_compare_exchange:
     Val1 = EmitPointerWithAlignment(E->getVal1());
@@ -873,6 +887,7 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
 
   case AtomicExpr::AO__c11_atomic_fetch_add:
   case AtomicExpr::AO__c11_atomic_fetch_sub:
+  case AtomicExpr::AO__hip_atomic_fetch_add:
   case AtomicExpr::AO__opencl_atomic_fetch_add:
   case AtomicExpr::AO__opencl_atomic_fetch_sub:
     if (MemTy->isPointerType()) {
@@ -901,7 +916,9 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
   case AtomicExpr::AO__c11_atomic_store:
   case AtomicExpr::AO__c11_atomic_exchange:
   case AtomicExpr::AO__opencl_atomic_store:
+  case AtomicExpr::AO__hip_atomic_store:
   case AtomicExpr::AO__opencl_atomic_exchange:
+  case AtomicExpr::AO__hip_atomic_exchange:
   case AtomicExpr::AO__atomic_store_n:
   case AtomicExpr::AO__atomic_exchange_n:
   case AtomicExpr::AO__c11_atomic_fetch_and:
@@ -916,8 +933,11 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
   case AtomicExpr::AO__opencl_atomic_fetch_min:
   case AtomicExpr::AO__opencl_atomic_fetch_max:
   case AtomicExpr::AO__atomic_fetch_and:
+  case AtomicExpr::AO__hip_atomic_fetch_and:
   case AtomicExpr::AO__atomic_fetch_or:
+  case AtomicExpr::AO__hip_atomic_fetch_or:
   case AtomicExpr::AO__atomic_fetch_xor:
+  case AtomicExpr::AO__hip_atomic_fetch_xor:
   case AtomicExpr::AO__atomic_fetch_nand:
   case AtomicExpr::AO__atomic_and_fetch:
   case AtomicExpr::AO__atomic_or_fetch:
@@ -926,7 +946,9 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
   case AtomicExpr::AO__atomic_max_fetch:
   case AtomicExpr::AO__atomic_min_fetch:
   case AtomicExpr::AO__atomic_fetch_max:
+  case AtomicExpr::AO__hip_atomic_fetch_max:
   case AtomicExpr::AO__atomic_fetch_min:
+  case AtomicExpr::AO__hip_atomic_fetch_min:
     Val1 = EmitValToTemp(*this, E->getVal1());
     break;
   }
@@ -968,11 +990,14 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
     case AtomicExpr::AO__c11_atomic_fetch_add:
     case AtomicExpr::AO__opencl_atomic_fetch_add:
     case AtomicExpr::AO__atomic_fetch_add:
+    case AtomicExpr::AO__hip_atomic_fetch_add:
     case AtomicExpr::AO__c11_atomic_fetch_and:
     case AtomicExpr::AO__opencl_atomic_fetch_and:
+    case AtomicExpr::AO__hip_atomic_fetch_and:
     case AtomicExpr::AO__atomic_fetch_and:
     case AtomicExpr::AO__c11_atomic_fetch_or:
     case AtomicExpr::AO__opencl_atomic_fetch_or:
+    case AtomicExpr::AO__hip_atomic_fetch_or:
     case AtomicExpr::AO__atomic_fetch_or:
     case AtomicExpr::AO__c11_atomic_fetch_nand:
     case AtomicExpr::AO__atomic_fetch_nand:
@@ -984,6 +1009,7 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
     case AtomicExpr::AO__opencl_atomic_fetch_min:
     case AtomicExpr::AO__opencl_atomic_fetch_max:
     case AtomicExpr::AO__atomic_fetch_xor:
+    case AtomicExpr::AO__hip_atomic_fetch_xor:
     case AtomicExpr::AO__c11_atomic_fetch_max:
     case AtomicExpr::AO__c11_atomic_fetch_min:
     case AtomicExpr::AO__atomic_add_fetch:
@@ -993,7 +1019,9 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
     case AtomicExpr::AO__atomic_sub_fetch:
     case AtomicExpr::AO__atomic_xor_fetch:
     case AtomicExpr::AO__atomic_fetch_max:
+    case AtomicExpr::AO__hip_atomic_fetch_max:
     case AtomicExpr::AO__atomic_fetch_min:
+    case AtomicExpr::AO__hip_atomic_fetch_min:
     case AtomicExpr::AO__atomic_max_fetch:
     case AtomicExpr::AO__atomic_min_fetch:
       // For these, only library calls for certain sizes exist.
@@ -1014,10 +1042,15 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
     case AtomicExpr::AO__c11_atomic_exchange:
     case AtomicExpr::AO__c11_atomic_compare_exchange_weak:
     case AtomicExpr::AO__c11_atomic_compare_exchange_strong:
+    case AtomicExpr::AO__hip_atomic_compare_exchange_strong:
     case AtomicExpr::AO__opencl_atomic_load:
+    case AtomicExpr::AO__hip_atomic_load:
     case AtomicExpr::AO__opencl_atomic_store:
+    case AtomicExpr::AO__hip_atomic_store:
     case AtomicExpr::AO__opencl_atomic_exchange:
+    case AtomicExpr::AO__hip_atomic_exchange:
     case AtomicExpr::AO__opencl_atomic_compare_exchange_weak:
+    case AtomicExpr::AO__hip_atomic_compare_exchange_weak:
     case AtomicExpr::AO__opencl_atomic_compare_exchange_strong:
     case AtomicExpr::AO__atomic_load_n:
     case AtomicExpr::AO__atomic_store_n:
@@ -1079,7 +1112,9 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
     case AtomicExpr::AO__c11_atomic_compare_exchange_weak:
     case AtomicExpr::AO__c11_atomic_compare_exchange_strong:
     case AtomicExpr::AO__opencl_atomic_compare_exchange_weak:
+    case AtomicExpr::AO__hip_atomic_compare_exchange_weak:
     case AtomicExpr::AO__opencl_atomic_compare_exchange_strong:
+    case AtomicExpr::AO__hip_atomic_compare_exchange_strong:
     case AtomicExpr::AO__atomic_compare_exchange:
     case AtomicExpr::AO__atomic_compare_exchange_n:
       LibCallName = "__atomic_compare_exchange";
@@ -1101,6 +1136,7 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
     case AtomicExpr::AO__opencl_atomic_exchange:
     case AtomicExpr::AO__atomic_exchange_n:
     case AtomicExpr::AO__atomic_exchange:
+    case AtomicExpr::AO__hip_atomic_exchange:
       LibCallName = "__atomic_exchange";
       AddDirectArgument(*this, Args, UseOptimizedLibcall, Val1.getPointer(),
                         MemTy, E->getExprLoc(), TInfo.Width);
@@ -1109,6 +1145,7 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
     // void __atomic_store_N(T *mem, T val, int order)
     case AtomicExpr::AO__c11_atomic_store:
     case AtomicExpr::AO__opencl_atomic_store:
+    case AtomicExpr::AO__hip_atomic_store:
     case AtomicExpr::AO__atomic_store:
     case AtomicExpr::AO__atomic_store_n:
       LibCallName = "__atomic_store";
@@ -1121,6 +1158,7 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
     // T __atomic_load_N(T *mem, int order)
     case AtomicExpr::AO__c11_atomic_load:
     case AtomicExpr::AO__opencl_atomic_load:
+    case AtomicExpr::AO__hip_atomic_load:
     case AtomicExpr::AO__atomic_load:
     case AtomicExpr::AO__atomic_load_n:
       LibCallName = "__atomic_load";
@@ -1133,6 +1171,7 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
     case AtomicExpr::AO__c11_atomic_fetch_add:
     case AtomicExpr::AO__opencl_atomic_fetch_add:
     case AtomicExpr::AO__atomic_fetch_add:
+    case AtomicExpr::AO__hip_atomic_fetch_add:
       LibCallName = "__atomic_fetch_add";
       AddDirectArgument(*this, Args, UseOptimizedLibcall, Val1.getPointer(),
                         LoweredMemTy, E->getExprLoc(), TInfo.Width);
@@ -1144,6 +1183,7 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
       LLVM_FALLTHROUGH;
     case AtomicExpr::AO__c11_atomic_fetch_and:
     case AtomicExpr::AO__opencl_atomic_fetch_and:
+    case AtomicExpr::AO__hip_atomic_fetch_and:
     case AtomicExpr::AO__atomic_fetch_and:
       LibCallName = "__atomic_fetch_and";
       AddDirectArgument(*this, Args, UseOptimizedLibcall, Val1.getPointer(),
@@ -1156,6 +1196,7 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
       LLVM_FALLTHROUGH;
     case AtomicExpr::AO__c11_atomic_fetch_or:
     case AtomicExpr::AO__opencl_atomic_fetch_or:
+    case AtomicExpr::AO__hip_atomic_fetch_or:
     case AtomicExpr::AO__atomic_fetch_or:
       LibCallName = "__atomic_fetch_or";
       AddDirectArgument(*this, Args, UseOptimizedLibcall, Val1.getPointer(),
@@ -1180,6 +1221,7 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
       LLVM_FALLTHROUGH;
     case AtomicExpr::AO__c11_atomic_fetch_xor:
     case AtomicExpr::AO__opencl_atomic_fetch_xor:
+    case AtomicExpr::AO__hip_atomic_fetch_xor:
     case AtomicExpr::AO__atomic_fetch_xor:
       LibCallName = "__atomic_fetch_xor";
       AddDirectArgument(*this, Args, UseOptimizedLibcall, Val1.getPointer(),
@@ -1190,6 +1232,7 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
       LLVM_FALLTHROUGH;
     case AtomicExpr::AO__c11_atomic_fetch_min:
     case AtomicExpr::AO__atomic_fetch_min:
+    case AtomicExpr::AO__hip_atomic_fetch_min:
     case AtomicExpr::AO__opencl_atomic_fetch_min:
       LibCallName = E->getValueType()->isSignedIntegerType()
                         ? "__atomic_fetch_min"
@@ -1202,6 +1245,7 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
       LLVM_FALLTHROUGH;
     case AtomicExpr::AO__c11_atomic_fetch_max:
     case AtomicExpr::AO__atomic_fetch_max:
+    case AtomicExpr::AO__hip_atomic_fetch_max:
     case AtomicExpr::AO__opencl_atomic_fetch_max:
       LibCallName = E->getValueType()->isSignedIntegerType()
                         ? "__atomic_fetch_max"
@@ -1291,10 +1335,12 @@ RValue CodeGenFunction::EmitAtomicExpr(AtomicExpr *E) {
 
   bool IsStore = E->getOp() == AtomicExpr::AO__c11_atomic_store ||
                  E->getOp() == AtomicExpr::AO__opencl_atomic_store ||
+                 E->getOp() == AtomicExpr::AO__hip_atomic_store ||
                  E->getOp() == AtomicExpr::AO__atomic_store ||
                  E->getOp() == AtomicExpr::AO__atomic_store_n;
   bool IsLoad = E->getOp() == AtomicExpr::AO__c11_atomic_load ||
                 E->getOp() == AtomicExpr::AO__opencl_atomic_load ||
+                E->getOp() == AtomicExpr::AO__hip_atomic_load ||
                 E->getOp() == AtomicExpr::AO__atomic_load ||
                 E->getOp() == AtomicExpr::AO__atomic_load_n;
 
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 849423c8b9ba..5d6df59cc405 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -170,8 +170,9 @@ static Value *EmitNontemporalStore(CodeGenFunction &CGF, const CallExpr *E) {
 
   // Convert the type of the pointer to a pointer to the stored type.
   Val = CGF.EmitToMemory(Val, E->getArg(0)->getType());
+  unsigned SrcAddrSpace = Address->getType()->getPointerAddressSpace();
   Value *BC = CGF.Builder.CreateBitCast(
-      Address, llvm::PointerType::getUnqual(Val->getType()), "cast");
+      Address, llvm::PointerType::get(Val->getType(), SrcAddrSpace), "cast");
   LValue LV = CGF.MakeNaturalAlignAddrLValue(BC, E->getArg(0)->getType());
   LV.setNontemporal(true);
   CGF.EmitStoreOfScalar(Val, LV, false);
diff --git a/clang/lib/CodeGen/CGStmtOpenMP.cpp b/clang/lib/CodeGen/CGStmtOpenMP.cpp
index 4f14459e4d28..f6853a22cd36 100644
--- a/clang/lib/CodeGen/CGStmtOpenMP.cpp
+++ b/clang/lib/CodeGen/CGStmtOpenMP.cpp
@@ -4510,6 +4510,9 @@ void CodeGenFunction::EmitOMPTaskBasedDirective(
         Address Replacement(CGF.Builder.CreateLoad(Pair.second),
                             CGF.getContext().getDeclAlign(Pair.first));
         Scope.addPrivate(Pair.first, [Replacement]() { return Replacement; });
+        if (auto *DI = CGF.getDebugInfo())
+          DI->EmitDeclareOfAutoVariable(Pair.first, Pair.second.getPointer(),
+                                        CGF.Builder, /*UsePointerValue*/ true);
       }
       // Adjust mapping for internal locals by mapping actual memory instead of
       // a pointer to this memory.
diff --git a/clang/lib/CodeGen/CodeGenModule.cpp b/clang/lib/CodeGen/CodeGenModule.cpp
index 59f3e0270571..9ba1a5c25e81 100644
--- a/clang/lib/CodeGen/CodeGenModule.cpp
+++ b/clang/lib/CodeGen/CodeGenModule.cpp
@@ -710,10 +710,25 @@ void CodeGenModule::Release() {
                               1);
   }
 
-  if (Arch == llvm::Triple::aarch64 || Arch == llvm::Triple::aarch64_32 ||
+  // Add module metadata for return address signing (ignoring
+  // non-leaf/all) and stack tagging. These are actually turned on by function
+  // attributes, but we use module metadata to emit build attributes. This is
+  // needed for LTO, where the function attributes are inside bitcode
+  // serialised into a global variable by the time build attributes are
+  // emitted, so we can't access them.
+  if (Context.getTargetInfo().hasFeature("ptrauth") &&
+      LangOpts.getSignReturnAddressScope() !=
+          LangOptions::SignReturnAddressScopeKind::None)
+    getModule().addModuleFlag(llvm::Module::Override,
+                              "sign-return-address-buildattr", 1);
+  if (LangOpts.Sanitize.has(SanitizerKind::MemTag))
+    getModule().addModuleFlag(llvm::Module::Override,
+                              "tag-stack-memory-buildattr", 1);
+
+  if (Arch == llvm::Triple::thumb || Arch == llvm::Triple::thumbeb ||
+      Arch == llvm::Triple::aarch64 || Arch == llvm::Triple::aarch64_32 ||
       Arch == llvm::Triple::aarch64_be) {
-    getModule().addModuleFlag(llvm::Module::Error,
-                              "branch-target-enforcement",
+    getModule().addModuleFlag(llvm::Module::Error, "branch-target-enforcement",
                               LangOpts.BranchTargetEnforcement);
 
     getModule().addModuleFlag(llvm::Module::Error, "sign-return-address",
@@ -722,9 +737,11 @@ void CodeGenModule::Release() {
     getModule().addModuleFlag(llvm::Module::Error, "sign-return-address-all",
                               LangOpts.isSignReturnAddressScopeAll());
 
-    getModule().addModuleFlag(llvm::Module::Error,
-                              "sign-return-address-with-bkey",
-                              !LangOpts.isSignReturnAddressWithAKey());
+    if (Arch != llvm::Triple::thumb && Arch != llvm::Triple::thumbeb) {
+      getModule().addModuleFlag(llvm::Module::Error,
+                                "sign-return-address-with-bkey",
+                                !LangOpts.isSignReturnAddressWithAKey());
+    }
   }
 
   if (!CodeGenOpts.MemoryProfileOutput.empty()) {
@@ -1266,6 +1283,20 @@ static bool isUniqueInternalLinkageDecl(GlobalDecl GD,
          (CGM.getFunctionLinkage(GD) == llvm::GlobalValue::InternalLinkage);
 }
 
+static void AppendTargetClonesMangling(const CodeGenModule &CGM,
+                                       const TargetClonesAttr *Attr,
+                                       unsigned VersionIndex,
+                                       raw_ostream &Out) {
+  Out << '.';
+  StringRef FeatureStr = Attr->getFeatureStr(VersionIndex);
+  if (FeatureStr.startswith("arch="))
+    Out << "arch_" << FeatureStr.substr(sizeof("arch=") - 1);
+  else
+    Out << FeatureStr;
+
+  Out << '.' << Attr->getMangledIndex(VersionIndex);
+}
+
 static std::string getMangledNameImpl(CodeGenModule &CGM, GlobalDecl GD,
                                       const NamedDecl *ND,
                                       bool OmitMultiVersionMangling = false) {
@@ -1319,6 +1350,10 @@ static std::string getMangledNameImpl(CodeGenModule &CGM, GlobalDecl GD,
       case MultiVersionKind::Target:
         AppendTargetMangling(CGM, FD->getAttr<TargetAttr>(), Out);
         break;
+      case MultiVersionKind::TargetClones:
+        AppendTargetClonesMangling(CGM, FD->getAttr<TargetClonesAttr>(),
+                                   GD.getMultiVersionIndex(), Out);
+        break;
       case MultiVersionKind::None:
         llvm_unreachable("None multiversion type isn't valid here");
       }
@@ -1983,8 +2018,9 @@ bool CodeGenModule::GetCPUAndFeaturesAttributes(GlobalDecl GD,
   FD = FD ? FD->getMostRecentDecl() : FD;
   const auto *TD = FD ? FD->getAttr<TargetAttr>() : nullptr;
   const auto *SD = FD ? FD->getAttr<CPUSpecificAttr>() : nullptr;
+  const auto *TC = FD ? FD->getAttr<TargetClonesAttr>() : nullptr;
   bool AddedAttr = false;
-  if (TD || SD) {
+  if (TD || SD || TC) {
     llvm::StringMap<bool> FeatureMap;
     getContext().getFunctionFeatureMap(FeatureMap, GD);
 
@@ -3226,6 +3262,12 @@ void CodeGenModule::EmitMultiVersionFunctionDefinition(GlobalDecl GD,
     for (unsigned I = 0; I < Spec->cpus_size(); ++I)
       EmitGlobalFunctionDefinition(GD.getWithMultiVersionIndex(I), nullptr);
     // Requires multiple emits.
+  } else if (FD->isTargetClonesMultiVersion()) {
+    auto *Clone = FD->getAttr<TargetClonesAttr>();
+    for (unsigned I = 0; I < Clone->featuresStrs_size(); ++I)
+      if (Clone->isFirstOfVersion(I))
+        EmitGlobalFunctionDefinition(GD.getWithMultiVersionIndex(I), nullptr);
+    EmitTargetClonesResolver(GD);
   } else
     EmitGlobalFunctionDefinition(GD, GV);
 }
@@ -3307,6 +3349,63 @@ llvm::GlobalValue::LinkageTypes getMultiversionLinkage(CodeGenModule &CGM,
   return llvm::GlobalValue::WeakODRLinkage;
 }
 
