diff --git a/clang/lib/CodeGen/CGExprScalar.cpp b/clang/lib/CodeGen/CGExprScalar.cpp
@@ -723,7 +723,9 @@ class ScalarExprEmitter
     if (Ops.Ty->isSignedIntegerOrEnumerationType()) {
       switch (CGF.getLangOpts().getSignedOverflowBehavior()) {
       case LangOptions::SOB_Defined:
-        return Builder.CreateMul(Ops.LHS, Ops.RHS, "mul");
+        if (!CGF.SanOpts.has(SanitizerKind::SignedIntegerOverflow))
+          return Builder.CreateMul(Ops.LHS, Ops.RHS, "mul");
+        [[fallthrough]];
       case LangOptions::SOB_Undefined:
         if (!CGF.SanOpts.has(SanitizerKind::SignedIntegerOverflow))
           return Builder.CreateNSWMul(Ops.LHS, Ops.RHS, "mul");
@@ -2568,7 +2570,9 @@ llvm::Value *ScalarExprEmitter::EmitIncDecConsiderOverflowBehavior(
   StringRef Name = IsInc ? "inc" : "dec";
   switch (CGF.getLangOpts().getSignedOverflowBehavior()) {
   case LangOptions::SOB_Defined:
-    return Builder.CreateAdd(InVal, Amount, Name);
+    if (!CGF.SanOpts.has(SanitizerKind::SignedIntegerOverflow))
+      return Builder.CreateAdd(InVal, Amount, Name);
+    [[fallthrough]];
   case LangOptions::SOB_Undefined:
     if (!CGF.SanOpts.has(SanitizerKind::SignedIntegerOverflow))
       return Builder.CreateNSWAdd(InVal, Amount, Name);
@@ -3913,7 +3917,9 @@ Value *ScalarExprEmitter::EmitAdd(const BinOpInfo &op) {
   if (op.Ty->isSignedIntegerOrEnumerationType()) {
     switch (CGF.getLangOpts().getSignedOverflowBehavior()) {
     case LangOptions::SOB_Defined:
-      return Builder.CreateAdd(op.LHS, op.RHS, "add");
+      if (!CGF.SanOpts.has(SanitizerKind::SignedIntegerOverflow))
+        return Builder.CreateAdd(op.LHS, op.RHS, "add");
+      [[fallthrough]];
     case LangOptions::SOB_Undefined:
       if (!CGF.SanOpts.has(SanitizerKind::SignedIntegerOverflow))
         return Builder.CreateNSWAdd(op.LHS, op.RHS, "add");
@@ -4067,7 +4073,9 @@ Value *ScalarExprEmitter::EmitSub(const BinOpInfo &op) {
     if (op.Ty->isSignedIntegerOrEnumerationType()) {
       switch (CGF.getLangOpts().getSignedOverflowBehavior()) {
       case LangOptions::SOB_Defined:
-        return Builder.CreateSub(op.LHS, op.RHS, "sub");
+        if (!CGF.SanOpts.has(SanitizerKind::SignedIntegerOverflow))
+          return Builder.CreateSub(op.LHS, op.RHS, "sub");
+        [[fallthrough]];
       case LangOptions::SOB_Undefined:
         if (!CGF.SanOpts.has(SanitizerKind::SignedIntegerOverflow))
           return Builder.CreateNSWSub(op.LHS, op.RHS, "sub");

diff --git a/clang/lib/CodeGen/CGRecordLayoutBuilder.cpp b/clang/lib/CodeGen/CGRecordLayoutBuilder.cpp
@@ -95,7 +95,7 @@ struct CGRecordLowering {
   CGRecordLowering(CodeGenTypes &Types, const RecordDecl *D, bool Packed);
   // Short helper routines.
   /// Constructs a MemberInfo instance from an offset and llvm::Type *.
-  MemberInfo StorageInfo(CharUnits Offset, llvm::Type *Data) {
+  static MemberInfo StorageInfo(CharUnits Offset, llvm::Type *Data) {
     return MemberInfo(Offset, MemberInfo::Field, Data);
   }
 
@@ -104,7 +104,7 @@ struct CGRecordLowering {
   /// fields of the same formal type.  We want to emit a layout with
   /// these discrete storage units instead of combining them into a
   /// continuous run.
-  bool isDiscreteBitFieldABI() {
+  bool isDiscreteBitFieldABI() const {
     return Context.getTargetInfo().getCXXABI().isMicrosoft() ||
            D->isMsStruct(Context);
   }
@@ -121,60 +121,60 @@ struct CGRecordLowering {
   /// other bases, which complicates layout in specific ways.
   ///
   /// Note specifically that the ms_struct attribute doesn't change this.
-  bool isOverlappingVBaseABI() {
+  bool isOverlappingVBaseABI() const {
     return !Context.getTargetInfo().getCXXABI().isMicrosoft();
   }
 