+void CodeGenModule::EmitTargetClonesResolver(GlobalDecl GD) {
+  const auto *FD = cast<FunctionDecl>(GD.getDecl());
+  assert(FD && "Not a FunctionDecl?");
+  const auto *TC = FD->getAttr<TargetClonesAttr>();
+  assert(TC && "Not a target_clones Function?");
+
+  QualType CanonTy = Context.getCanonicalType(FD->getType());
+  llvm::Type *DeclTy = getTypes().ConvertType(CanonTy);
+
+  if (const auto *CXXFD = dyn_cast<CXXMethodDecl>(FD)) {
+    const CGFunctionInfo &FInfo = getTypes().arrangeCXXMethodDeclaration(CXXFD);
+    DeclTy = getTypes().GetFunctionType(FInfo);
+  }
+
+  llvm::Function *ResolverFunc;
+  if (getTarget().supportsIFunc()) {
+    auto *IFunc = cast<llvm::GlobalIFunc>(
+        GetOrCreateMultiVersionResolver(GD, DeclTy, FD));
+    ResolverFunc = cast<llvm::Function>(IFunc->getResolver());
+  } else
+    ResolverFunc =
+        cast<llvm::Function>(GetOrCreateMultiVersionResolver(GD, DeclTy, FD));
+
+  SmallVector<CodeGenFunction::MultiVersionResolverOption, 10> Options;
+  for (unsigned VersionIndex = 0; VersionIndex < TC->featuresStrs_size();
+       ++VersionIndex) {
+    if (!TC->isFirstOfVersion(VersionIndex))
+      continue;
+    StringRef Version = TC->getFeatureStr(VersionIndex);
+    StringRef MangledName =
+        getMangledName(GD.getWithMultiVersionIndex(VersionIndex));
+    llvm::Constant *Func = GetGlobalValue(MangledName);
+    assert(Func &&
+           "Should have already been created before calling resolver emit");
+
+    StringRef Architecture;
+    llvm::SmallVector<StringRef, 1> Feature;
+
+    if (Version.startswith("arch="))
+      Architecture = Version.drop_front(sizeof("arch=") - 1);
+    else if (Version != "default")
+      Feature.push_back(Version);
+
+    Options.emplace_back(cast<llvm::Function>(Func), Architecture, Feature);
+  }
+
+  const TargetInfo &TI = getTarget();
+  std::stable_sort(
+      Options.begin(), Options.end(),
+      [&TI](const CodeGenFunction::MultiVersionResolverOption &LHS,
+            const CodeGenFunction::MultiVersionResolverOption &RHS) {
+        return TargetMVPriority(TI, LHS) > TargetMVPriority(TI, RHS);
+      });
+  CodeGenFunction CGF(*this);
+  CGF.EmitMultiVersionResolver(ResolverFunc, Options);
+}
+
 void CodeGenModule::emitMultiVersionFunctions() {
   std::vector<GlobalDecl> MVFuncsToEmit;
   MultiVersionFuncs.swap(MVFuncsToEmit);
@@ -3511,8 +3610,25 @@ llvm::Constant *CodeGenModule::GetOrCreateMultiVersionResolver(
   // Since this is the first time we've created this IFunc, make sure
   // that we put this multiversioned function into the list to be
   // replaced later if necessary (target multiversioning only).
-  if (!FD->isCPUDispatchMultiVersion() && !FD->isCPUSpecificMultiVersion())
+  if (FD->isTargetMultiVersion())
     MultiVersionFuncs.push_back(GD);
+  else if (FD->isTargetClonesMultiVersion()) {
+    // In target_clones multiversioning, make sure we emit this if used.
+    auto DDI =
+        DeferredDecls.find(getMangledName(GD.getWithMultiVersionIndex(0)));
+    if (DDI != DeferredDecls.end()) {
+      addDeferredDeclToEmit(GD);
+      DeferredDecls.erase(DDI);
+    } else {
+      // Emit the symbol of the 1st variant, so that the deferred decls know we
+      // need it, otherwise the only global value will be the resolver/ifunc,
+      // which end up getting broken if we search for them with GetGlobalValue'.
+      GetOrCreateLLVMFunction(
+          getMangledName(GD.getWithMultiVersionIndex(0)), DeclTy, FD,
+          /*ForVTable=*/false, /*DontDefer=*/true,
+          /*IsThunk=*/false, llvm::AttributeList(), ForDefinition);
+    }
+  }
 
   if (getTarget().supportsIFunc()) {
     llvm::Type *ResolverType = llvm::FunctionType::get(
diff --git a/clang/lib/CodeGen/CodeGenModule.h b/clang/lib/CodeGen/CodeGenModule.h
index fbed22376c82..e1c7f486d334 100644
--- a/clang/lib/CodeGen/CodeGenModule.h
+++ b/clang/lib/CodeGen/CodeGenModule.h
@@ -1500,6 +1500,7 @@ private:
   void EmitAliasDefinition(GlobalDecl GD);
   void emitIFuncDefinition(GlobalDecl GD);
   void emitCPUDispatchDefinition(GlobalDecl GD);
+  void EmitTargetClonesResolver(GlobalDecl GD);
   void EmitObjCPropertyImplementations(const ObjCImplementationDecl *D);
   void EmitObjCIvarInitializations(ObjCImplementationDecl *D);
 
diff --git a/clang/lib/CodeGen/TargetInfo.cpp b/clang/lib/CodeGen/TargetInfo.cpp
index 302dc653c46e..36e0319c8ab9 100644
--- a/clang/lib/CodeGen/TargetInfo.cpp
+++ b/clang/lib/CodeGen/TargetInfo.cpp
@@ -6364,6 +6364,26 @@ public:
     const FunctionDecl *FD = dyn_cast_or_null<FunctionDecl>(D);
     if (!FD)
       return;
+    auto *Fn = cast<llvm::Function>(GV);
+
+    if (const auto *TA = FD->getAttr<TargetAttr>()) {
+      ParsedTargetAttr Attr = TA->parse();
+      if (!Attr.BranchProtection.empty()) {
+        TargetInfo::BranchProtectionInfo BPI;
+        StringRef DiagMsg;
+        (void)CGM.getTarget().validateBranchProtection(Attr.BranchProtection,
+                                                       BPI, DiagMsg);
+
+        static const char *SignReturnAddrStr[] = {"none", "non-leaf", "all"};
+        assert(static_cast<unsigned>(BPI.SignReturnAddr) <= 2 &&
+               "Unexpected SignReturnAddressScopeKind");
+        Fn->addFnAttr("sign-return-address",
+                      SignReturnAddrStr[static_cast<int>(BPI.SignReturnAddr)]);
+
+        Fn->addFnAttr("branch-target-enforcement",
+                      BPI.BranchTargetEnforcement ? "true" : "false");
+      }
+    }
 
     const ARMInterruptAttr *Attr = FD->getAttr<ARMInterruptAttr>();
     if (!Attr)
@@ -6379,8 +6399,6 @@ public:
     case ARMInterruptAttr::UNDEF:   Kind = "UNDEF"; break;
     }
 
-    llvm::Function *Fn = cast<llvm::Function>(GV);
-
     Fn->addFnAttr("interrupt", Kind);
 
     ARMABIInfo::ABIKind ABI = cast<ARMABIInfo>(getABIInfo()).getABIKind();
@@ -9339,17 +9357,25 @@ AMDGPUTargetCodeGenInfo::getLLVMSyncScopeID(const LangOptions &LangOpts,
                                             llvm::LLVMContext &Ctx) const {
   std::string Name;
   switch (Scope) {
+  case SyncScope::HIPSingleThread:
+    Name = "singlethread";
+    break;
+  case SyncScope::HIPWavefront:
+  case SyncScope::OpenCLSubGroup:
+    Name = "wavefront";
+    break;
+  case SyncScope::HIPWorkgroup:
   case SyncScope::OpenCLWorkGroup:
     Name = "workgroup";
     break;
+  case SyncScope::HIPAgent:
   case SyncScope::OpenCLDevice:
     Name = "agent";
     break;
+  case SyncScope::HIPSystem:
   case SyncScope::OpenCLAllSVMDevices:
     Name = "";
     break;
-  case SyncScope::OpenCLSubGroup:
-    Name = "wavefront";
   }
 
   if (Ordering != llvm::AtomicOrdering::SequentiallyConsistent) {
diff --git a/clang/lib/Driver/Driver.cpp b/clang/lib/Driver/Driver.cpp
index 8023d03013a1..d501bd026219 100644
--- a/clang/lib/Driver/Driver.cpp
+++ b/clang/lib/Driver/Driver.cpp
@@ -38,6 +38,7 @@
 #include "ToolChains/NaCl.h"
 #include "ToolChains/NetBSD.h"
 #include "ToolChains/OpenBSD.h"
+#include "ToolChains/PPCFreeBSD.h"
 #include "ToolChains/PPCLinux.h"
 #include "ToolChains/PS4CPU.h"
 #include "ToolChains/RISCVToolchain.h"
@@ -5302,7 +5303,11 @@ const ToolChain &Driver::getToolChain(const ArgList &Args,
       TC = std::make_unique<toolchains::NetBSD>(*this, Target, Args);
       break;
     case llvm::Triple::FreeBSD:
-      TC = std::make_unique<toolchains::FreeBSD>(*this, Target, Args);
+      if (Target.isPPC())
+        TC = std::make_unique<toolchains::PPCFreeBSDToolChain>(*this, Target,
+                                                               Args);
+      else
+        TC = std::make_unique<toolchains::FreeBSD>(*this, Target, Args);
       break;
     case llvm::Triple::Minix:
       TC = std::make_unique<toolchains::Minix>(*this, Target, Args);
diff --git a/clang/lib/Driver/ToolChains/Arch/AArch64.cpp b/clang/lib/Driver/ToolChains/Arch/AArch64.cpp
index 0b60d097b9ca..abc32f22d2a1 100644
--- a/clang/lib/Driver/ToolChains/Arch/AArch64.cpp
+++ b/clang/lib/Driver/ToolChains/Arch/AArch64.cpp
@@ -225,7 +225,7 @@ void aarch64::getAArch64TargetFeatures(const Driver &D,
   bool success = true;
   // Enable NEON by default.
   Features.push_back("+neon");
-  llvm::StringRef WaMArch = "";
+  llvm::StringRef WaMArch;
   if (ForAS)
     for (const auto *A :
          Args.filtered(options::OPT_Wa_COMMA, options::OPT_Xassembler))
@@ -235,7 +235,7 @@ void aarch64::getAArch64TargetFeatures(const Driver &D,
   // Call getAArch64ArchFeaturesFromMarch only if "-Wa,-march=" or
   // "-Xassembler -march" is detected. Otherwise it may return false
   // and causes Clang to error out.
-  if (WaMArch.size())
+  if (!WaMArch.empty())
     success = getAArch64ArchFeaturesFromMarch(D, WaMArch, Args, Features);
   else if ((A = Args.getLastArg(options::OPT_march_EQ)))
     success = getAArch64ArchFeaturesFromMarch(D, A->getValue(), Args, Features);
@@ -259,8 +259,15 @@ void aarch64::getAArch64TargetFeatures(const Driver &D,
     success = getAArch64MicroArchFeaturesFromMcpu(
         D, getAArch64TargetCPU(Args, Triple, A), Args, Features);
 
-  if (!success)
-    D.Diag(diag::err_drv_clang_unsupported) << A->getAsString(Args);
+  if (!success) {
+    auto Diag = D.Diag(diag::err_drv_clang_unsupported);
+    // If "-Wa,-march=" is used, 'WaMArch' will contain the argument's value,
+    // while 'A' is uninitialized. Only dereference 'A' in the other case.
+    if (!WaMArch.empty())
+      Diag << "-march=" + WaMArch.str();
+    else
+      Diag << A->getAsString(Args);
+  }
 
   if (Args.getLastArg(options::OPT_mgeneral_regs_only)) {
     Features.push_back("-fp-armv8");
diff --git a/clang/lib/Driver/ToolChains/Clang.cpp b/clang/lib/Driver/ToolChains/Clang.cpp
index 55518cd7926f..c5aaa067c4f5 100644
--- a/clang/lib/Driver/ToolChains/Clang.cpp
+++ b/clang/lib/Driver/ToolChains/Clang.cpp
@@ -403,7 +403,7 @@ shouldUseExceptionTablesForObjCExceptions(const ObjCRuntime &runtime,
 }
 
 /// Adds exception related arguments to the driver command arguments. There's a
-/// master flag, -fexceptions and also language specific flags to enable/disable
+/// main flag, -fexceptions and also language specific flags to enable/disable
 /// C++ and Objective-C exceptions. This makes it possible to for example
 /// disable C++ exceptions but enable Objective-C exceptions.
 static bool addExceptionArgs(const ArgList &Args, types::ID InputType,
@@ -1603,6 +1603,49 @@ void RenderARMABI(const Driver &D, const llvm::Triple &Triple,
 }
 }
 
+static void CollectARMPACBTIOptions(const Driver &D, const ArgList &Args,
+                                    ArgStringList &CmdArgs, bool isAArch64) {
+  const Arg *A = isAArch64
+                     ? Args.getLastArg(options::OPT_msign_return_address_EQ,
+                                       options::OPT_mbranch_protection_EQ)
+                     : Args.getLastArg(options::OPT_mbranch_protection_EQ);
+  if (!A)
+    return;
+
+  StringRef Scope, Key;
+  bool IndirectBranches;
+
+  if (A->getOption().matches(options::OPT_msign_return_address_EQ)) {
+    Scope = A->getValue();
+    if (!Scope.equals("none") && !Scope.equals("non-leaf") &&
+        !Scope.equals("all"))
+      D.Diag(diag::err_invalid_branch_protection)
+          << Scope << A->getAsString(Args);
+    Key = "a_key";
+    IndirectBranches = false;
+  } else {
+    StringRef DiagMsg;
+    llvm::ARM::ParsedBranchProtection PBP;
+    if (!llvm::ARM::parseBranchProtection(A->getValue(), PBP, DiagMsg))
+      D.Diag(diag::err_invalid_branch_protection)
+          << DiagMsg << A->getAsString(Args);
+    if (!isAArch64 && PBP.Key == "b_key")
+      D.Diag(diag::warn_unsupported_branch_protection)
+          << "b-key" << A->getAsString(Args);
+    Scope = PBP.Scope;
+    Key = PBP.Key;
+    IndirectBranches = PBP.BranchTargetEnforcement;
+  }
+
+  CmdArgs.push_back(
+      Args.MakeArgString(Twine("-msign-return-address=") + Scope));
+  if (!Scope.equals("none"))
+    CmdArgs.push_back(
+        Args.MakeArgString(Twine("-msign-return-address-key=") + Key));
+  if (IndirectBranches)
+    CmdArgs.push_back("-mbranch-target-enforce");
+}
+
 void Clang::AddARMTargetArgs(const llvm::Triple &Triple, const ArgList &Args,
                              ArgStringList &CmdArgs, bool KernelOrKext) const {
   RenderARMABI(getToolChain().getDriver(), Triple, Args, CmdArgs);
@@ -1644,6 +1687,10 @@ void Clang::AddARMTargetArgs(const llvm::Triple &Triple, const ArgList &Args,
     CmdArgs.push_back("-mcmse");
 
   AddAAPCSVolatileBitfieldArgs(Args, CmdArgs);
+
+  // Enable/disable return address signing and indirect branch targets.
+  CollectARMPACBTIOptions(getToolChain().getDriver(), Args, CmdArgs,
+                          false /*isAArch64*/);
 }
 
 void Clang::RenderTargetOptions(const llvm::Triple &EffectiveTriple,
@@ -1783,40 +1830,8 @@ void Clang::AddAArch64TargetArgs(const ArgList &Args,
   }
 
   // Enable/disable return address signing and indirect branch targets.
-  if (Arg *A = Args.getLastArg(options::OPT_msign_return_address_EQ,
-                               options::OPT_mbranch_protection_EQ)) {
-
-    const Driver &D = getToolChain().getDriver();
-
-    StringRef Scope, Key;
-    bool IndirectBranches;
-
-    if (A->getOption().matches(options::OPT_msign_return_address_EQ)) {
-      Scope = A->getValue();
-      if (!Scope.equals("none") && !Scope.equals("non-leaf") &&
-          !Scope.equals("all"))
-        D.Diag(diag::err_invalid_branch_protection)
-            << Scope << A->getAsString(Args);
-      Key = "a_key";
-      IndirectBranches = false;
-    } else {
-      StringRef Err;
-      llvm::AArch64::ParsedBranchProtection PBP;
-      if (!llvm::AArch64::parseBranchProtection(A->getValue(), PBP, Err))
-        D.Diag(diag::err_invalid_branch_protection)
-            << Err << A->getAsString(Args);
-      Scope = PBP.Scope;
-      Key = PBP.Key;
-      IndirectBranches = PBP.BranchTargetEnforcement;
-    }
-
-    CmdArgs.push_back(
-        Args.MakeArgString(Twine("-msign-return-address=") + Scope));
-    CmdArgs.push_back(
-        Args.MakeArgString(Twine("-msign-return-address-key=") + Key));
-    if (IndirectBranches)
-      CmdArgs.push_back("-mbranch-target-enforce");
-  }
+  CollectARMPACBTIOptions(getToolChain().getDriver(), Args, CmdArgs,
+                          true /*isAArch64*/);
 
   // Handle -msve_vector_bits=<bits>
   if (Arg *A = Args.getLastArg(options::OPT_msve_vector_bits_EQ)) {
@@ -5821,9 +5836,12 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA,
   Args.AddLastArg(CmdArgs, options::OPT_fvisibility_inlines_hidden_static_local_var,
                            options::OPT_fno_visibility_inlines_hidden_static_local_var);
   Args.AddLastArg(CmdArgs, options::OPT_fvisibility_global_new_delete_hidden);
-  Args.AddLastArg(CmdArgs, options::OPT_fnew_infallible);
   Args.AddLastArg(CmdArgs, options::OPT_ftlsmodel_EQ);
 
+  if (Args.hasFlag(options::OPT_fnew_infallible,
+                   options::OPT_fno_new_infallible, false))
+    CmdArgs.push_back("-fnew-infallible");
+
   if (Args.hasFlag(options::OPT_fno_operator_names,
                    options::OPT_foperator_names, false))
     CmdArgs.push_back("-fno-operator-names");
@@ -5886,7 +5904,7 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA,
       // runtime.
       if (Args.hasFlag(options::OPT_fopenmp_target_new_runtime,
                        options::OPT_fno_openmp_target_new_runtime,
-                       /*Default=*/false))
+                       /*Default=*/!getToolChain().getTriple().isAMDGCN()))
         CmdArgs.push_back("-fopenmp-target-new-runtime");
 
       // When in OpenMP offloading mode, enable debugging on the device.
@@ -6659,6 +6677,35 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA,
     A->claim();
   }
 
+  // Turn -fplugin-arg-pluginname-key=value into
+  // -plugin-arg-pluginname key=value
+  // GCC has an actual plugin_argument struct with key/value pairs that it
+  // passes to its plugins, but we don't, so just pass it on as-is.
+  //
+  // The syntax for -fplugin-arg- is ambiguous if both plugin name and
+  // argument key are allowed to contain dashes. GCC therefore only
+  // allows dashes in the key. We do the same.
+  for (const Arg *A : Args.filtered(options::OPT_fplugin_arg)) {
+    auto ArgValue = StringRef(A->getValue());
+    auto FirstDashIndex = ArgValue.find('-');
+    StringRef PluginName = ArgValue.substr(0, FirstDashIndex);
+    StringRef Arg = ArgValue.substr(FirstDashIndex + 1);
+
+    A->claim();
+    if (FirstDashIndex == StringRef::npos || Arg.empty()) {
+      if (PluginName.empty()) {
+        D.Diag(diag::warn_drv_missing_plugin_name) << A->getAsString(Args);
+      } else {
+        D.Diag(diag::warn_drv_missing_plugin_arg)
+            << PluginName << A->getAsString(Args);
+      }
+      continue;
+    }
+
+    CmdArgs.push_back(Args.MakeArgString(Twine("-plugin-arg-") + PluginName));
+    CmdArgs.push_back(Args.MakeArgString(Arg));
+  }
+
   // Forward -fpass-plugin=name.so to -cc1.
   for (const Arg *A : Args.filtered(options::OPT_fpass_plugin_EQ)) {
     CmdArgs.push_back(
diff --git a/clang/lib/Driver/ToolChains/Cuda.cpp b/clang/lib/Driver/ToolChains/Cuda.cpp
index 5397c7a9a0e6..ee573b89bed1 100644
--- a/clang/lib/Driver/ToolChains/Cuda.cpp
+++ b/clang/lib/Driver/ToolChains/Cuda.cpp
@@ -745,7 +745,7 @@ void CudaToolChain::addClangTargetOptions(
 
     std::string BitcodeSuffix;
     if (DriverArgs.hasFlag(options::OPT_fopenmp_target_new_runtime,
-                           options::OPT_fno_openmp_target_new_runtime, false))
+                           options::OPT_fno_openmp_target_new_runtime, true))
       BitcodeSuffix = "new-nvptx-" + GpuArch.str();
     else
       BitcodeSuffix = "nvptx-" + GpuArch.str();
diff --git a/clang/lib/Driver/ToolChains/Flang.cpp b/clang/lib/Driver/ToolChains/Flang.cpp
index b82c5d7600df..c169e3d45793 100644
--- a/clang/lib/Driver/ToolChains/Flang.cpp
+++ b/clang/lib/Driver/ToolChains/Flang.cpp
@@ -32,7 +32,8 @@ void Flang::AddFortranDialectOptions(const ArgList &Args,
                 options::OPT_fxor_operator, options::OPT_fno_xor_operator,
                 options::OPT_falternative_parameter_statement,
                 options::OPT_fdefault_real_8, options::OPT_fdefault_integer_8,
-                options::OPT_fdefault_double_8, options::OPT_flarge_sizes});
+                options::OPT_fdefault_double_8, options::OPT_flarge_sizes,
+                options::OPT_fno_automatic});
 }
 
 void Flang::AddPreprocessingOptions(const ArgList &Args,
diff --git a/clang/lib/Driver/ToolChains/FreeBSD.cpp b/clang/lib/Driver/ToolChains/FreeBSD.cpp
index dc05f9893465..d08ea282f6df 100644
--- a/clang/lib/Driver/ToolChains/FreeBSD.cpp
+++ b/clang/lib/Driver/ToolChains/FreeBSD.cpp
@@ -391,7 +391,8 @@ FreeBSD::FreeBSD(const Driver &D, const llvm::Triple &Triple,
 }
 
 ToolChain::CXXStdlibType FreeBSD::GetDefaultCXXStdlibType() const {
-  if (getTriple().getOSMajorVersion() >= 10)
+  unsigned Major = getTriple().getOSMajorVersion();
+  if (Major >= 10 || Major == 0)
     return ToolChain::CST_Libcxx;
   return ToolChain::CST_Libstdcxx;
 }
diff --git a/clang/lib/Driver/ToolChains/Linux.cpp b/clang/lib/Driver/ToolChains/Linux.cpp
index 0224383e63a1..198774506e5e 100644
--- a/clang/lib/Driver/ToolChains/Linux.cpp
+++ b/clang/lib/Driver/ToolChains/Linux.cpp
@@ -421,6 +421,9 @@ std::string Linux::getDynamicLinker(const ArgList &Args) const {
         (Triple.getEnvironment() == llvm::Triple::MuslEABIHF ||
          tools::arm::getARMFloatABI(*this, Args) == tools::arm::FloatABI::Hard))
       ArchName += "hf";
+    if (Arch == llvm::Triple::ppc &&
+        Triple.getSubArch() == llvm::Triple::PPCSubArch_spe)
+      ArchName = "powerpc-sf";
 
     return "/lib/ld-musl-" + ArchName + ".so.1";
   }
diff --git a/clang/lib/Driver/ToolChains/PPCFreeBSD.cpp b/clang/lib/Driver/ToolChains/PPCFreeBSD.cpp
new file mode 100644
index 000000000000..8d381c4f1437
--- /dev/null
+++ b/clang/lib/Driver/ToolChains/PPCFreeBSD.cpp
@@ -0,0 +1,28 @@
+//===-- PPCFreeBSD.cpp - PowerPC ToolChain Implementations ------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "PPCFreeBSD.h"
+#include "clang/Driver/Driver.h"
+#include "clang/Driver/Options.h"
+#include "llvm/Support/Path.h"
+
+using namespace clang::driver::toolchains;
+using namespace llvm::opt;
+
+void PPCFreeBSDToolChain::AddClangSystemIncludeArgs(
+    const ArgList &DriverArgs, ArgStringList &CC1Args) const {
+  if (!DriverArgs.hasArg(clang::driver::options::OPT_nostdinc) &&
+      !DriverArgs.hasArg(options::OPT_nobuiltininc)) {
+    const Driver &D = getDriver();
+    SmallString<128> P(D.ResourceDir);
+    llvm::sys::path::append(P, "include", "ppc_wrappers");
+    addSystemInclude(DriverArgs, CC1Args, P);
+  }
+
+  FreeBSD::AddClangSystemIncludeArgs(DriverArgs, CC1Args);
+}
diff --git a/clang/lib/Driver/ToolChains/PPCFreeBSD.h b/clang/lib/Driver/ToolChains/PPCFreeBSD.h
new file mode 100644
index 000000000000..d5d9cf4e83a0
--- /dev/null
+++ b/clang/lib/Driver/ToolChains/PPCFreeBSD.h
@@ -0,0 +1,33 @@
+//===--- PPCFreeBSD.h - PowerPC ToolChain Implementations -------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_CLANG_LIB_DRIVER_TOOLCHAINS_PPC_FREEBSD_H
+#define LLVM_CLANG_LIB_DRIVER_TOOLCHAINS_PPC_FREEBSD_H
+
+#include "FreeBSD.h"
+
+namespace clang {
+namespace driver {
+namespace toolchains {
+
+class LLVM_LIBRARY_VISIBILITY PPCFreeBSDToolChain : public FreeBSD {
+public:
+  PPCFreeBSDToolChain(const Driver &D, const llvm::Triple &Triple,
+                      const llvm::opt::ArgList &Args)
+      : FreeBSD(D, Triple, Args) {}
+
+  void
+  AddClangSystemIncludeArgs(const llvm::opt::ArgList &DriverArgs,
+                            llvm::opt::ArgStringList &CC1Args) const override;
+};
+
+} // end namespace toolchains
+} // end namespace driver
+} // end namespace clang
+
+#endif // LLVM_CLANG_LIB_DRIVER_TOOLCHAINS_PPC_FREEBSD_H
diff --git a/clang/lib/Driver/ToolChains/PS4CPU.h b/clang/lib/Driver/ToolChains/PS4CPU.h
index 82f9523f84fb..4bedabaf267c 100644
--- a/clang/lib/Driver/ToolChains/PS4CPU.h
+++ b/clang/lib/Driver/ToolChains/PS4CPU.h
@@ -80,6 +80,7 @@ public:
     return LangOptions::SSPStrong;
   }
 
+  unsigned GetDefaultDwarfVersion() const override { return 4; }
   llvm::DebuggerKind getDefaultDebuggerTuning() const override {
     return llvm::DebuggerKind::SCE;
   }
diff --git a/clang/lib/Format/ContinuationIndenter.cpp b/clang/lib/Format/ContinuationIndenter.cpp
index 1e4f5690ef24..5073f5105d05 100644
--- a/clang/lib/Format/ContinuationIndenter.cpp
+++ b/clang/lib/Format/ContinuationIndenter.cpp
@@ -1984,9 +1984,17 @@ ContinuationIndenter::createBreakableToken(const FormatToken &Current,
   } else if (Current.is(TT_LineComment) &&
              (Current.Previous == nullptr ||
               Current.Previous->isNot(TT_ImplicitStringLiteral))) {
+    bool RegularComments = [&]() {
+      for (const FormatToken *T = &Current; T && T->is(TT_LineComment);
+           T = T->Next) {
+        if (!(T->TokenText.startswith("//") || T->TokenText.startswith("#")))
+          return false;
+      }
+      return true;
+    }();
     if (!Style.ReflowComments ||
         CommentPragmasRegex.match(Current.TokenText.substr(2)) ||
-        switchesFormatting(Current))
+        switchesFormatting(Current) || !RegularComments)
       return nullptr;
     return std::make_unique<BreakableLineCommentSection>(
         Current, StartColumn, /*InPPDirective=*/false, Encoding, Style);
@@ -2195,11 +2203,10 @@ ContinuationIndenter::breakProtrudingToken(const FormatToken &Current,
       // When breaking before a tab character, it may be moved by a few columns,
       // but will still be expanded to the next tab stop, so we don't save any
       // columns.
-      if (NewRemainingTokenColumns == RemainingTokenColumns) {
+      if (NewRemainingTokenColumns >= RemainingTokenColumns) {
         // FIXME: Do we need to adjust the penalty?
         break;
       }
-      assert(NewRemainingTokenColumns < RemainingTokenColumns);
 
       LLVM_DEBUG(llvm::dbgs() << "    Breaking at: " << TailOffset + Split.first
                               << ", " << Split.second << "\n");
diff --git a/clang/lib/Format/Format.cpp b/clang/lib/Format/Format.cpp
index 8ae29c54a762..17de1075aeaa 100644
--- a/clang/lib/Format/Format.cpp
+++ b/clang/lib/Format/Format.cpp
@@ -2988,9 +2988,8 @@ reformat(const FormatStyle &Style, StringRef Code,
   // JSON only needs the formatting passing.
   if (Style.isJson()) {
     std::vector<tooling::Range> Ranges(1, tooling::Range(0, Code.size()));
-    auto Env =
-        Environment::make(Code, FileName, Ranges, FirstStartColumn,
-                                      NextStartColumn, LastStartColumn);
+    auto Env = Environment::make(Code, FileName, Ranges, FirstStartColumn,
+                                 NextStartColumn, LastStartColumn);
     if (!Env)
       return {};
     // Perform the actual formatting pass.
@@ -3118,9 +3117,7 @@ tooling::Replacements fixNamespaceEndComments(const FormatStyle &Style,
   auto Env = Environment::make(Code, FileName, Ranges);
   if (!Env)
     return {};
-  return NamespaceEndCommentsFixer(*Env, Style)
-      .process()
-      .first;
+  return NamespaceEndCommentsFixer(*Env, Style).process().first;
 }
 
 tooling::Replacements sortUsingDeclarations(const FormatStyle &Style,
@@ -3130,9 +3127,7 @@ tooling::Replacements sortUsingDeclarations(const FormatStyle &Style,
   auto Env = Environment::make(Code, FileName, Ranges);
   if (!Env)
     return {};
-  return UsingDeclarationsSorter(*Env, Style)
-      .process()
-      .first;
+  return UsingDeclarationsSorter(*Env, Style).process().first;
 }
 
 LangOptions getFormattingLangOpts(const FormatStyle &Style) {
diff --git a/clang/lib/Format/FormatToken.h b/clang/lib/Format/FormatToken.h
index 06d51dd95f50..1a2858018fde 100644
--- a/clang/lib/Format/FormatToken.h
+++ b/clang/lib/Format/FormatToken.h
@@ -76,6 +76,7 @@ namespace format {
   TYPE(LineComment)                                                            \
   TYPE(MacroBlockBegin)                                                        \
   TYPE(MacroBlockEnd)                                                          \
+  TYPE(ModulePartitionColon)                                                   \
   TYPE(NamespaceMacro)                                                         \
   TYPE(NonNullAssertion)                                                       \
   TYPE(NullCoalescingEqual)                                                    \
diff --git a/clang/lib/Format/FormatTokenLexer.cpp b/clang/lib/Format/FormatTokenLexer.cpp
index 8075756cca03..64fbd2d5d45b 100644
--- a/clang/lib/Format/FormatTokenLexer.cpp
+++ b/clang/lib/Format/FormatTokenLexer.cpp
@@ -37,27 +37,40 @@ FormatTokenLexer::FormatTokenLexer(
                       getFormattingLangOpts(Style)));
   Lex->SetKeepWhitespaceMode(true);
 
-  for (const std::string &ForEachMacro : Style.ForEachMacros)
-    Macros.insert({&IdentTable.get(ForEachMacro), TT_ForEachMacro});
-  for (const std::string &IfMacro : Style.IfMacros)
-    Macros.insert({&IdentTable.get(IfMacro), TT_IfMacro});
-  for (const std::string &AttributeMacro : Style.AttributeMacros)
-    Macros.insert({&IdentTable.get(AttributeMacro), TT_AttributeMacro});
-  for (const std::string &StatementMacro : Style.StatementMacros)
-    Macros.insert({&IdentTable.get(StatementMacro), TT_StatementMacro});
-  for (const std::string &TypenameMacro : Style.TypenameMacros)
-    Macros.insert({&IdentTable.get(TypenameMacro), TT_TypenameMacro});
-  for (const std::string &NamespaceMacro : Style.NamespaceMacros)
-    Macros.insert({&IdentTable.get(NamespaceMacro), TT_NamespaceMacro});
+  for (const std::string &ForEachMacro : Style.ForEachMacros) {
+    auto Identifier = &IdentTable.get(ForEachMacro);
+    Macros.insert({Identifier, TT_ForEachMacro});
+  }
+  for (const std::string &IfMacro : Style.IfMacros) {
+    auto Identifier = &IdentTable.get(IfMacro);
+    Macros.insert({Identifier, TT_IfMacro});
+  }
+  for (const std::string &AttributeMacro : Style.AttributeMacros) {
+    auto Identifier = &IdentTable.get(AttributeMacro);
+    Macros.insert({Identifier, TT_AttributeMacro});
+  }
+  for (const std::string &StatementMacro : Style.StatementMacros) {
+    auto Identifier = &IdentTable.get(StatementMacro);
+    Macros.insert({Identifier, TT_StatementMacro});
+  }
+  for (const std::string &TypenameMacro : Style.TypenameMacros) {
+    auto Identifier = &IdentTable.get(TypenameMacro);
+    Macros.insert({Identifier, TT_TypenameMacro});
+  }
+  for (const std::string &NamespaceMacro : Style.NamespaceMacros) {
+    auto Identifier = &IdentTable.get(NamespaceMacro);
+    Macros.insert({Identifier, TT_NamespaceMacro});
+  }
   for (const std::string &WhitespaceSensitiveMacro :
        Style.WhitespaceSensitiveMacros) {
-    Macros.insert(
-        {&IdentTable.get(WhitespaceSensitiveMacro), TT_UntouchableMacroFunc});
+    auto Identifier = &IdentTable.get(WhitespaceSensitiveMacro);
+    Macros.insert({Identifier, TT_UntouchableMacroFunc});
   }
   for (const std::string &StatementAttributeLikeMacro :
-       Style.StatementAttributeLikeMacros)
-    Macros.insert({&IdentTable.get(StatementAttributeLikeMacro),
-                   TT_StatementAttributeLikeMacro});
+       Style.StatementAttributeLikeMacros) {
+    auto Identifier = &IdentTable.get(StatementAttributeLikeMacro);
+    Macros.insert({Identifier, TT_StatementAttributeLikeMacro});
+  }
 }
 
 ArrayRef<FormatToken *> FormatTokenLexer::lex() {
@@ -739,6 +752,8 @@ bool FormatTokenLexer::tryMerge_TMacro() {
   Tokens.pop_back();
   Tokens.pop_back();
   Tokens.back() = String;
+  if (FirstInLineIndex >= Tokens.size())
+    FirstInLineIndex = Tokens.size() - 1;
   return true;
 }
 
diff --git a/clang/lib/Format/SortJavaScriptImports.cpp b/clang/lib/Format/SortJavaScriptImports.cpp
index 515cfce725a4..77dc0d683e5f 100644
--- a/clang/lib/Format/SortJavaScriptImports.cpp
+++ b/clang/lib/Format/SortJavaScriptImports.cpp
@@ -553,9 +553,7 @@ tooling::Replacements sortJavaScriptImports(const FormatStyle &Style,
   auto Env = Environment::make(Code, FileName, Ranges);
   if (!Env)
     return {};
-  return JavaScriptImportSorter(*Env, Style)
-      .process()
-      .first;
+  return JavaScriptImportSorter(*Env, Style).process().first;
 }
 
 } // end namespace format
diff --git a/clang/lib/Format/TokenAnalyzer.cpp b/clang/lib/Format/TokenAnalyzer.cpp
index a619c6d939e9..d83e837ca134 100644
--- a/clang/lib/Format/TokenAnalyzer.cpp
+++ b/clang/lib/Format/TokenAnalyzer.cpp
@@ -37,7 +37,7 @@ namespace format {
 
 // FIXME: Instead of printing the diagnostic we should store it and have a
 // better way to return errors through the format APIs.
-class FatalDiagnosticConsumer: public DiagnosticConsumer {
+class FatalDiagnosticConsumer : public DiagnosticConsumer {
 public:
   void HandleDiagnostic(DiagnosticsEngine::Level DiagLevel,
                         const Diagnostic &Info) override {
@@ -71,7 +71,8 @@ Environment::make(StringRef Code, StringRef FileName,
   }
   // Validate that we can get the buffer data without a fatal error.
   Env->SM.getBufferData(Env->ID);
-  if (Diags.fatalError()) return nullptr;
+  if (Diags.fatalError())
+    return nullptr;
   return Env;
 }
 
@@ -80,8 +81,7 @@ Environment::Environment(StringRef Code, StringRef FileName,
                          unsigned LastStartColumn)
     : VirtualSM(new SourceManagerForFile(FileName, Code)), SM(VirtualSM->get()),
       ID(VirtualSM->get().getMainFileID()), FirstStartColumn(FirstStartColumn),
-      NextStartColumn(NextStartColumn), LastStartColumn(LastStartColumn) {
-}
+      NextStartColumn(NextStartColumn), LastStartColumn(LastStartColumn) {}
 