   /// Wraps llvm::Type::getIntNTy with some implicit arguments.
-  llvm::Type *getIntNType(uint64_t NumBits) {
+  llvm::Type *getIntNType(uint64_t NumBits) const {
     unsigned AlignedBits = llvm::alignTo(NumBits, Context.getCharWidth());
     return llvm::Type::getIntNTy(Types.getLLVMContext(), AlignedBits);
   }
   /// Get the LLVM type sized as one character unit.
-  llvm::Type *getCharType() {
+  llvm::Type *getCharType() const {
     return llvm::Type::getIntNTy(Types.getLLVMContext(),
                                  Context.getCharWidth());
   }
   /// Gets an llvm type of size NumChars and alignment 1.
-  llvm::Type *getByteArrayType(CharUnits NumChars) {
+  llvm::Type *getByteArrayType(CharUnits NumChars) const {
     assert(!NumChars.isZero() && "Empty byte arrays aren't allowed.");
     llvm::Type *Type = getCharType();
     return NumChars == CharUnits::One() ? Type :
         (llvm::Type *)llvm::ArrayType::get(Type, NumChars.getQuantity());
   }
   /// Gets the storage type for a field decl and handles storage
   /// for itanium bitfields that are smaller than their declared type.
-  llvm::Type *getStorageType(const FieldDecl *FD) {
+  llvm::Type *getStorageType(const FieldDecl *FD) const {
     llvm::Type *Type = Types.ConvertTypeForMem(FD->getType());
     if (!FD->isBitField()) return Type;
     if (isDiscreteBitFieldABI()) return Type;
     return getIntNType(std::min(FD->getBitWidthValue(Context),
                              (unsigned)Context.toBits(getSize(Type))));
   }
   /// Gets the llvm Basesubobject type from a CXXRecordDecl.
-  llvm::Type *getStorageType(const CXXRecordDecl *RD) {
+  llvm::Type *getStorageType(const CXXRecordDecl *RD) const {
     return Types.getCGRecordLayout(RD).getBaseSubobjectLLVMType();
   }
-  CharUnits bitsToCharUnits(uint64_t BitOffset) {
+  CharUnits bitsToCharUnits(uint64_t BitOffset) const {
     return Context.toCharUnitsFromBits(BitOffset);
   }
-  CharUnits getSize(llvm::Type *Type) {
+  CharUnits getSize(llvm::Type *Type) const {
     return CharUnits::fromQuantity(DataLayout.getTypeAllocSize(Type));
   }
-  CharUnits getAlignment(llvm::Type *Type) {
+  CharUnits getAlignment(llvm::Type *Type) const {
     return CharUnits::fromQuantity(DataLayout.getABITypeAlign(Type));
   }
-  bool isZeroInitializable(const FieldDecl *FD) {
+  bool isZeroInitializable(const FieldDecl *FD) const {
     return Types.isZeroInitializable(FD->getType());
   }
-  bool isZeroInitializable(const RecordDecl *RD) {
+  bool isZeroInitializable(const RecordDecl *RD) const {
     return Types.isZeroInitializable(RD);
   }
   void appendPaddingBytes(CharUnits Size) {
     if (!Size.isZero())
       FieldTypes.push_back(getByteArrayType(Size));
   }
-  uint64_t getFieldBitOffset(const FieldDecl *FD) {
+  uint64_t getFieldBitOffset(const FieldDecl *FD) const {
     return Layout.getFieldOffset(FD->getFieldIndex());
   }
   // Layout routines.