 TokenAnalyzer::TokenAnalyzer(const Environment &Env, const FormatStyle &Style)
     : Style(Style), Env(Env),
diff --git a/clang/lib/Format/TokenAnnotator.cpp b/clang/lib/Format/TokenAnnotator.cpp
index 3897241cb858..a94d8cdc3b04 100644
--- a/clang/lib/Format/TokenAnnotator.cpp
+++ b/clang/lib/Format/TokenAnnotator.cpp
@@ -314,10 +314,11 @@ private:
     //
     // void (*FunctionPointer)(void);
     // void (&FunctionReference)(void);
+    // void (&&FunctionReference)(void);
     // void (^ObjCBlock)(void);
     bool MightBeFunctionType = !Contexts[Contexts.size() - 2].IsExpression;
     bool ProbablyFunctionType =
-        CurrentToken->isOneOf(tok::star, tok::amp, tok::caret);
+        CurrentToken->isOneOf(tok::star, tok::amp, tok::ampamp, tok::caret);
     bool HasMultipleLines = false;
     bool HasMultipleParametersOnALine = false;
     bool MightBeObjCForRangeLoop =
@@ -902,9 +903,13 @@ private:
           break;
         }
       }
-      if (Contexts.back().ColonIsDictLiteral ||
-          Style.Language == FormatStyle::LK_Proto ||
-          Style.Language == FormatStyle::LK_TextProto) {
+      if (Line.First->isOneOf(Keywords.kw_module, Keywords.kw_import) ||
+          Line.First->startsSequence(tok::kw_export, Keywords.kw_module) ||
+          Line.First->startsSequence(tok::kw_export, Keywords.kw_import)) {
+        Tok->setType(TT_ModulePartitionColon);
+      } else if (Contexts.back().ColonIsDictLiteral ||
+                 Style.Language == FormatStyle::LK_Proto ||
+                 Style.Language == FormatStyle::LK_TextProto) {
         Tok->setType(TT_DictLiteral);
         if (Style.Language == FormatStyle::LK_TextProto) {
           if (FormatToken *Previous = Tok->getPreviousNonComment())
@@ -946,11 +951,15 @@ private:
                  !Line.First->isOneOf(tok::kw_enum, tok::kw_case,
                                       tok::kw_default)) {
         FormatToken *Prev = Tok->getPreviousNonComment();
+        if (!Prev)
+          break;
         if (Prev->isOneOf(tok::r_paren, tok::kw_noexcept))
           Tok->setType(TT_CtorInitializerColon);
         else if (Prev->is(tok::kw_try)) {
           // Member initializer list within function try block.
           FormatToken *PrevPrev = Prev->getPreviousNonComment();
+          if (!PrevPrev)
+            break;
           if (PrevPrev && PrevPrev->isOneOf(tok::r_paren, tok::kw_noexcept))
             Tok->setType(TT_CtorInitializerColon);
         } else
@@ -995,6 +1004,8 @@ private:
         if (CurrentToken && CurrentToken->is(Keywords.kw_await))
           next();
       }
+      if (Style.isCpp() && CurrentToken && CurrentToken->is(tok::kw_co_await))
+        next();
       Contexts.back().ColonIsForRangeExpr = true;
       next();
       if (!parseParens())
@@ -1578,6 +1589,8 @@ private:
         if (TemplateCloser->is(tok::l_paren)) {
           // No Matching Paren yet so skip to matching paren
           TemplateCloser = untilMatchingParen(TemplateCloser);
+          if (!TemplateCloser)
+            break;
         }
         if (TemplateCloser->is(tok::less))
           NestingLevel++;
@@ -2336,16 +2349,15 @@ void TokenAnnotator::setCommentLineLevels(
     if (NextNonCommentLine && CommentLine &&
         NextNonCommentLine->First->NewlinesBefore <= 1 &&
         NextNonCommentLine->First->OriginalColumn ==
-        AL->First->OriginalColumn) {
+            AL->First->OriginalColumn) {
       // Align comments for preprocessor lines with the # in column 0 if
       // preprocessor lines are not indented. Otherwise, align with the next
       // line.
-      AL->Level =
-          (Style.IndentPPDirectives != FormatStyle::PPDIS_BeforeHash &&
-           (NextNonCommentLine->Type == LT_PreprocessorDirective ||
-            NextNonCommentLine->Type == LT_ImportStatement))
-              ? 0
-              : NextNonCommentLine->Level;
+      AL->Level = (Style.IndentPPDirectives != FormatStyle::PPDIS_BeforeHash &&
+                   (NextNonCommentLine->Type == LT_PreprocessorDirective ||
+                    NextNonCommentLine->Type == LT_ImportStatement))
+                      ? 0
+                      : NextNonCommentLine->Level;
     } else {
       NextNonCommentLine = AL->First->isNot(tok::r_brace) ? AL : nullptr;
     }
@@ -2639,8 +2651,8 @@ void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) {
     if (Current->Role)
       Current->Role->precomputeFormattingInfos(Current);
     if (Current->MatchingParen &&
-        Current->MatchingParen->opensBlockOrBlockTypeList(Style)) {
-      assert(IndentLevel > 0);
+        Current->MatchingParen->opensBlockOrBlockTypeList(Style) &&
+        IndentLevel > 0) {
       --IndentLevel;
     }
     Current->IndentLevel = IndentLevel;
@@ -2942,6 +2954,14 @@ bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line,
   if (Left.is(tok::kw_auto) && Right.isOneOf(tok::l_paren, tok::l_brace))
     return false;
 
+  // operator co_await(x)
+  if (Right.is(tok::l_paren) && Left.is(tok::kw_co_await) && Left.Previous &&
+      Left.Previous->is(tok::kw_operator))
+    return false;
+  // co_await (x), co_yield (x), co_return (x)
+  if (Left.isOneOf(tok::kw_co_await, tok::kw_co_yield, tok::kw_co_return) &&
+      Right.isNot(tok::semi))
+    return true;
   // requires clause Concept1<T> && Concept2<T>
   if (Left.is(TT_ConstraintJunctions) && Right.is(tok::identifier))
     return true;
@@ -3159,9 +3179,13 @@ bool TokenAnnotator::spaceRequiredBetween(const AnnotatedLine &Line,
     if (Left.isIf(Line.Type != LT_PreprocessorDirective))
       return Style.SpaceBeforeParensOptions.AfterControlStatements ||
              spaceRequiredBeforeParens(Right);
+
+    // TODO add Operator overloading specific Options to
+    // SpaceBeforeParensOptions
+    if (Right.is(TT_OverloadedOperatorLParen))
+      return spaceRequiredBeforeParens(Right);
     // Function declaration or definition
-    if (Line.MightBeFunctionDecl && (Left.is(TT_FunctionDeclarationName) ||
-                                     Right.is(TT_OverloadedOperatorLParen))) {
+    if (Line.MightBeFunctionDecl && (Left.is(TT_FunctionDeclarationName))) {
       if (Line.mightBeFunctionDefinition())
         return Style.SpaceBeforeParensOptions.AfterFunctionDefinitionName ||
                spaceRequiredBeforeParens(Right);
@@ -3238,9 +3262,35 @@ bool TokenAnnotator::spaceRequiredBefore(const AnnotatedLine &Line,
   auto HasExistingWhitespace = [&Right]() {
     return Right.WhitespaceRange.getBegin() != Right.WhitespaceRange.getEnd();
   };
+
   if (Right.Tok.getIdentifierInfo() && Left.Tok.getIdentifierInfo())
     return true; // Never ever merge two identifiers.
+
+  // Leave a space between * and /* to avoid C4138 `comment end` found outside
+  // of comment.
+  if (Left.is(tok::star) && Right.is(tok::comment))
+    return true;
+
   if (Style.isCpp()) {
+    // Space between import <iostream>.
+    // or import .....;
+    if (Left.is(Keywords.kw_import) && Right.isOneOf(tok::less, tok::ellipsis))
+      return true;
+    // No space between module :.
+    if (Left.isOneOf(Keywords.kw_module, Keywords.kw_import) &&
+        Right.is(TT_ModulePartitionColon))
+      return true;
+    // No space between import foo:bar but keep a space between import :bar;
+    if (Left.is(tok::identifier) && Right.is(TT_ModulePartitionColon))
+      return false;
+    // No space between :bar;
+    if (Left.is(TT_ModulePartitionColon) &&
+        Right.isOneOf(tok::identifier, tok::kw_private))
+      return false;
+    if (Left.is(tok::ellipsis) && Right.is(tok::identifier) &&
+        Line.First->is(Keywords.kw_import))
+      return false;
+
     if (Left.is(tok::kw_operator))
       return Right.is(tok::coloncolon);
     if (Right.is(tok::l_brace) && Right.is(BK_BracedInit) &&
diff --git a/clang/lib/Format/UnwrappedLineFormatter.cpp b/clang/lib/Format/UnwrappedLineFormatter.cpp
index 299536cd806e..d099cfee9dea 100644
--- a/clang/lib/Format/UnwrappedLineFormatter.cpp
+++ b/clang/lib/Format/UnwrappedLineFormatter.cpp
@@ -320,9 +320,9 @@ private:
     }
     // Try to merge a control statement block with left brace wrapped
     if (I[1]->First->is(tok::l_brace) &&
-        (TheLine->First->isOneOf(tok::kw_if, tok::kw_while, tok::kw_for,
-                                 tok::kw_switch, tok::kw_try, tok::kw_do,
-                                 TT_ForEachMacro) ||
+        (TheLine->First->isOneOf(tok::kw_if, tok::kw_else, tok::kw_while,
+                                 tok::kw_for, tok::kw_switch, tok::kw_try,
+                                 tok::kw_do, TT_ForEachMacro) ||
          (TheLine->First->is(tok::r_brace) && TheLine->First->Next &&
           TheLine->First->Next->isOneOf(tok::kw_else, tok::kw_catch))) &&
         Style.BraceWrapping.AfterControlStatement ==
@@ -335,7 +335,7 @@ private:
                  ? 1
                  : 0;
     } else if (I[1]->First->is(tok::l_brace) &&
-               TheLine->First->isOneOf(tok::kw_if, tok::kw_while,
+               TheLine->First->isOneOf(tok::kw_if, tok::kw_else, tok::kw_while,
                                        tok::kw_for)) {
       return (Style.BraceWrapping.AfterControlStatement ==
               FormatStyle::BWACS_Always)
@@ -569,7 +569,7 @@ private:
 
     // Check that the current line allows merging. This depends on whether we
     // are in a control flow statements as well as several style flags.
-    if (Line.First->isOneOf(tok::kw_else, tok::kw_case) ||
+    if (Line.First->is(tok::kw_case) ||
         (Line.First->Next && Line.First->Next->is(tok::kw_else)))
       return 0;
     // default: in switch statement
@@ -578,20 +578,21 @@ private:
       if (Tok && Tok->is(tok::colon))
         return 0;
     }
-    if (Line.First->isOneOf(tok::kw_if, tok::kw_while, tok::kw_do, tok::kw_try,
-                            tok::kw___try, tok::kw_catch, tok::kw___finally,
-                            tok::kw_for, tok::r_brace, Keywords.kw___except)) {
+    if (Line.First->isOneOf(tok::kw_if, tok::kw_else, tok::kw_while, tok::kw_do,
+                            tok::kw_try, tok::kw___try, tok::kw_catch,
+                            tok::kw___finally, tok::kw_for, tok::r_brace,
+                            Keywords.kw___except)) {
       if (Style.AllowShortBlocksOnASingleLine == FormatStyle::SBS_Never)
         return 0;
       // Don't merge when we can't except the case when
       // the control statement block is empty
       if (!Style.AllowShortIfStatementsOnASingleLine &&
-          Line.startsWith(tok::kw_if) &&
+          Line.First->isOneOf(tok::kw_if, tok::kw_else) &&
           !Style.BraceWrapping.AfterControlStatement &&
           !I[1]->First->is(tok::r_brace))
         return 0;
       if (!Style.AllowShortIfStatementsOnASingleLine &&
-          Line.startsWith(tok::kw_if) &&
+          Line.First->isOneOf(tok::kw_if, tok::kw_else) &&
           Style.BraceWrapping.AfterControlStatement ==
               FormatStyle::BWACS_Always &&
           I + 2 != E && !I[2]->First->is(tok::r_brace))
@@ -676,7 +677,7 @@ private:
         // { <-- current Line
         //   baz();
         // }
-        if (Line.First == Line.Last &&
+        if (Line.First == Line.Last && Line.First->isNot(TT_FunctionLBrace) &&
             Style.BraceWrapping.AfterControlStatement ==
                 FormatStyle::BWACS_MultiLine)
           return 0;
diff --git a/clang/lib/Format/UnwrappedLineParser.cpp b/clang/lib/Format/UnwrappedLineParser.cpp
index 28d925858f77..5b9fe267aae6 100644
--- a/clang/lib/Format/UnwrappedLineParser.cpp
+++ b/clang/lib/Format/UnwrappedLineParser.cpp
@@ -28,9 +28,28 @@ namespace format {
 class FormatTokenSource {
 public:
   virtual ~FormatTokenSource() {}
+
+  // Returns the next token in the token stream.
   virtual FormatToken *getNextToken() = 0;
 
+  // Returns the token precedint the token returned by the last call to
+  // getNextToken() in the token stream, or nullptr if no such token exists.
+  virtual FormatToken *getPreviousToken() = 0;
+
+  // Returns the token that would be returned by the next call to
+  // getNextToken().
+  virtual FormatToken *peekNextToken() = 0;
+
+  // Returns whether we are at the end of the file.
+  // This can be different from whether getNextToken() returned an eof token
+  // when the FormatTokenSource is a view on a part of the token stream.
+  virtual bool isEOF() = 0;
+
+  // Gets the current position in the token stream, to be used by setPosition().
   virtual unsigned getPosition() = 0;
+
+  // Resets the token stream to the state it was in when getPosition() returned
+  // Position, and return the token at that position in the stream.
   virtual FormatToken *setPosition(unsigned Position) = 0;
 };
 
@@ -108,6 +127,18 @@ public:
     return Token;
   }
 
+  FormatToken *getPreviousToken() override {
+    return PreviousTokenSource->getPreviousToken();
+  }
+
+  FormatToken *peekNextToken() override {
+    if (eof())
+      return &FakeEOF;
+    return PreviousTokenSource->peekNextToken();
+  }
+
+  bool isEOF() override { return PreviousTokenSource->isEOF(); }
+
   unsigned getPosition() override { return PreviousTokenSource->getPosition(); }
 
   FormatToken *setPosition(unsigned Position) override {
@@ -199,16 +230,45 @@ public:
       : Tokens(Tokens), Position(-1) {}
 
   FormatToken *getNextToken() override {
+    if (Position >= 0 && Tokens[Position]->is(tok::eof)) {
+      LLVM_DEBUG({
+        llvm::dbgs() << "Next ";
+        dbgToken(Position);
+      });
+      return Tokens[Position];
+    }
     ++Position;
+    LLVM_DEBUG({
+      llvm::dbgs() << "Next ";
+      dbgToken(Position);
+    });
     return Tokens[Position];
   }
 
+  FormatToken *getPreviousToken() override {
+    assert(Position > 0);
+    return Tokens[Position - 1];
+  }
+
+  FormatToken *peekNextToken() override {
+    int Next = Position + 1;
+    LLVM_DEBUG({
+      llvm::dbgs() << "Peeking ";
+      dbgToken(Next);
+    });
+    return Tokens[Next];
+  }
+
+  bool isEOF() override { return Tokens[Position]->is(tok::eof); }
+
   unsigned getPosition() override {
+    LLVM_DEBUG(llvm::dbgs() << "Getting Position: " << Position << "\n");
     assert(Position >= 0);
     return Position;
   }
 
   FormatToken *setPosition(unsigned P) override {
+    LLVM_DEBUG(llvm::dbgs() << "Setting Position: " << P << "\n");
     Position = P;
     return Tokens[Position];
   }
@@ -216,6 +276,13 @@ public:
   void reset() { Position = -1; }
 
 private:
+  void dbgToken(int Position, llvm::StringRef Indent = "") {
+    FormatToken *Tok = Tokens[Position];
+    llvm::dbgs() << Indent << "[" << Position
+                 << "] Token: " << Tok->Tok.getName() << " / " << Tok->TokenText
+                 << ", Macro: " << !!Tok->MacroCtx << "\n";
+  }
+
   ArrayRef<FormatToken *> Tokens;
   int Position;
 };
@@ -399,7 +466,7 @@ void UnwrappedLineParser::parseLevel(bool HasOpeningBrace) {
       FormatToken *Next;
       do {
         Next = Tokens->getNextToken();
-      } while (Next && Next->is(tok::comment));
+      } while (Next->is(tok::comment));
       FormatTok = Tokens->setPosition(StoredPosition);
       if (Next && Next->isNot(tok::colon)) {
         // default not followed by ':' is not a case label; treat it like
@@ -875,10 +942,7 @@ void UnwrappedLineParser::parsePPEndIf() {
   parsePPUnknown();
   // If the #endif of a potential include guard is the last thing in the file,
   // then we found an include guard.
-  unsigned TokenPosition = Tokens->getPosition();
-  FormatToken *PeekNext = AllTokens[TokenPosition];
-  if (IncludeGuard == IG_Defined && PPBranchLevel == -1 &&
-      PeekNext->is(tok::eof) &&
+  if (IncludeGuard == IG_Defined && PPBranchLevel == -1 && Tokens->isEOF() &&
       Style.IndentPPDirectives != FormatStyle::PPDIS_None)
     IncludeGuard = IG_Found;
 }
@@ -1050,6 +1114,35 @@ static bool isC78ParameterDecl(const FormatToken *Tok, const FormatToken *Next,
   return Tok->Previous && Tok->Previous->isOneOf(tok::l_paren, tok::comma);
 }
 
+void UnwrappedLineParser::parseModuleImport() {
+  nextToken();
+  while (!eof()) {
+    if (FormatTok->is(tok::colon)) {
+      FormatTok->setType(TT_ModulePartitionColon);
+    }
+    // Handle import <foo/bar.h> as we would an include statement.
+    else if (FormatTok->is(tok::less)) {
+      nextToken();
+      while (!FormatTok->isOneOf(tok::semi, tok::greater, tok::eof)) {
+        // Mark tokens up to the trailing line comments as implicit string
+        // literals.
+        if (FormatTok->isNot(tok::comment) &&
+            !FormatTok->TokenText.startswith("//"))
+          FormatTok->setType(TT_ImplicitStringLiteral);
+        nextToken();
+      }
+    }
+    if (FormatTok->is(tok::semi)) {
+      nextToken();
+      break;
+    }
+    nextToken();
+  }
+
+  addUnwrappedLine();
+  return;
+}
+
 // readTokenWithJavaScriptASI reads the next token and terminates the current
 // line if JavaScript Automatic Semicolon Insertion must
 // happen between the current token and the next token.
@@ -1097,7 +1190,6 @@ void UnwrappedLineParser::readTokenWithJavaScriptASI() {
 }
 
 void UnwrappedLineParser::parseStructuralElement(bool IsTopLevel) {
-  assert(!FormatTok->is(tok::l_brace));
   if (Style.Language == FormatStyle::LK_TableGen &&
       FormatTok->is(tok::pp_include)) {
     nextToken();
@@ -1249,6 +1341,10 @@ void UnwrappedLineParser::parseStructuralElement(bool IsTopLevel) {
         addUnwrappedLine();
         return;
       }
+      if (Style.isCpp()) {
+        parseModuleImport();
+        return;
+      }
     }
     if (Style.isCpp() &&
         FormatTok->isOneOf(Keywords.kw_signals, Keywords.kw_qsignals,
@@ -1402,9 +1498,7 @@ void UnwrappedLineParser::parseStructuralElement(bool IsTopLevel) {
       // declaration.
       if (!IsTopLevel || !Style.isCpp() || !Previous || FormatTok->is(tok::eof))
         break;
-      const unsigned Position = Tokens->getPosition() + 1;
-      assert(Position < AllTokens.size());
-      if (isC78ParameterDecl(FormatTok, AllTokens[Position], Previous)) {
+      if (isC78ParameterDecl(FormatTok, Tokens->peekNextToken(), Previous)) {
         addUnwrappedLine();
         return;
       }
@@ -1488,7 +1582,7 @@ void UnwrappedLineParser::parseStructuralElement(bool IsTopLevel) {
           unsigned StoredPosition = Tokens->getPosition();
           FormatToken *Next = Tokens->getNextToken();
           FormatTok = Tokens->setPosition(StoredPosition);
-          if (Next && !mustBeJSIdent(Keywords, Next)) {
+          if (!mustBeJSIdent(Keywords, Next)) {
             nextToken();
             break;
           }
@@ -2099,8 +2193,8 @@ void UnwrappedLineParser::parseIfThenElse() {
       parseBlock();
       addUnwrappedLine();
     } else if (FormatTok->Tok.is(tok::kw_if)) {
-      FormatToken *Previous = AllTokens[Tokens->getPosition() - 1];
-      bool PrecededByComment = Previous->is(tok::comment);
+      FormatToken *Previous = Tokens->getPreviousToken();
+      bool PrecededByComment = Previous && Previous->is(tok::comment);
       if (PrecededByComment) {
         addUnwrappedLine();
         ++Line->Level;
@@ -2305,6 +2399,8 @@ void UnwrappedLineParser::parseForOrWhileLoop() {
   if (Style.Language == FormatStyle::LK_JavaScript &&
       FormatTok->is(Keywords.kw_await))
     nextToken();
+  if (Style.isCpp() && FormatTok->is(tok::kw_co_await))
+    nextToken();
   if (FormatTok->Tok.is(tok::l_paren))
     parseParens();
   if (FormatTok->Tok.is(tok::l_brace)) {
@@ -2653,23 +2749,25 @@ bool UnwrappedLineParser::tryToParseSimpleAttribute() {
   ScopedTokenPosition AutoPosition(Tokens);
   FormatToken *Tok = Tokens->getNextToken();
   // We already read the first [ check for the second.
-  if (Tok && !Tok->is(tok::l_square)) {
+  if (!Tok->is(tok::l_square)) {
     return false;
   }
   // Double check that the attribute is just something
   // fairly simple.
-  while (Tok) {
+  while (Tok->isNot(tok::eof)) {
     if (Tok->is(tok::r_square)) {
       break;
     }
     Tok = Tokens->getNextToken();
   }
+  if (Tok->is(tok::eof))
+    return false;
   Tok = Tokens->getNextToken();
-  if (Tok && !Tok->is(tok::r_square)) {
+  if (!Tok->is(tok::r_square)) {
     return false;
   }
   Tok = Tokens->getNextToken();
-  if (Tok && Tok->is(tok::semi)) {
+  if (Tok->is(tok::semi)) {
     return false;
   }
   return true;
@@ -2682,7 +2780,7 @@ void UnwrappedLineParser::parseJavaEnumBody() {
   unsigned StoredPosition = Tokens->getPosition();
   bool IsSimple = true;
   FormatToken *Tok = Tokens->getNextToken();
-  while (Tok) {
+  while (!Tok->is(tok::eof)) {
     if (Tok->is(tok::r_brace))
       break;
     if (Tok->isOneOf(tok::l_brace, tok::semi)) {
@@ -3292,6 +3390,20 @@ void UnwrappedLineParser::readToken(int LevelDifference) {
   do {
     FormatTok = Tokens->getNextToken();
     assert(FormatTok);
+    while (FormatTok->getType() == TT_ConflictStart ||
+           FormatTok->getType() == TT_ConflictEnd ||
+           FormatTok->getType() == TT_ConflictAlternative) {
+      if (FormatTok->getType() == TT_ConflictStart) {
+        conditionalCompilationStart(/*Unreachable=*/false);
+      } else if (FormatTok->getType() == TT_ConflictAlternative) {
+        conditionalCompilationAlternative();
+      } else if (FormatTok->getType() == TT_ConflictEnd) {
+        conditionalCompilationEnd();
+      }
+      FormatTok = Tokens->getNextToken();
+      FormatTok->MustBreakBefore = true;
+    }
+
     while (!Line->InPPDirective && FormatTok->Tok.is(tok::hash) &&
            (FormatTok->HasUnescapedNewline || FormatTok->IsFirst)) {
       distributeComments(Comments, FormatTok);
@@ -3313,19 +3425,6 @@ void UnwrappedLineParser::readToken(int LevelDifference) {
       flushComments(isOnNewLine(*FormatTok));
       parsePPDirective();
     }
-    while (FormatTok->getType() == TT_ConflictStart ||
-           FormatTok->getType() == TT_ConflictEnd ||
-           FormatTok->getType() == TT_ConflictAlternative) {
-      if (FormatTok->getType() == TT_ConflictStart) {
-        conditionalCompilationStart(/*Unreachable=*/false);
-      } else if (FormatTok->getType() == TT_ConflictAlternative) {
-        conditionalCompilationAlternative();
-      } else if (FormatTok->getType() == TT_ConflictEnd) {
-        conditionalCompilationEnd();
-      }
-      FormatTok = Tokens->getNextToken();
-      FormatTok->MustBreakBefore = true;
-    }
 
     if (!PPStack.empty() && (PPStack.back().Kind == PP_Unreachable) &&
         !Line->InPPDirective) {
diff --git a/clang/lib/Format/UnwrappedLineParser.h b/clang/lib/Format/UnwrappedLineParser.h
index bcae0f3ad258..b4c082654597 100644
--- a/clang/lib/Format/UnwrappedLineParser.h
+++ b/clang/lib/Format/UnwrappedLineParser.h
@@ -110,6 +110,7 @@ private:
   void parseCaseLabel();
   void parseSwitch();
   void parseNamespace();
+  void parseModuleImport();
   void parseNew();
   void parseAccessSpecifier();
   bool parseEnum();
diff --git a/clang/lib/Format/WhitespaceManager.cpp b/clang/lib/Format/WhitespaceManager.cpp
index 74136d2f5caa..fae8a1c3fdc6 100644
--- a/clang/lib/Format/WhitespaceManager.cpp
+++ b/clang/lib/Format/WhitespaceManager.cpp
@@ -372,8 +372,6 @@ AlignTokenSequence(const FormatStyle &Style, unsigned Start, unsigned End,
     if (ContinuedStringLiteral)
       Changes[i].Spaces += Shift;
 
-    assert(Shift >= 0);
-
     Changes[i].StartOfTokenColumn += Shift;
     if (i + 1 != Changes.size())
       Changes[i + 1].PreviousEndOfTokenColumn += Shift;
@@ -915,7 +913,8 @@ void WhitespaceManager::alignTrailingComments(unsigned Start, unsigned End,
               Changes[i].StartOfBlockComment->StartOfTokenColumn -
               Changes[i].StartOfTokenColumn;
     }
-    assert(Shift >= 0);
+    if (Shift < 0)
+      continue;
     Changes[i].Spaces += Shift;
     if (i + 1 != Changes.size())
       Changes[i + 1].PreviousEndOfTokenColumn += Shift;
@@ -1270,10 +1269,10 @@ WhitespaceManager::linkCells(CellDescriptions &&CellDesc) {
 void WhitespaceManager::generateChanges() {
   for (unsigned i = 0, e = Changes.size(); i != e; ++i) {
     const Change &C = Changes[i];
-    if (i > 0) {
-      assert(Changes[i - 1].OriginalWhitespaceRange.getBegin() !=
-                 C.OriginalWhitespaceRange.getBegin() &&
-             "Generating two replacements for the same location");
+    if (i > 0 && Changes[i - 1].OriginalWhitespaceRange.getBegin() ==
+                     C.OriginalWhitespaceRange.getBegin()) {
+      // Do not generate two replacements for the same location.
+      continue;
     }
     if (C.CreateReplacement) {
       std::string ReplacementText = C.PreviousLinePostfix;
diff --git a/clang/lib/Frontend/InitPreprocessor.cpp b/clang/lib/Frontend/InitPreprocessor.cpp
index 0ecb024fc6b9..0c153446142e 100644
--- a/clang/lib/Frontend/InitPreprocessor.cpp
+++ b/clang/lib/Frontend/InitPreprocessor.cpp
@@ -505,6 +505,11 @@ static void InitializeStandardPredefinedMacros(const TargetInfo &TI,
   if (LangOpts.HIP) {
     Builder.defineMacro("__HIP__");
     Builder.defineMacro("__HIPCC__");
+    Builder.defineMacro("__HIP_MEMORY_SCOPE_SINGLETHREAD", "1");
+    Builder.defineMacro("__HIP_MEMORY_SCOPE_WAVEFRONT", "2");
+    Builder.defineMacro("__HIP_MEMORY_SCOPE_WORKGROUP", "3");
+    Builder.defineMacro("__HIP_MEMORY_SCOPE_AGENT", "4");
+    Builder.defineMacro("__HIP_MEMORY_SCOPE_SYSTEM", "5");
     if (LangOpts.CUDAIsDevice)
       Builder.defineMacro("__HIP_DEVICE_COMPILE__");
   }
diff --git a/clang/lib/Frontend/PrecompiledPreamble.cpp b/clang/lib/Frontend/PrecompiledPreamble.cpp
index af82ab3f5558..8aa80a4c96fb 100644
--- a/clang/lib/Frontend/PrecompiledPreamble.cpp
+++ b/clang/lib/Frontend/PrecompiledPreamble.cpp
@@ -412,10 +412,13 @@ llvm::ErrorOr<PrecompiledPreamble> PrecompiledPreamble::Build(
   std::unique_ptr<PrecompilePreambleAction> Act;
   Act.reset(new PrecompilePreambleAction(
       StoreInMemory ? &Storage.asMemory().Data : nullptr, Callbacks));
-  Callbacks.BeforeExecute(*Clang);
   if (!Act->BeginSourceFile(*Clang.get(), Clang->getFrontendOpts().Inputs[0]))
     return BuildPreambleError::BeginSourceFileFailed;
 
+  // Performed after BeginSourceFile to ensure Clang->Preprocessor can be
+  // referenced in the callback.
+  Callbacks.BeforeExecute(*Clang);
+
   std::unique_ptr<PPCallbacks> DelegatedPPCallbacks =
       Callbacks.createPPCallbacks();
   if (DelegatedPPCallbacks)
diff --git a/clang/lib/Frontend/Rewrite/RewriteModernObjC.cpp b/clang/lib/Frontend/Rewrite/RewriteModernObjC.cpp
index 626ec4d71ccd..b4487f004715 100644
--- a/clang/lib/Frontend/Rewrite/RewriteModernObjC.cpp
+++ b/clang/lib/Frontend/Rewrite/RewriteModernObjC.cpp
@@ -633,7 +633,7 @@ static bool IsHeaderFile(const std::string &Filename) {
     return false;
   }
 
-  std::string Ext = std::string(Filename.begin()+DotPos+1, Filename.end());
+  std::string Ext = Filename.substr(DotPos + 1);
   // C header: .h
   // C++ header: .hh or .H;
   return Ext == "h" || Ext == "hh" || Ext == "H";
diff --git a/clang/lib/Frontend/Rewrite/RewriteObjC.cpp b/clang/lib/Frontend/Rewrite/RewriteObjC.cpp
index 0750d36b02ac..b2ecb42c43dd 100644
--- a/clang/lib/Frontend/Rewrite/RewriteObjC.cpp
+++ b/clang/lib/Frontend/Rewrite/RewriteObjC.cpp
@@ -569,7 +569,7 @@ static bool IsHeaderFile(const std::string &Filename) {
     return false;
   }
 
-  std::string Ext = std::string(Filename.begin()+DotPos+1, Filename.end());
+  std::string Ext = Filename.substr(DotPos + 1);
   // C header: .h
   // C++ header: .hh or .H;
   return Ext == "h" || Ext == "hh" || Ext == "H";
diff --git a/clang/lib/Headers/altivec.h b/clang/lib/Headers/altivec.h
index fb808d7b0a4f..55195b0781fb 100644
--- a/clang/lib/Headers/altivec.h
+++ b/clang/lib/Headers/altivec.h
@@ -19,6 +19,10 @@
 #define __CR6_EQ_REV 1
 #define __CR6_LT 2
 #define __CR6_LT_REV 3
+#define __CR6_GT 4
+#define __CR6_GT_REV 5
+#define __CR6_SO 6
+#define __CR6_SO_REV 7
 
 /* Constants for vec_test_data_class */
 #define __VEC_CLASS_FP_SUBNORMAL_N (1 << 0)
@@ -8413,9 +8417,20 @@ static __inline__ vector float __ATTRS_o_ai vec_round(vector float __a) {
 }
 
 #ifdef __VSX__
+#ifdef __XL_COMPAT_ALTIVEC__
+static __inline__ vector double __ATTRS_o_ai vec_rint(vector double __a);
+static __inline__ vector double __ATTRS_o_ai vec_round(vector double __a) {
+  double __fpscr = __builtin_readflm();
+  __builtin_setrnd(0);
+  vector double __rounded = vec_rint(__a);
+  __builtin_setflm(__fpscr);
+  return __rounded;
+}
+#else
 static __inline__ vector double __ATTRS_o_ai vec_round(vector double __a) {
   return __builtin_vsx_xvrdpi(__a);
 }
+#endif
 
 /* vec_rint */
 
@@ -19026,6 +19041,51 @@ vec_sra(vector signed __int128 __a, vector unsigned __int128 __b) {
 #endif /* __SIZEOF_INT128__ */
 #endif /* __POWER10_VECTOR__ */
 
+#ifdef __POWER8_VECTOR__
+#define __bcdadd(__a, __b, __ps) __builtin_ppc_bcdadd((__a), (__b), (__ps))
+#define __bcdsub(__a, __b, __ps) __builtin_ppc_bcdsub((__a), (__b), (__ps))
+
+static __inline__ long __bcdadd_ofl(vector unsigned char __a,
+                                    vector unsigned char __b) {
+  return __builtin_ppc_bcdadd_p(__CR6_SO, __a, __b);
+}
+
+static __inline__ long __bcdsub_ofl(vector unsigned char __a,
+                                    vector unsigned char __b) {
+  return __builtin_ppc_bcdsub_p(__CR6_SO, __a, __b);
+}
+
+static __inline__ long __bcd_invalid(vector unsigned char __a) {
+  return __builtin_ppc_bcdsub_p(__CR6_SO, __a, __a);
+}
+
+static __inline__ long __bcdcmpeq(vector unsigned char __a,
+                                  vector unsigned char __b) {
+  return __builtin_ppc_bcdsub_p(__CR6_EQ, __a, __b);
+}
+
+static __inline__ long __bcdcmplt(vector unsigned char __a,
+                                  vector unsigned char __b) {
+  return __builtin_ppc_bcdsub_p(__CR6_LT, __a, __b);
+}
+
+static __inline__ long __bcdcmpgt(vector unsigned char __a,
+                                  vector unsigned char __b) {
+  return __builtin_ppc_bcdsub_p(__CR6_GT, __a, __b);
+}
+
+static __inline__ long __bcdcmple(vector unsigned char __a,
+                                  vector unsigned char __b) {
+  return __builtin_ppc_bcdsub_p(__CR6_GT_REV, __a, __b);
+}
+
+static __inline__ long __bcdcmpge(vector unsigned char __a,
+                                  vector unsigned char __b) {
+  return __builtin_ppc_bcdsub_p(__CR6_LT_REV, __a, __b);
+}
+
+#endif // __POWER8_VECTOR__
+
 #undef __ATTRS_o_ai
 
 #endif /* __ALTIVEC_H */
diff --git a/clang/lib/Headers/ppc_wrappers/emmintrin.h b/clang/lib/Headers/ppc_wrappers/emmintrin.h
index 4dcb8485e2e9..82a71788b27a 100644
--- a/clang/lib/Headers/ppc_wrappers/emmintrin.h
+++ b/clang/lib/Headers/ppc_wrappers/emmintrin.h
@@ -35,7 +35,7 @@
 #ifndef EMMINTRIN_H_
 #define EMMINTRIN_H_
 
-#if defined(__linux__) && defined(__ppc64__)
+#if defined(__ppc64__) && (defined(__linux__) || defined(__FreeBSD__))
 
 #include <altivec.h>
 
@@ -2319,6 +2319,7 @@ _mm_castsi128_pd(__m128i __A)
 
 #else
 #include_next <emmintrin.h>
-#endif /* defined(__linux__) && defined(__ppc64__) */
+#endif /* defined(__ppc64__) && (defined(__linux__) || defined(__FreeBSD__))   \
+        */
 
 #endif /* EMMINTRIN_H_ */
diff --git a/clang/lib/Headers/ppc_wrappers/mm_malloc.h b/clang/lib/Headers/ppc_wrappers/mm_malloc.h
index 24b14c8e07c0..86cf1a0f7618 100644
--- a/clang/lib/Headers/ppc_wrappers/mm_malloc.h
+++ b/clang/lib/Headers/ppc_wrappers/mm_malloc.h
@@ -10,7 +10,7 @@
 #ifndef _MM_MALLOC_H_INCLUDED
 #define _MM_MALLOC_H_INCLUDED
 
-#if defined(__linux__) && defined(__ppc64__)
+#if defined(__ppc64__) && (defined(__linux__) || defined(__FreeBSD__))
 
 #include <stdlib.h>
 
diff --git a/clang/lib/Headers/ppc_wrappers/mmintrin.h b/clang/lib/Headers/ppc_wrappers/mmintrin.h
index c55c44726f00..54e4ee9f4468 100644
--- a/clang/lib/Headers/ppc_wrappers/mmintrin.h
+++ b/clang/lib/Headers/ppc_wrappers/mmintrin.h
@@ -35,7 +35,7 @@
 #ifndef _MMINTRIN_H_INCLUDED
 #define _MMINTRIN_H_INCLUDED
 
-#if defined(__linux__) && defined(__ppc64__)
+#if defined(__ppc64__) && (defined(__linux__) || defined(__FreeBSD__))
 
 #include <altivec.h>
 /* The Intel API is flexible enough that we must allow aliasing with other
@@ -1445,6 +1445,7 @@ extern __inline __m64
 
 #else
 #include_next <mmintrin.h>
-#endif /* defined(__linux__) && defined(__ppc64__) */
+#endif /* defined(__ppc64__) && (defined(__linux__) || defined(__FreeBSD__))   \
+        */
 
 #endif /* _MMINTRIN_H_INCLUDED */
diff --git a/clang/lib/Headers/ppc_wrappers/pmmintrin.h b/clang/lib/Headers/ppc_wrappers/pmmintrin.h
index 6d93383d5412..8d4046bd43f1 100644
--- a/clang/lib/Headers/ppc_wrappers/pmmintrin.h
+++ b/clang/lib/Headers/ppc_wrappers/pmmintrin.h
@@ -38,7 +38,7 @@
 #ifndef PMMINTRIN_H_
 #define PMMINTRIN_H_
 
-#if defined(__linux__) && defined(__ppc64__)
+#if defined(__ppc64__) && (defined(__linux__) || defined(__FreeBSD__))
 
 /* We need definitions from the SSE2 and SSE header files*/
 #include <emmintrin.h>
@@ -145,6 +145,7 @@ _mm_lddqu_si128 (__m128i const *__P)
 
 #else
 #include_next <pmmintrin.h>
-#endif /* defined(__linux__) && defined(__ppc64__) */
+#endif /* defined(__ppc64__) && (defined(__linux__) || defined(__FreeBSD__))   \
+        */
 
 #endif /* PMMINTRIN_H_ */
diff --git a/clang/lib/Headers/ppc_wrappers/smmintrin.h b/clang/lib/Headers/ppc_wrappers/smmintrin.h
index f41264b27584..674703245a69 100644
--- a/clang/lib/Headers/ppc_wrappers/smmintrin.h
+++ b/clang/lib/Headers/ppc_wrappers/smmintrin.h
@@ -29,7 +29,7 @@
 #ifndef SMMINTRIN_H_
 #define SMMINTRIN_H_
 
-#if defined(__linux__) && defined(__ppc64__)
+#if defined(__ppc64__) && (defined(__linux__) || defined(__FreeBSD__))
 
 #include <altivec.h>
 #include <tmmintrin.h>
@@ -104,6 +104,7 @@ extern __inline __m128i
 
 #else
 #include_next <smmintrin.h>
-#endif /* defined(__linux__) && defined(__ppc64__) */
+#endif /* defined(__ppc64__) && (defined(__linux__) || defined(__FreeBSD__))   \
+        */
 
 #endif /* _SMMINTRIN_H_ */
diff --git a/clang/lib/Headers/ppc_wrappers/tmmintrin.h b/clang/lib/Headers/ppc_wrappers/tmmintrin.h
index b5a935d5e47e..ebef7b8192d7 100644
--- a/clang/lib/Headers/ppc_wrappers/tmmintrin.h
+++ b/clang/lib/Headers/ppc_wrappers/tmmintrin.h
@@ -25,7 +25,7 @@
 #ifndef TMMINTRIN_H_
 #define TMMINTRIN_H_
 
-#if defined(__linux__) && defined(__ppc64__)
+#if defined(__ppc64__) && (defined(__linux__) || defined(__FreeBSD__))
 
 #include <altivec.h>
 
@@ -490,6 +490,7 @@ _mm_mulhrs_pi16 (__m64 __A, __m64 __B)
 
 #else
 #include_next <tmmintrin.h>
-#endif /* defined(__linux__) && defined(__ppc64__) */
+#endif /* defined(__ppc64__) && (defined(__linux__) || defined(__FreeBSD__))   \
+        */
 
 #endif /* TMMINTRIN_H_ */
diff --git a/clang/lib/Headers/ppc_wrappers/xmmintrin.h b/clang/lib/Headers/ppc_wrappers/xmmintrin.h
index 0e45b96769f8..956603d36408 100644
--- a/clang/lib/Headers/ppc_wrappers/xmmintrin.h
+++ b/clang/lib/Headers/ppc_wrappers/xmmintrin.h
@@ -34,7 +34,7 @@
 #ifndef _XMMINTRIN_H_INCLUDED
 #define _XMMINTRIN_H_INCLUDED
 
-#if defined(__linux__) && defined(__ppc64__)
+#if defined(__ppc64__) && (defined(__linux__) || defined(__FreeBSD__))
 
 /* Define four value permute mask */
 #define _MM_SHUFFLE(w,x,y,z) (((w) << 6) | ((x) << 4) | ((y) << 2) | (z))
@@ -1838,6 +1838,7 @@ do {									\
 
 #else
 #include_next <xmmintrin.h>
-#endif /* defined(__linux__) && defined(__ppc64__) */
+#endif /* defined(__ppc64__) && (defined(__linux__) || defined(__FreeBSD__))   \
+        */
 
 #endif /* _XMMINTRIN_H_INCLUDED */
diff --git a/clang/lib/Headers/stdatomic.h b/clang/lib/Headers/stdatomic.h
index 665551ea69a4..1e47bcb2bacf 100644
--- a/clang/lib/Headers/stdatomic.h
+++ b/clang/lib/Headers/stdatomic.h
@@ -12,8 +12,12 @@
 
 /* If we're hosted, fall back to the system's stdatomic.h. FreeBSD, for
  * example, already has a Clang-compatible stdatomic.h header.
+ *
+ * Exclude the MSVC path as well as the MSVC header as of the 14.31.30818
+ * explicitly disallows `stdatomic.h` in the C mode via an `#error`.  Fallback
+ * to the clang resource header until that is fully supported.
  */
-#if __STDC_HOSTED__ && __has_include_next(<stdatomic.h>)
+#if __STDC_HOSTED__ && __has_include_next(<stdatomic.h>) && !defined(_MSC_VER)
 # include_next <stdatomic.h>
 #else
 
diff --git a/clang/lib/Parse/ParseDecl.cpp b/clang/lib/Parse/ParseDecl.cpp
index a0871062395e..1bdeccc4cbf5 100644
--- a/clang/lib/Parse/ParseDecl.cpp
+++ b/clang/lib/Parse/ParseDecl.cpp
@@ -6978,13 +6978,13 @@ void Parser::ParseParameterDeclarationClause(
       //
       // We care about case 1) where the declarator type should be known, and
       // the identifier should be null.
-      if (!ParmDeclarator.isInvalidType() && !ParmDeclarator.hasName()) {
-        if (Tok.getIdentifierInfo() &&
-            Tok.getIdentifierInfo()->isKeyword(getLangOpts())) {
-          Diag(Tok, diag::err_keyword_as_parameter) << PP.getSpelling(Tok);
-          // Consume the keyword.
-          ConsumeToken();
-        }
+      if (!ParmDeclarator.isInvalidType() && !ParmDeclarator.hasName() &&
+          Tok.isNot(tok::raw_identifier) && !Tok.isAnnotation() &&
+          Tok.getIdentifierInfo() &&
+          Tok.getIdentifierInfo()->isKeyword(getLangOpts())) {
+        Diag(Tok, diag::err_keyword_as_parameter) << PP.getSpelling(Tok);
+        // Consume the keyword.
+        ConsumeToken();
       }
       // Inform the actions module about the parameter declarator, so it gets
       // added to the current scope.
diff --git a/clang/lib/Parse/ParseStmt.cpp b/clang/lib/Parse/ParseStmt.cpp
index bb8718671bb0..292ab03e8614 100644
--- a/clang/lib/Parse/ParseStmt.cpp
+++ b/clang/lib/Parse/ParseStmt.cpp
@@ -2108,6 +2108,9 @@ StmtResult Parser::ParseForStatement(SourceLocation *TrailingElseLoc) {
     CoawaitLoc = SourceLocation();
   }
 
+  if (CoawaitLoc.isValid() && getLangOpts().CPlusPlus20)
+    Diag(CoawaitLoc, diag::warn_deprecated_for_co_await);
+
   // We need to perform most of the semantic analysis for a C++0x for-range
   // statememt before parsing the body, in order to be able to deduce the type
   // of an auto-typed loop variable.
diff --git a/clang/lib/Sema/AnalysisBasedWarnings.cpp b/clang/lib/Sema/AnalysisBasedWarnings.cpp
index 8544a4fccf4c..b4dcc9759b99 100644
--- a/clang/lib/Sema/AnalysisBasedWarnings.cpp
+++ b/clang/lib/Sema/AnalysisBasedWarnings.cpp
@@ -464,7 +464,7 @@ static ControlFlowKind CheckFallThrough(AnalysisDeclContext &AC) {
     // No more CFGElements in the block?
     if (ri == re) {
       const Stmt *Term = B.getTerminatorStmt();
-      if (Term && isa<CXXTryStmt>(Term)) {
+      if (Term && (isa<CXXTryStmt>(Term) || isa<ObjCAtTryStmt>(Term))) {
         HasAbnormalEdge = true;
         continue;
       }
diff --git a/clang/lib/Sema/SemaChecking.cpp b/clang/lib/Sema/SemaChecking.cpp
index a0f6702a5f82..33e2b3b5027d 100644
--- a/clang/lib/Sema/SemaChecking.cpp
+++ b/clang/lib/Sema/SemaChecking.cpp
@@ -5297,6 +5297,7 @@ static bool isValidOrderingForOp(int64_t Ordering, AtomicExpr::AtomicOp Op) {
 
   case AtomicExpr::AO__c11_atomic_load:
   case AtomicExpr::AO__opencl_atomic_load:
+  case AtomicExpr::AO__hip_atomic_load:
   case AtomicExpr::AO__atomic_load_n:
   case AtomicExpr::AO__atomic_load:
     return OrderingCABI != llvm::AtomicOrderingCABI::release &&
@@ -5304,6 +5305,7 @@ static bool isValidOrderingForOp(int64_t Ordering, AtomicExpr::AtomicOp Op) {
 
   case AtomicExpr::AO__c11_atomic_store:
   case AtomicExpr::AO__opencl_atomic_store:
+  case AtomicExpr::AO__hip_atomic_store:
   case AtomicExpr::AO__atomic_store:
   case AtomicExpr::AO__atomic_store_n:
     return OrderingCABI != llvm::AtomicOrderingCABI::consume &&
@@ -5380,6 +5382,8 @@ ExprResult Sema::BuildAtomicExpr(SourceRange CallRange, SourceRange ExprRange,
                 "need to update code for modified C11 atomics");
   bool IsOpenCL = Op >= AtomicExpr::AO__opencl_atomic_init &&
                   Op <= AtomicExpr::AO__opencl_atomic_fetch_max;
+  bool IsHIP = Op >= AtomicExpr::AO__hip_atomic_load &&
+               Op <= AtomicExpr::AO__hip_atomic_fetch_max;
   bool IsC11 = (Op >= AtomicExpr::AO__c11_atomic_init &&
                Op <= AtomicExpr::AO__c11_atomic_fetch_min) ||
                IsOpenCL;
@@ -5397,6 +5401,7 @@ ExprResult Sema::BuildAtomicExpr(SourceRange CallRange, SourceRange ExprRange,
 
   case AtomicExpr::AO__c11_atomic_load:
   case AtomicExpr::AO__opencl_atomic_load:
+  case AtomicExpr::AO__hip_atomic_load:
   case AtomicExpr::AO__atomic_load_n:
     Form = Load;
     break;
@@ -5407,11 +5412,14 @@ ExprResult Sema::BuildAtomicExpr(SourceRange CallRange, SourceRange ExprRange,
 
   case AtomicExpr::AO__c11_atomic_store:
   case AtomicExpr::AO__opencl_atomic_store:
+  case AtomicExpr::AO__hip_atomic_store:
   case AtomicExpr::AO__atomic_store:
   case AtomicExpr::AO__atomic_store_n:
     Form = Copy;
     break;
-
+  case AtomicExpr::AO__hip_atomic_fetch_add:
+  case AtomicExpr::AO__hip_atomic_fetch_min:
+  case AtomicExpr::AO__hip_atomic_fetch_max:
   case AtomicExpr::AO__c11_atomic_fetch_add:
   case AtomicExpr::AO__c11_atomic_fetch_sub:
   case AtomicExpr::AO__opencl_atomic_fetch_add:
@@ -5426,6 +5434,9 @@ ExprResult Sema::BuildAtomicExpr(SourceRange CallRange, SourceRange ExprRange,
   case AtomicExpr::AO__c11_atomic_fetch_and:
   case AtomicExpr::AO__c11_atomic_fetch_or:
   case AtomicExpr::AO__c11_atomic_fetch_xor:
+  case AtomicExpr::AO__hip_atomic_fetch_and:
+  case AtomicExpr::AO__hip_atomic_fetch_or:
+  case AtomicExpr::AO__hip_atomic_fetch_xor:
   case AtomicExpr::AO__c11_atomic_fetch_nand:
   case AtomicExpr::AO__opencl_atomic_fetch_and:
   case AtomicExpr::AO__opencl_atomic_fetch_or:
@@ -5452,6 +5463,7 @@ ExprResult Sema::BuildAtomicExpr(SourceRange CallRange, SourceRange ExprRange,
     break;
 
   case AtomicExpr::AO__c11_atomic_exchange:
+  case AtomicExpr::AO__hip_atomic_exchange:
   case AtomicExpr::AO__opencl_atomic_exchange:
   case AtomicExpr::AO__atomic_exchange_n:
     Form = Xchg;
@@ -5463,8 +5475,10 @@ ExprResult Sema::BuildAtomicExpr(SourceRange CallRange, SourceRange ExprRange,
 
   case AtomicExpr::AO__c11_atomic_compare_exchange_strong:
   case AtomicExpr::AO__c11_atomic_compare_exchange_weak:
+  case AtomicExpr::AO__hip_atomic_compare_exchange_strong:
   case AtomicExpr::AO__opencl_atomic_compare_exchange_strong:
   case AtomicExpr::AO__opencl_atomic_compare_exchange_weak:
+  case AtomicExpr::AO__hip_atomic_compare_exchange_weak:
     Form = C11CmpXchg;
     break;
 
@@ -5475,7 +5489,7 @@ ExprResult Sema::BuildAtomicExpr(SourceRange CallRange, SourceRange ExprRange,
   }
 
   unsigned AdjustedNumArgs = NumArgs[Form];
-  if (IsOpenCL && Op != AtomicExpr::AO__opencl_atomic_init)
+  if ((IsOpenCL || IsHIP) && Op != AtomicExpr::AO__opencl_atomic_init)
     ++AdjustedNumArgs;
   // Check we have the right number of arguments.
   if (Args.size() < AdjustedNumArgs) {
@@ -5532,8 +5546,8 @@ ExprResult Sema::BuildAtomicExpr(SourceRange CallRange, SourceRange ExprRange,
 
   // For an arithmetic operation, the implied arithmetic must be well-formed.
   if (Form == Arithmetic) {
-    // GCC does not enforce these rules for GNU atomics, but we do, because if
-    // we didn't it would be very confusing. FIXME:  For whom? How so?
+    // GCC does not enforce these rules for GNU atomics, but we do to help catch
+    // trivial type errors.
     auto IsAllowedValueType = [&](QualType ValType) {
       if (ValType->isIntegerType())
         return true;
@@ -5574,8 +5588,9 @@ ExprResult Sema::BuildAtomicExpr(SourceRange CallRange, SourceRange ExprRange,
   if (!IsC11 && !AtomTy.isTriviallyCopyableType(Context) &&
       !AtomTy->isScalarType()) {
     // For GNU atomics, require a trivially-copyable type. This is not part of
-    // the GNU atomics specification, but we enforce it, because if we didn't it
-    // would be very confusing. FIXME:  For whom? How so?
+    // the GNU atomics specification but we enforce it for consistency with
+    // other atomics which generally all require a trivially-copyable type. This
+    // is because atomics just copy bits.
     Diag(ExprRange.getBegin(), diag::err_atomic_op_needs_trivial_copy)
         << Ptr->getType() << Ptr->getSourceRange();
     return ExprError();
@@ -5614,7 +5629,7 @@ ExprResult Sema::BuildAtomicExpr(SourceRange CallRange, SourceRange ExprRange,
   // arguments are actually passed as pointers.
   QualType ByValType = ValType; // 'CP'
   bool IsPassedByAddress = false;
-  if (!IsC11 && !IsN) {
+  if (!IsC11 && !IsHIP && !IsN) {
     ByValType = Ptr->getType();
     IsPassedByAddress = true;
   }
@@ -5793,11 +5808,14 @@ ExprResult Sema::BuildAtomicExpr(SourceRange CallRange, SourceRange ExprRange,
   if ((Op == AtomicExpr::AO__c11_atomic_load ||
        Op == AtomicExpr::AO__c11_atomic_store ||
        Op == AtomicExpr::AO__opencl_atomic_load ||
-       Op == AtomicExpr::AO__opencl_atomic_store ) &&
+       Op == AtomicExpr::AO__hip_atomic_load ||
+       Op == AtomicExpr::AO__opencl_atomic_store ||
+       Op == AtomicExpr::AO__hip_atomic_store) &&
       Context.AtomicUsesUnsupportedLibcall(AE))
     Diag(AE->getBeginLoc(), diag::err_atomic_load_store_uses_lib)
         << ((Op == AtomicExpr::AO__c11_atomic_load ||
-             Op == AtomicExpr::AO__opencl_atomic_load)
+             Op == AtomicExpr::AO__opencl_atomic_load ||
+             Op == AtomicExpr::AO__hip_atomic_load)
                 ? 0
                 : 1);
 
diff --git a/clang/lib/Sema/SemaDecl.cpp b/clang/lib/Sema/SemaDecl.cpp
index af174ac1ca1a..7be71ca49ea2 100644
--- a/clang/lib/Sema/SemaDecl.cpp
+++ b/clang/lib/Sema/SemaDecl.cpp
@@ -10268,13 +10268,9 @@ static bool checkNonMultiVersionCompatAttributes(Sema &S,
                                                  const FunctionDecl *FD,
                                                  const FunctionDecl *CausedFD,
                                                  MultiVersionKind MVType) {
-  bool IsCPUSpecificCPUDispatchMVType =
-      MVType == MultiVersionKind::CPUDispatch ||
-      MVType == MultiVersionKind::CPUSpecific;
-  const auto Diagnose = [FD, CausedFD, IsCPUSpecificCPUDispatchMVType](
-                            Sema &S, const Attr *A) {
+  const auto Diagnose = [FD, CausedFD, MVType](Sema &S, const Attr *A) {
     S.Diag(FD->getLocation(), diag::err_multiversion_disallowed_other_attr)
-        << IsCPUSpecificCPUDispatchMVType << A;
+        << static_cast<unsigned>(MVType) << A;
     if (CausedFD)
       S.Diag(CausedFD->getLocation(), diag::note_multiversioning_caused_here);
     return true;
@@ -10292,6 +10288,10 @@ static bool checkNonMultiVersionCompatAttributes(Sema &S,
       if (MVType != MultiVersionKind::Target)
         return Diagnose(S, A);
       break;
+    case attr::TargetClones:
+      if (MVType != MultiVersionKind::TargetClones)
+        return Diagnose(S, A);
+      break;
     default:
       if (!AttrCompatibleWithMultiVersion(A->getKind(), MVType))
         return Diagnose(S, A);
@@ -10318,6 +10318,7 @@ bool Sema::areMultiversionVariantFunctionsCompatible(
     DefaultedFuncs = 6,
     ConstexprFuncs = 7,
     ConstevalFuncs = 8,
+    Lambda = 9,
   };
   enum Different {
     CallingConv = 0,
@@ -10445,7 +10446,7 @@ static bool CheckMultiVersionAdditionalRules(Sema &S, const FunctionDecl *OldFD,
                           S.PDiag(diag::note_multiversioning_caused_here)),
       PartialDiagnosticAt(NewFD->getLocation(),
                           S.PDiag(diag::err_multiversion_doesnt_support)
-                              << IsCPUSpecificCPUDispatchMVType),
+                              << static_cast<unsigned>(MVType)),
       PartialDiagnosticAt(NewFD->getLocation(),
                           S.PDiag(diag::err_multiversion_diff)),
       /*TemplatesSupported=*/false,
@@ -10574,21 +10575,30 @@ static bool CheckTargetCausesMultiVersioning(
   return false;
 }
 
+static bool MultiVersionTypesCompatible(MultiVersionKind Old,
+                                        MultiVersionKind New) {
+  if (Old == New || Old == MultiVersionKind::None ||
+      New == MultiVersionKind::None)
+    return true;
+
+  return (Old == MultiVersionKind::CPUDispatch &&
+          New == MultiVersionKind::CPUSpecific) ||
+         (Old == MultiVersionKind::CPUSpecific &&
+          New == MultiVersionKind::CPUDispatch);
+}
+
 /// Check the validity of a new function declaration being added to an existing
 /// multiversioned declaration collection.
 static bool CheckMultiVersionAdditionalDecl(
     Sema &S, FunctionDecl *OldFD, FunctionDecl *NewFD,
     MultiVersionKind NewMVType, const TargetAttr *NewTA,
     const CPUDispatchAttr *NewCPUDisp, const CPUSpecificAttr *NewCPUSpec,
-    bool &Redeclaration, NamedDecl *&OldDecl, bool &MergeTypeWithPrevious,
-    LookupResult &Previous) {
+    const TargetClonesAttr *NewClones, bool &Redeclaration, NamedDecl *&OldDecl,
+    bool &MergeTypeWithPrevious, LookupResult &Previous) {
 
   MultiVersionKind OldMVType = OldFD->getMultiVersionKind();
   // Disallow mixing of multiversioning types.
-  if ((OldMVType == MultiVersionKind::Target &&
-       NewMVType != MultiVersionKind::Target) ||
-      (NewMVType == MultiVersionKind::Target &&
-       OldMVType != MultiVersionKind::Target)) {
+  if (!MultiVersionTypesCompatible(OldMVType, NewMVType)) {
     S.Diag(NewFD->getLocation(), diag::err_multiversion_types_mixed);
     S.Diag(OldFD->getLocation(), diag::note_previous_declaration);
     NewFD->setInvalidDecl();
@@ -10613,7 +10623,12 @@ static bool CheckMultiVersionAdditionalDecl(
     if (S.IsOverload(NewFD, CurFD, UseMemberUsingDeclRules))
       continue;
 
-    if (NewMVType == MultiVersionKind::Target) {
+    switch (NewMVType) {
+    case MultiVersionKind::None:
+      assert(OldMVType == MultiVersionKind::TargetClones &&
+             "Only target_clones can be omitted in subsequent declarations");
+      break;
+    case MultiVersionKind::Target: {
       const auto *CurTA = CurFD->getAttr<TargetAttr>();
       if (CurTA->getFeaturesStr() == NewTA->getFeaturesStr()) {
         NewFD->setIsMultiVersion();
@@ -10629,7 +10644,30 @@ static bool CheckMultiVersionAdditionalDecl(
         NewFD->setInvalidDecl();
         return true;
       }
-    } else {
+      break;
+    }
+    case MultiVersionKind::TargetClones: {
+      const auto *CurClones = CurFD->getAttr<TargetClonesAttr>();
+      Redeclaration = true;
+      OldDecl = CurFD;
+      MergeTypeWithPrevious = true;
+      NewFD->setIsMultiVersion();
+
+      if (CurClones && NewClones &&
+          (CurClones->featuresStrs_size() != NewClones->featuresStrs_size() ||
+           !std::equal(CurClones->featuresStrs_begin(),
+                       CurClones->featuresStrs_end(),
+                       NewClones->featuresStrs_begin()))) {
+        S.Diag(NewFD->getLocation(), diag::err_target_clone_doesnt_match);
+        S.Diag(CurFD->getLocation(), diag::note_previous_declaration);
+        NewFD->setInvalidDecl();
+        return true;
+      }
+
+      return false;
+    }
+    case MultiVersionKind::CPUSpecific:
+    case MultiVersionKind::CPUDispatch: {
       const auto *CurCPUSpec = CurFD->getAttr<CPUSpecificAttr>();
       const auto *CurCPUDisp = CurFD->getAttr<CPUDispatchAttr>();
       // Handle CPUDispatch/CPUSpecific versions.
@@ -10684,8 +10722,8 @@ static bool CheckMultiVersionAdditionalDecl(
           }
         }
       }
-      // If the two decls aren't the same MVType, there is no possible error
-      // condition.
+      break;
+    }
     }
   }
 
@@ -10721,7 +10759,6 @@ static bool CheckMultiVersionAdditionalDecl(
   return false;
 }
 
-
 /// Check the validity of a mulitversion function declaration.
 /// Also sets the multiversion'ness' of the function itself.
 ///
@@ -10735,23 +10772,14 @@ static bool CheckMultiVersionFunction(Sema &S, FunctionDecl *NewFD,
   const auto *NewTA = NewFD->getAttr<TargetAttr>();
   const auto *NewCPUDisp = NewFD->getAttr<CPUDispatchAttr>();
   const auto *NewCPUSpec = NewFD->getAttr<CPUSpecificAttr>();
-
-  // Mixing Multiversioning types is prohibited.
-  if ((NewTA && NewCPUDisp) || (NewTA && NewCPUSpec) ||
-      (NewCPUDisp && NewCPUSpec)) {
-    S.Diag(NewFD->getLocation(), diag::err_multiversion_types_mixed);
-    NewFD->setInvalidDecl();
-    return true;
-  }
-
-  MultiVersionKind  MVType = NewFD->getMultiVersionKind();
+  const auto *NewClones = NewFD->getAttr<TargetClonesAttr>();
+  MultiVersionKind MVType = NewFD->getMultiVersionKind();
 
   // Main isn't allowed to become a multiversion function, however it IS
   // permitted to have 'main' be marked with the 'target' optimization hint.
   if (NewFD->isMain()) {
-    if ((MVType == MultiVersionKind::Target && NewTA->isDefaultVersion()) ||
-        MVType == MultiVersionKind::CPUDispatch ||
-        MVType == MultiVersionKind::CPUSpecific) {
+    if (MVType != MultiVersionKind::None &&
+        !(MVType == MultiVersionKind::Target && !NewTA->isDefaultVersion())) {
       S.Diag(NewFD->getLocation(), diag::err_multiversion_not_allowed_on_main);
       NewFD->setInvalidDecl();
       return true;
@@ -10774,13 +10802,35 @@ static bool CheckMultiVersionFunction(Sema &S, FunctionDecl *NewFD,
   if (!OldFD->isMultiVersion() && MVType == MultiVersionKind::None)
     return false;
 
-  if (OldFD->isMultiVersion() && MVType == MultiVersionKind::None) {
+  // Multiversioned redeclarations aren't allowed to omit the attribute, except
+  // for target_clones.
+  if (OldFD->isMultiVersion() && MVType == MultiVersionKind::None &&
+      OldFD->getMultiVersionKind() != MultiVersionKind::TargetClones) {
     S.Diag(NewFD->getLocation(), diag::err_multiversion_required_in_redecl)
         << (OldFD->getMultiVersionKind() != MultiVersionKind::Target);
     NewFD->setInvalidDecl();
     return true;
   }
 
+  if (!OldFD->isMultiVersion()) {
+    switch (MVType) {
+    case MultiVersionKind::Target:
+      return CheckTargetCausesMultiVersioning(S, OldFD, NewFD, NewTA,
+                                              Redeclaration, OldDecl,
+                                              MergeTypeWithPrevious, Previous);
+    case MultiVersionKind::TargetClones:
+      if (OldFD->isUsed(false)) {
+        NewFD->setInvalidDecl();
+        return S.Diag(NewFD->getLocation(), diag::err_multiversion_after_used);
+      }
+      OldFD->setIsMultiVersion();
+      break;
+    case MultiVersionKind::CPUDispatch:
+    case MultiVersionKind::CPUSpecific:
+    case MultiVersionKind::None:
+      break;
+    }
+  }
   // Handle the target potentially causes multiversioning case.
   if (!OldFD->isMultiVersion() && MVType == MultiVersionKind::Target)
     return CheckTargetCausesMultiVersioning(S, OldFD, NewFD, NewTA,
@@ -10791,8 +10841,8 @@ static bool CheckMultiVersionFunction(Sema &S, FunctionDecl *NewFD,
   // appropriate attribute in the current function decl.  Resolve that these are
   // still compatible with previous declarations.
   return CheckMultiVersionAdditionalDecl(
-      S, OldFD, NewFD, MVType, NewTA, NewCPUDisp, NewCPUSpec, Redeclaration,
-      OldDecl, MergeTypeWithPrevious, Previous);
+      S, OldFD, NewFD, MVType, NewTA, NewCPUDisp, NewCPUSpec, NewClones,
+      Redeclaration, OldDecl, MergeTypeWithPrevious, Previous);
 }
 
 /// Perform semantic checking of a new function declaration.
diff --git a/clang/lib/Sema/SemaDeclAttr.cpp b/clang/lib/Sema/SemaDeclAttr.cpp
index ef889a36bd55..4df8687aff89 100644
--- a/clang/lib/Sema/SemaDeclAttr.cpp
+++ b/clang/lib/Sema/SemaDeclAttr.cpp
@@ -1965,6 +1965,28 @@ static void handleRestrictAttr(Sema &S, Decl *D, const ParsedAttr &AL) {
 }
 
 static void handleCPUSpecificAttr(Sema &S, Decl *D, const ParsedAttr &AL) {
+  // Ensure we don't combine these with themselves, since that causes some
+  // confusing behavior.
+  if (AL.getParsedKind() == ParsedAttr::AT_CPUDispatch) {
+    if (checkAttrMutualExclusion<CPUSpecificAttr>(S, D, AL))
+      return;
+
+    if (const auto *Other = D->getAttr<CPUDispatchAttr>()) {
+      S.Diag(AL.getLoc(), diag::err_disallowed_duplicate_attribute) << AL;
+      S.Diag(Other->getLocation(), diag::note_conflicting_attribute);
+      return;
+    }
+  } else if (AL.getParsedKind() == ParsedAttr::AT_CPUSpecific) {
+    if (checkAttrMutualExclusion<CPUDispatchAttr>(S, D, AL))
+      return;
+
+    if (const auto *Other = D->getAttr<CPUSpecificAttr>()) {
+      S.Diag(AL.getLoc(), diag::err_disallowed_duplicate_attribute) << AL;
+      S.Diag(Other->getLocation(), diag::note_conflicting_attribute);
+      return;
+    }
+  }
+
   FunctionDecl *FD = cast<FunctionDecl>(D);
 
   if (const auto *MD = dyn_cast<CXXMethodDecl>(D)) {
@@ -3211,54 +3233,57 @@ static void handleCodeSegAttr(Sema &S, Decl *D, const ParsedAttr &AL) {
 bool Sema::checkTargetAttr(SourceLocation LiteralLoc, StringRef AttrStr) {
   enum FirstParam { Unsupported, Duplicate, Unknown };
   enum SecondParam { None, Architecture, Tune };
+  enum ThirdParam { Target, TargetClones };
   if (AttrStr.contains("fpmath="))
     return Diag(LiteralLoc, diag::warn_unsupported_target_attribute)
-           << Unsupported << None << "fpmath=";
+           << Unsupported << None << "fpmath=" << Target;
 
   // Diagnose use of tune if target doesn't support it.
   if (!Context.getTargetInfo().supportsTargetAttributeTune() &&
       AttrStr.contains("tune="))
     return Diag(LiteralLoc, diag::warn_unsupported_target_attribute)
-           << Unsupported << None << "tune=";
+           << Unsupported << None << "tune=" << Target;
 
   ParsedTargetAttr ParsedAttrs = TargetAttr::parse(AttrStr);
 
   if (!ParsedAttrs.Architecture.empty() &&
       !Context.getTargetInfo().isValidCPUName(ParsedAttrs.Architecture))
     return Diag(LiteralLoc, diag::warn_unsupported_target_attribute)
-           << Unknown << Architecture << ParsedAttrs.Architecture;
+           << Unknown << Architecture << ParsedAttrs.Architecture << Target;
 
   if (!ParsedAttrs.Tune.empty() &&
       !Context.getTargetInfo().isValidCPUName(ParsedAttrs.Tune))
     return Diag(LiteralLoc, diag::warn_unsupported_target_attribute)
-           << Unknown << Tune << ParsedAttrs.Tune;
+           << Unknown << Tune << ParsedAttrs.Tune << Target;
 
   if (ParsedAttrs.DuplicateArchitecture)
     return Diag(LiteralLoc, diag::warn_unsupported_target_attribute)
-           << Duplicate << None << "arch=";
+           << Duplicate << None << "arch=" << Target;
   if (ParsedAttrs.DuplicateTune)
     return Diag(LiteralLoc, diag::warn_unsupported_target_attribute)
-           << Duplicate << None << "tune=";
+           << Duplicate << None << "tune=" << Target;
 
   for (const auto &Feature : ParsedAttrs.Features) {
     auto CurFeature = StringRef(Feature).drop_front(); // remove + or -.
     if (!Context.getTargetInfo().isValidFeatureName(CurFeature))
       return Diag(LiteralLoc, diag::warn_unsupported_target_attribute)
-             << Unsupported << None << CurFeature;
+             << Unsupported << None << CurFeature << Target;
   }
 
   TargetInfo::BranchProtectionInfo BPI;
-  StringRef Error;
-  if (!ParsedAttrs.BranchProtection.empty() &&
-      !Context.getTargetInfo().validateBranchProtection(
-          ParsedAttrs.BranchProtection, BPI, Error)) {
-    if (Error.empty())
+  StringRef DiagMsg;
+  if (ParsedAttrs.BranchProtection.empty())
+    return false;
+  if (!Context.getTargetInfo().validateBranchProtection(
+          ParsedAttrs.BranchProtection, BPI, DiagMsg)) {
+    if (DiagMsg.empty())
       return Diag(LiteralLoc, diag::warn_unsupported_target_attribute)
-             << Unsupported << None << "branch-protection";
-    else
-      return Diag(LiteralLoc, diag::err_invalid_branch_protection_spec)
-             << Error;
+             << Unsupported << None << "branch-protection" << Target;
+    return Diag(LiteralLoc, diag::err_invalid_branch_protection_spec)
+           << DiagMsg;
   }
+  if (!DiagMsg.empty())
+    Diag(LiteralLoc, diag::warn_unsupported_branch_protection_spec) << DiagMsg;
 
   return false;
 }
@@ -3274,6 +3299,107 @@ static void handleTargetAttr(Sema &S, Decl *D, const ParsedAttr &AL) {
   D->addAttr(NewAttr);
 }
 
+bool Sema::checkTargetClonesAttrString(SourceLocation LiteralLoc, StringRef Str,
+                                       const StringLiteral *Literal,
+                                       bool &HasDefault, bool &HasCommas,
+                                       SmallVectorImpl<StringRef> &Strings) {
+  enum FirstParam { Unsupported, Duplicate, Unknown };
+  enum SecondParam { None, Architecture, Tune };
+  enum ThirdParam { Target, TargetClones };
+  HasCommas = HasCommas || Str.contains(',');
+  // Warn on empty at the beginning of a string.
+  if (Str.size() == 0)
+    return Diag(LiteralLoc, diag::warn_unsupported_target_attribute)
+           << Unsupported << None << "" << TargetClones;
+
+  std::pair<StringRef, StringRef> Parts = {{}, Str};
+  while (!Parts.second.empty()) {
+    Parts = Parts.second.split(',');
+    StringRef Cur = Parts.first.trim();
+    SourceLocation CurLoc = Literal->getLocationOfByte(
+        Cur.data() - Literal->getString().data(), getSourceManager(),
+        getLangOpts(), Context.getTargetInfo());
+
+    bool DefaultIsDupe = false;
+    if (Cur.empty())
+      return Diag(CurLoc, diag::warn_unsupported_target_attribute)
+             << Unsupported << None << "" << TargetClones;
+
+    if (Cur.startswith("arch=")) {
+      if (!Context.getTargetInfo().isValidCPUName(
+              Cur.drop_front(sizeof("arch=") - 1)))
+        return Diag(CurLoc, diag::warn_unsupported_target_attribute)
+               << Unsupported << Architecture
+               << Cur.drop_front(sizeof("arch=") - 1) << TargetClones;
+    } else if (Cur == "default") {
+      DefaultIsDupe = HasDefault;
+      HasDefault = true;
+    } else if (!Context.getTargetInfo().isValidFeatureName(Cur))
+      return Diag(CurLoc, diag::warn_unsupported_target_attribute)
+             << Unsupported << None << Cur << TargetClones;
+
+    if (llvm::find(Strings, Cur) != Strings.end() || DefaultIsDupe)
+      Diag(CurLoc, diag::warn_target_clone_duplicate_options);
+    // Note: Add even if there are duplicates, since it changes name mangling.
+    Strings.push_back(Cur);
+  }
+
+  if (Str.rtrim().endswith(","))
+    return Diag(LiteralLoc, diag::warn_unsupported_target_attribute)
+           << Unsupported << None << "" << TargetClones;
+  return false;
+}
+
+static void handleTargetClonesAttr(Sema &S, Decl *D, const ParsedAttr &AL) {
+  // Ensure we don't combine these with themselves, since that causes some
+  // confusing behavior.
+  if (const auto *Other = D->getAttr<TargetClonesAttr>()) {
+    S.Diag(AL.getLoc(), diag::err_disallowed_duplicate_attribute) << AL;
+    S.Diag(Other->getLocation(), diag::note_conflicting_attribute);
+    return;
+  }
+  if (checkAttrMutualExclusion<TargetClonesAttr>(S, D, AL))
+    return;
+
+  SmallVector<StringRef, 2> Strings;
+  bool HasCommas = false, HasDefault = false;
+
+  for (unsigned I = 0, E = AL.getNumArgs(); I != E; ++I) {
+    StringRef CurStr;
+    SourceLocation LiteralLoc;
+    if (!S.checkStringLiteralArgumentAttr(AL, I, CurStr, &LiteralLoc) ||
+        S.checkTargetClonesAttrString(
+            LiteralLoc, CurStr,
+            cast<StringLiteral>(AL.getArgAsExpr(I)->IgnoreParenCasts()),
+            HasDefault, HasCommas, Strings))
+      return;
+  }
+
+  if (HasCommas && AL.getNumArgs() > 1)
+    S.Diag(AL.getLoc(), diag::warn_target_clone_mixed_values);
+
+  if (!HasDefault) {
+    S.Diag(AL.getLoc(), diag::err_target_clone_must_have_default);
+    return;
+  }
+
+  // FIXME: We could probably figure out how to get this to work for lambdas
+  // someday.
+  if (const auto *MD = dyn_cast<CXXMethodDecl>(D)) {
+    if (MD->getParent()->isLambda()) {
+      S.Diag(D->getLocation(), diag::err_multiversion_doesnt_support)
+          << static_cast<unsigned>(MultiVersionKind::TargetClones)
+          << /*Lambda*/ 9;
+      return;
+    }
+  }
+
+  cast<FunctionDecl>(D)->setIsMultiVersion();
+  TargetClonesAttr *NewAttr = ::new (S.Context)
+      TargetClonesAttr(S.Context, AL, Strings.data(), Strings.size());
+  D->addAttr(NewAttr);
+}
+
 static void handleMinVectorWidthAttr(Sema &S, Decl *D, const ParsedAttr &AL) {
   Expr *E = AL.getArgAsExpr(0);
   uint32_t VecWidth;
@@ -8217,6 +8343,9 @@ static void ProcessDeclAttribute(Sema &S, Scope *scope, Decl *D,
   case ParsedAttr::AT_Target:
     handleTargetAttr(S, D, AL);
     break;
+  case ParsedAttr::AT_TargetClones:
+    handleTargetClonesAttr(S, D, AL);
+    break;
   case ParsedAttr::AT_MinVectorWidth:
     handleMinVectorWidthAttr(S, D, AL);
     break;
diff --git a/clang/lib/Sema/SemaExpr.cpp b/clang/lib/Sema/SemaExpr.cpp
index 8592335e20d3..b305d4e5b92f 100644
--- a/clang/lib/Sema/SemaExpr.cpp
+++ b/clang/lib/Sema/SemaExpr.cpp
@@ -16566,6 +16566,17 @@ Sema::PushExpressionEvaluationContext(
     ExpressionEvaluationContextRecord::ExpressionKind ExprContext) {
   ExprEvalContexts.emplace_back(NewContext, ExprCleanupObjects.size(), Cleanup,
                                 LambdaContextDecl, ExprContext);
+
+  // Discarded statements and immediate contexts nested in other
+  // discarded statements or immediate context are themselves
+  // a discarded statement or an immediate context, respectively.
+  ExprEvalContexts.back().InDiscardedStatement =
+      ExprEvalContexts[ExprEvalContexts.size() - 2]
+          .isDiscardedStatementContext();
+  ExprEvalContexts.back().InImmediateFunctionContext =
+      ExprEvalContexts[ExprEvalContexts.size() - 2]
+          .isImmediateFunctionContext();
+
   Cleanup.reset();
   if (!MaybeODRUseExprs.empty())
     std::swap(MaybeODRUseExprs, ExprEvalContexts.back().SavedMaybeODRUseExprs);
@@ -18965,6 +18976,10 @@ bool Sema::DiagIfReachable(SourceLocation Loc, ArrayRef<const Stmt *> Stmts,
 /// during overload resolution or within sizeof/alignof/typeof/typeid.
 bool Sema::DiagRuntimeBehavior(SourceLocation Loc, ArrayRef<const Stmt*> Stmts,
                                const PartialDiagnostic &PD) {
+
+  if (ExprEvalContexts.back().isDiscardedStatementContext())
+    return false;
+
   switch (ExprEvalContexts.back().Context) {
   case ExpressionEvaluationContext::Unevaluated:
   case ExpressionEvaluationContext::UnevaluatedList:
diff --git a/clang/lib/Sema/SemaExprCXX.cpp b/clang/lib/Sema/SemaExprCXX.cpp
index 635252584562..d25f329f85e4 100644
--- a/clang/lib/Sema/SemaExprCXX.cpp
+++ b/clang/lib/Sema/SemaExprCXX.cpp
@@ -1508,8 +1508,9 @@ Sema::BuildCXXTypeConstructExpr(TypeSourceInfo *TInfo,
   }
 
   // Only construct objects with object types.
-  // There doesn't seem to be an explicit rule for this but functions are
-  // not objects, so they cannot take initializers.
+  // The standard doesn't explicitly forbid function types here, but that's an
+  // obvious oversight, as there's no way to dynamically construct a function
+  // in general.
   if (Ty->isFunctionType())
     return ExprError(Diag(TyBeginLoc, diag::err_init_for_function_type)
                        << Ty << FullRange);
diff --git a/clang/lib/Sema/SemaStmt.cpp b/clang/lib/Sema/SemaStmt.cpp
index 3c820829864d..1d90759f2406 100644
--- a/clang/lib/Sema/SemaStmt.cpp
+++ b/clang/lib/Sema/SemaStmt.cpp
@@ -3563,8 +3563,7 @@ StmtResult Sema::ActOnCapScopeReturnStmt(SourceLocation ReturnLoc,
   bool HasDeducedReturnType =
       CurLambda && hasDeducedReturnType(CurLambda->CallOperator);
 
-  if (ExprEvalContexts.back().Context ==
-          ExpressionEvaluationContext::DiscardedStatement &&
+  if (ExprEvalContexts.back().isDiscardedStatementContext() &&
       (HasDeducedReturnType || CurCap->HasImplicitReturnType)) {
     if (RetValExp) {
       ExprResult ER =
@@ -3880,8 +3879,7 @@ Sema::ActOnReturnStmt(SourceLocation ReturnLoc, Expr *RetValExp,
   if (RetVal.isInvalid())
     return StmtError();
   StmtResult R = BuildReturnStmt(ReturnLoc, RetVal.get());
-  if (R.isInvalid() || ExprEvalContexts.back().Context ==
-                           ExpressionEvaluationContext::DiscardedStatement)
+  if (R.isInvalid() || ExprEvalContexts.back().isDiscardedStatementContext())
     return R;
 
   if (VarDecl *VD =
@@ -3966,8 +3964,7 @@ StmtResult Sema::BuildReturnStmt(SourceLocation ReturnLoc, Expr *RetValExp) {
 
   // C++1z: discarded return statements are not considered when deducing a
   // return type.
-  if (ExprEvalContexts.back().Context ==
-          ExpressionEvaluationContext::DiscardedStatement &&
+  if (ExprEvalContexts.back().isDiscardedStatementContext() &&
       FnRetType->getContainedAutoType()) {
     if (RetValExp) {
       ExprResult ER =
diff --git a/clang/lib/StaticAnalyzer/Core/CoreEngine.cpp b/clang/lib/StaticAnalyzer/Core/CoreEngine.cpp
index bc939d252800..d57bab154b61 100644
--- a/clang/lib/StaticAnalyzer/Core/CoreEngine.cpp
+++ b/clang/lib/StaticAnalyzer/Core/CoreEngine.cpp
@@ -686,8 +686,8 @@ SwitchNodeBuilder::generateDefaultCaseNode(ProgramStateRef St,
   assert(Src->succ_rbegin() != Src->succ_rend());
   CFGBlock *DefaultBlock = *Src->succ_rbegin();
 
-  // Sanity check for default blocks that are unreachable and not caught
-  // by earlier stages.
+  // Basic correctness check for default blocks that are unreachable and not
+  // caught by earlier stages.
   if (!DefaultBlock)
     return nullptr;
 
diff --git a/clang/lib/StaticAnalyzer/Core/RangeConstraintManager.cpp b/clang/lib/StaticAnalyzer/Core/RangeConstraintManager.cpp
index 74403a160b8e..23c67c64f975 100644
--- a/clang/lib/StaticAnalyzer/Core/RangeConstraintManager.cpp
+++ b/clang/lib/StaticAnalyzer/Core/RangeConstraintManager.cpp
@@ -2191,6 +2191,42 @@ LLVM_NODISCARD ProgramStateRef reAssume(ProgramStateRef State,
                                      Constraint->getMaxValue(), true);
 }
 
+// Simplify the given symbol with the help of the SValBuilder. In
+// SValBuilder::symplifySval, we traverse the symbol tree and query the
+// constraint values for the sub-trees and if a value is a constant we do the
+// constant folding. Compound symbols might collapse to simpler symbol tree
+// that is still possible to further simplify. Thus, we do the simplification on
+// a new symbol tree until we reach the simplest form, i.e. the fixpoint.
+//
+// Consider the following symbol `(b * b) * b * b` which has this tree:
+//       *
+//      / \
+//     *   b
+//    /  \
+//   /    b
+// (b * b)
+// Now, if the `b * b == 1` new constraint is added then during the first
+// iteration we have the following transformations:
+//       *                  *
+//      / \                / \
+//     *   b     -->      b   b
+//    /  \
+//   /    b
+//  1
+// We need another iteration to reach the final result `1`.
+LLVM_NODISCARD
+static SVal simplifyUntilFixpoint(SValBuilder &SVB, ProgramStateRef State,
+                                  const SymbolRef Sym) {
+  SVal Val = SVB.makeSymbolVal(Sym);
+  SVal SimplifiedVal = SVB.simplifySVal(State, Val);
+  // Do the simplification until we can.
+  while (SimplifiedVal != Val) {
+    Val = SimplifiedVal;
+    SimplifiedVal = SVB.simplifySVal(State, Val);
+  }
+  return SimplifiedVal;
+}
+
 // Iterate over all symbols and try to simplify them. Once a symbol is
 // simplified then we check if we can merge the simplified symbol's equivalence
 // class to this class. This way, we simplify not just the symbols but the
@@ -2202,7 +2238,8 @@ EquivalenceClass::simplify(SValBuilder &SVB, RangeSet::Factory &F,
   SymbolSet ClassMembers = Class.getClassMembers(State);
   for (const SymbolRef &MemberSym : ClassMembers) {
 
-    const SVal SimplifiedMemberVal = simplifyToSVal(State, MemberSym);
+    const SVal SimplifiedMemberVal =
+        simplifyUntilFixpoint(SVB, State, MemberSym);
     const SymbolRef SimplifiedMemberSym = SimplifiedMemberVal.getAsSymbol();
 
     // The symbol is collapsed to a constant, check if the current State is
diff --git a/clang/lib/StaticAnalyzer/Core/SimpleSValBuilder.cpp b/clang/lib/StaticAnalyzer/Core/SimpleSValBuilder.cpp
index 681a1f64eadc..4ca35dd06ae5 100644
--- a/clang/lib/StaticAnalyzer/Core/SimpleSValBuilder.cpp
+++ b/clang/lib/StaticAnalyzer/Core/SimpleSValBuilder.cpp
@@ -372,6 +372,15 @@ SVal SimpleSValBuilder::evalBinOpNN(ProgramStateRef state,
   NonLoc InputLHS = lhs;
   NonLoc InputRHS = rhs;
 
+  // Constraints may have changed since the creation of a bound SVal. Check if
+  // the values can be simplified based on those new constraints.
+  SVal simplifiedLhs = simplifySVal(state, lhs);
+  SVal simplifiedRhs = simplifySVal(state, rhs);
+  if (auto simplifiedLhsAsNonLoc = simplifiedLhs.getAs<NonLoc>())
+    lhs = *simplifiedLhsAsNonLoc;
+  if (auto simplifiedRhsAsNonLoc = simplifiedRhs.getAs<NonLoc>())
+    rhs = *simplifiedRhsAsNonLoc;
+
   // Handle trivial case where left-side and right-side are the same.
   if (lhs == rhs)
     switch (op) {
@@ -619,16 +628,6 @@ SVal SimpleSValBuilder::evalBinOpNN(ProgramStateRef state,
         }
       }
 
-      // Does the symbolic expression simplify to a constant?
-      // If so, "fold" the constant by setting 'lhs' to a ConcreteInt
-      // and try again.
-      SVal simplifiedLhs = simplifySVal(state, lhs);
-      if (simplifiedLhs != lhs)
-        if (auto simplifiedLhsAsNonLoc = simplifiedLhs.getAs<NonLoc>()) {
-          lhs = *simplifiedLhsAsNonLoc;
-          continue;
-        }
-
       // Is the RHS a constant?
       if (const llvm::APSInt *RHSValue = getKnownValue(state, rhs))
         return MakeSymIntVal(Sym, op, *RHSValue, resultTy);
@@ -1103,7 +1102,6 @@ const llvm::APSInt *SimpleSValBuilder::getKnownValue(ProgramStateRef state,
   if (SymbolRef Sym = V.getAsSymbol())
     return state->getConstraintManager().getSymVal(state, Sym);
 
-  // FIXME: Add support for SymExprs.
   return nullptr;
 }
 
@@ -1135,6 +1133,24 @@ SVal SimpleSValBuilder::simplifySVal(ProgramStateRef State, SVal V) {
       return cache(Sym, SVB.makeSymbolVal(Sym));
     }
 
+    // Return the known const value for the Sym if available, or return Undef
+    // otherwise.
+    SVal getConst(SymbolRef Sym) {
+      const llvm::APSInt *Const =
+          State->getConstraintManager().getSymVal(State, Sym);
+      if (Const)
+        return Loc::isLocType(Sym->getType()) ? (SVal)SVB.makeIntLocVal(*Const)
+                                              : (SVal)SVB.makeIntVal(*Const);
+      return UndefinedVal();
+    }
+
+    SVal getConstOrVisit(SymbolRef Sym) {
+      const SVal Ret = getConst(Sym);
+      if (Ret.isUndef())
+        return Visit(Sym);
+      return Ret;
+    }
+
   public:
     Simplifier(ProgramStateRef State)
         : State(State), SVB(State->getStateManager().getSValBuilder()) {}
@@ -1148,15 +1164,14 @@ SVal SimpleSValBuilder::simplifySVal(ProgramStateRef State, SVal V) {
       return SVB.makeSymbolVal(S);
     }
 
-    // TODO: Support SymbolCast. Support IntSymExpr when/if we actually
-    // start producing them.
+    // TODO: Support SymbolCast.
 
     SVal VisitSymIntExpr(const SymIntExpr *S) {
       auto I = Cached.find(S);
       if (I != Cached.end())
         return I->second;
 
-      SVal LHS = Visit(S->getLHS());
+      SVal LHS = getConstOrVisit(S->getLHS());
       if (isUnchanged(S->getLHS(), LHS))
         return skip(S);
 
@@ -1183,6 +1198,20 @@ SVal SimpleSValBuilder::simplifySVal(ProgramStateRef State, SVal V) {
           S, SVB.evalBinOp(State, S->getOpcode(), LHS, RHS, S->getType()));
     }
 
+    SVal VisitIntSymExpr(const IntSymExpr *S) {
+      auto I = Cached.find(S);
+      if (I != Cached.end())
+        return I->second;
+
+      SVal RHS = getConstOrVisit(S->getRHS());
+      if (isUnchanged(S->getRHS(), RHS))
+        return skip(S);
+
+      SVal LHS = SVB.makeIntVal(S->getLHS());
+      return cache(
+          S, SVB.evalBinOp(State, S->getOpcode(), LHS, RHS, S->getType()));
+    }
+
     SVal VisitSymSymExpr(const SymSymExpr *S) {
       auto I = Cached.find(S);
       if (I != Cached.end())
@@ -1196,8 +1225,9 @@ SVal SimpleSValBuilder::simplifySVal(ProgramStateRef State, SVal V) {
           Loc::isLocType(S->getRHS()->getType()))
         return skip(S);
 
-      SVal LHS = Visit(S->getLHS());
-      SVal RHS = Visit(S->getRHS());
+      SVal LHS = getConstOrVisit(S->getLHS());
+      SVal RHS = getConstOrVisit(S->getRHS());
+
       if (isUnchanged(S->getLHS(), LHS) && isUnchanged(S->getRHS(), RHS))
         return skip(S);
 
diff --git a/clang/lib/StaticAnalyzer/Frontend/AnalysisConsumer.cpp b/clang/lib/StaticAnalyzer/Frontend/AnalysisConsumer.cpp
index 31de49033ac2..f692c68045ee 100644
--- a/clang/lib/StaticAnalyzer/Frontend/AnalysisConsumer.cpp
+++ b/clang/lib/StaticAnalyzer/Frontend/AnalysisConsumer.cpp
@@ -591,16 +591,24 @@ AnalysisConsumer::getModeForDecl(Decl *D, AnalysisMode Mode) {
   // - Main source file: run both path-sensitive and non-path-sensitive checks.
   // - Header files: run non-path-sensitive checks only.
   // - System headers: don't run any checks.
-  SourceManager &SM = Ctx->getSourceManager();
-  const Stmt *Body = D->getBody();
-  SourceLocation SL = Body ? Body->getBeginLoc() : D->getLocation();
-  SL = SM.getExpansionLoc(SL);
-
-  if (!Opts->AnalyzeAll && !Mgr->isInCodeFile(SL)) {
-    if (SL.isInvalid() || SM.isInSystemHeader(SL))
-      return AM_None;
+  if (Opts->AnalyzeAll)
+    return Mode;
+
+  const SourceManager &SM = Ctx->getSourceManager();
+
+  const SourceLocation Loc = [&SM](Decl *D) -> SourceLocation {
+    const Stmt *Body = D->getBody();
+    SourceLocation SL = Body ? Body->getBeginLoc() : D->getLocation();
+    return SM.getExpansionLoc(SL);
+  }(D);
+
+  // Ignore system headers.
+  if (Loc.isInvalid() || SM.isInSystemHeader(Loc))
+    return AM_None;
+
+  // Disable path sensitive analysis in user-headers.
+  if (!Mgr->isInCodeFile(Loc))
     return Mode & ~AM_Path;
-  }
 
   return Mode;
 }
diff --git a/clang/lib/Tooling/DependencyScanning/DependencyScanningFilesystem.cpp b/clang/lib/Tooling/DependencyScanning/DependencyScanningFilesystem.cpp
index 40e8bd2b8776..f7c711690d7e 100644
--- a/clang/lib/Tooling/DependencyScanning/DependencyScanningFilesystem.cpp
+++ b/clang/lib/Tooling/DependencyScanning/DependencyScanningFilesystem.cpp
@@ -129,7 +129,7 @@ DependencyScanningFilesystemSharedCache::get(StringRef Key, bool Minimized) {
 ///
 /// This is kinda hacky, it would be better if we knew what kind of file Clang
 /// was expecting instead.
-static bool shouldMinimize(StringRef Filename) {
+static bool shouldMinimizeBasedOnExtension(StringRef Filename) {
   StringRef Ext = llvm::sys::path::extension(Filename);
   if (Ext.empty())
     return true; // C++ standard library
@@ -147,26 +147,43 @@ static bool shouldCacheStatFailures(StringRef Filename) {
   StringRef Ext = llvm::sys::path::extension(Filename);
   if (Ext.empty())
     return false; // This may be the module cache directory.
-  return shouldMinimize(Filename); // Only cache stat failures on source files.
+  // Only cache stat failures on source files.
+  return shouldMinimizeBasedOnExtension(Filename);
 }
 
-void DependencyScanningWorkerFilesystem::ignoreFile(StringRef RawFilename) {
+void DependencyScanningWorkerFilesystem::disableMinimization(
+    StringRef RawFilename) {
   llvm::SmallString<256> Filename;
   llvm::sys::path::native(RawFilename, Filename);
-  IgnoredFiles.insert(Filename);
+  NotToBeMinimized.insert(Filename);
 }
 
-bool DependencyScanningWorkerFilesystem::shouldIgnoreFile(
-    StringRef RawFilename) {
+bool DependencyScanningWorkerFilesystem::shouldMinimize(StringRef RawFilename) {
+  if (!shouldMinimizeBasedOnExtension(RawFilename))
+    return false;
+
   llvm::SmallString<256> Filename;
   llvm::sys::path::native(RawFilename, Filename);
-  return IgnoredFiles.contains(Filename);
+  return !NotToBeMinimized.contains(Filename);
+}
+
+CachedFileSystemEntry DependencyScanningWorkerFilesystem::createFileSystemEntry(
+    llvm::ErrorOr<llvm::vfs::Status> &&MaybeStatus, StringRef Filename,
+    bool ShouldMinimize) {
+  if (!MaybeStatus)
+    return CachedFileSystemEntry(MaybeStatus.getError());
+
+  if (MaybeStatus->isDirectory())
+    return CachedFileSystemEntry::createDirectoryEntry(std::move(*MaybeStatus));
+
+  return CachedFileSystemEntry::createFileEntry(Filename, getUnderlyingFS(),
+                                                ShouldMinimize);
 }
 
 llvm::ErrorOr<const CachedFileSystemEntry *>
 DependencyScanningWorkerFilesystem::getOrCreateFileSystemEntry(
     const StringRef Filename) {
-  bool ShouldMinimize = !shouldIgnoreFile(Filename) && shouldMinimize(Filename);
+  bool ShouldMinimize = shouldMinimize(Filename);
 
   if (const auto *Entry = Cache.getCachedEntry(Filename, ShouldMinimize))
     return Entry;
@@ -182,23 +199,15 @@ DependencyScanningWorkerFilesystem::getOrCreateFileSystemEntry(
     CachedFileSystemEntry &CacheEntry = SharedCacheEntry.Value;
 
     if (!CacheEntry.isValid()) {
-      llvm::vfs::FileSystem &FS = getUnderlyingFS();
-      auto MaybeStatus = FS.status(Filename);
-      if (!MaybeStatus) {
-        if (!shouldCacheStatFailures(Filename))
-          // HACK: We need to always restat non source files if the stat fails.
-          //   This is because Clang first looks up the module cache and module
-          //   files before building them, and then looks for them again. If we
-          //   cache the stat failure, it won't see them the second time.
-          return MaybeStatus.getError();
-        else
-          CacheEntry = CachedFileSystemEntry(MaybeStatus.getError());
-      } else if (MaybeStatus->isDirectory())
-        CacheEntry = CachedFileSystemEntry::createDirectoryEntry(
-            std::move(*MaybeStatus));
-      else
-        CacheEntry = CachedFileSystemEntry::createFileEntry(Filename, FS,
-                                                            ShouldMinimize);
+      auto MaybeStatus = getUnderlyingFS().status(Filename);
+      if (!MaybeStatus && !shouldCacheStatFailures(Filename))
+        // HACK: We need to always restat non source files if the stat fails.
+        //   This is because Clang first looks up the module cache and module
+        //   files before building them, and then looks for them again. If we
+        //   cache the stat failure, it won't see them the second time.
+        return MaybeStatus.getError();
+      CacheEntry = createFileSystemEntry(std::move(MaybeStatus), Filename,
+                                         ShouldMinimize);
     }
 
     Result = &CacheEntry;
diff --git a/clang/lib/Tooling/DependencyScanning/DependencyScanningWorker.cpp b/clang/lib/Tooling/DependencyScanning/DependencyScanningWorker.cpp
index 7fdc49271791..70bb6c5caf87 100644
--- a/clang/lib/Tooling/DependencyScanning/DependencyScanningWorker.cpp
+++ b/clang/lib/Tooling/DependencyScanning/DependencyScanningWorker.cpp
@@ -193,20 +193,19 @@ public:
 
     // Use the dependency scanning optimized file system if requested to do so.
     if (DepFS) {
-      DepFS->clearIgnoredFiles();
-      // Ignore any files that contributed to prebuilt modules. The implicit
-      // build validates the modules by comparing the reported sizes of their
-      // inputs to the current state of the filesystem. Minimization would throw
-      // this mechanism off.
+      DepFS->enableMinimizationOfAllFiles();
+      // Don't minimize any files that contributed to prebuilt modules. The
+      // implicit build validates the modules by comparing the reported sizes of
+      // their inputs to the current state of the filesystem. Minimization would
+      // throw this mechanism off.
       for (const auto &File : PrebuiltModulesInputFiles)
-        DepFS->ignoreFile(File.getKey());
-      // Add any filenames that were explicity passed in the build settings and
-      // that might be opened, as we want to ensure we don't run source
-      // minimization on them.
+        DepFS->disableMinimization(File.getKey());
+      // Don't minimize any files that were explicitly passed in the build
+      // settings and that might be opened.
       for (const auto &E : ScanInstance.getHeaderSearchOpts().UserEntries)
-        DepFS->ignoreFile(E.Path);
+        DepFS->disableMinimization(E.Path);
       for (const auto &F : ScanInstance.getHeaderSearchOpts().VFSOverlayFiles)
-        DepFS->ignoreFile(F);
+        DepFS->disableMinimization(F);
 
       // Support for virtual file system overlays on top of the caching
       // filesystem.
diff --git a/clang/utils/TableGen/ASTTableGen.cpp b/clang/utils/TableGen/ASTTableGen.cpp
index 3f6da40964e0..6aa8b28a942f 100644
--- a/clang/utils/TableGen/ASTTableGen.cpp
+++ b/clang/utils/TableGen/ASTTableGen.cpp
@@ -107,7 +107,7 @@ static void visitASTNodeRecursive(ASTNode node, ASTNode base,
 static void visitHierarchy(RecordKeeper &records,
                            StringRef nodeClassName,
                            ASTNodeHierarchyVisitor<ASTNode> visit) {
-  // Check for the node class, just as a sanity check.
+  // Check for the node class, just as a basic correctness check.
   if (!records.getClass(nodeClassName)) {
     PrintFatalError(Twine("cannot find definition for node class ")
                       + nodeClassName);