diff --git a/clang/lib/CodeGen/CodeGenFunction.h b/clang/lib/CodeGen/CodeGenFunction.h
@@ -4405,6 +4405,7 @@ class CodeGenFunction : public CodeGenTypeCache {
   llvm::Value *EmitX86BuiltinExpr(unsigned BuiltinID, const CallExpr *E);
   llvm::Value *EmitPPCBuiltinExpr(unsigned BuiltinID, const CallExpr *E);
   llvm::Value *EmitAMDGPUBuiltinExpr(unsigned BuiltinID, const CallExpr *E);
+  llvm::Value *EmitHLSLBuiltinExpr(unsigned BuiltinID, const CallExpr *E);
   llvm::Value *EmitScalarOrConstFoldImmArg(unsigned ICEArguments, unsigned Idx,
                                            const CallExpr *E);
   llvm::Value *EmitSystemZBuiltinExpr(unsigned BuiltinID, const CallExpr *E);
@@ -5013,10 +5014,10 @@ class CodeGenFunction : public CodeGenTypeCache {
   llvm::Value *EmitAArch64CpuInit();
   llvm::Value *
   FormAArch64ResolverCondition(const MultiVersionResolverOption &RO);
+  llvm::Value *EmitAArch64CpuSupports(const CallExpr *E);
   llvm::Value *EmitAArch64CpuSupports(ArrayRef<StringRef> FeatureStrs);
 };
 
-
 inline DominatingLLVMValue::saved_type
 DominatingLLVMValue::save(CodeGenFunction &CGF, llvm::Value *value) {
   if (!needsSaving(value)) return saved_type(value, false);

diff --git a/clang/lib/CodeGen/CodeGenModule.cpp b/clang/lib/CodeGen/CodeGenModule.cpp
@@ -915,7 +915,15 @@ void CodeGenModule::Release() {
         llvm::ConstantArray::get(ATy, UsedArray), "__clang_gpu_used_external");
     addCompilerUsedGlobal(GV);
   }
-
+  if (LangOpts.HIP) {
+    // Emit a unique ID so that host and device binaries from the same
+    // compilation unit can be associated.
+    auto *GV = new llvm::GlobalVariable(
+        getModule(), Int8Ty, false, llvm::GlobalValue::ExternalLinkage,
+        llvm::Constant::getNullValue(Int8Ty),
+        "__hip_cuid_" + getContext().getCUIDHash());
+    addCompilerUsedGlobal(GV);
+  }
   emitLLVMUsed();
   if (SanStats)
     SanStats->finish();

diff --git a/clang/lib/CodeGen/CodeGenPGO.cpp b/clang/lib/CodeGen/CodeGenPGO.cpp
@@ -165,6 +165,7 @@ struct MapRegionCounters : public RecursiveASTVisitor<MapRegionCounters> {
   llvm::DenseMap<const Stmt *, unsigned> &CounterMap;
   /// The next bitmap byte index to assign.
   unsigned NextMCDCBitmapIdx;
+  /// The state of MC/DC Coverage in this function.
   MCDC::State &MCDCState;
   /// Maximum number of supported MC/DC conditions in a boolean expression.
   unsigned MCDCMaxCond;
@@ -239,9 +240,12 @@ struct MapRegionCounters : public RecursiveASTVisitor<MapRegionCounters> {
     if (MCDCMaxCond == 0)
       return true;
 
-    /// At the top of the logical operator nest, reset the number of conditions.
-    if (LogOpStack.empty())
+    /// At the top of the logical operator nest, reset the number of conditions,
+    /// also forget previously seen split nesting cases.
+    if (LogOpStack.empty()) {
       NumCond = 0;
+      SplitNestedLogicalOp = false;
+    }
 
     if (const Expr *E = dyn_cast<Expr>(S)) {
       const BinaryOperator *BinOp = dyn_cast<BinaryOperator>(E->IgnoreParens());
@@ -292,7 +296,7 @@ struct MapRegionCounters : public RecursiveASTVisitor<MapRegionCounters> {
                 "contains an operation with a nested boolean expression. "
                 "Expression will not be covered");
             Diag.Report(S->getBeginLoc(), DiagID);
-            return false;
+            return true;
           }
 
           /// Was the maximum number of conditions encountered?
@@ -303,12 +307,12 @@ struct MapRegionCounters : public RecursiveASTVisitor<MapRegionCounters> {
                 "number of conditions (%0) exceeds max (%1). "
                 "Expression will not be covered");
             Diag.Report(S->getBeginLoc(), DiagID) << NumCond << MCDCMaxCond;
-            return false;
+            return true;
           }
 
           // Otherwise, allocate the number of bytes required for the bitmap
           // based on the number of conditions. Must be at least 1-byte long.
-          MCDCState.BitmapMap[BinOp] = NextMCDCBitmapIdx;
+          MCDCState.DecisionByStmt[BinOp].BitmapIdx = NextMCDCBitmapIdx;
           unsigned SizeInBits = std::max<unsigned>(1L << NumCond, CHAR_BIT);
           NextMCDCBitmapIdx += SizeInBits / CHAR_BIT;
         }
@@ -1031,7 +1035,7 @@ void CodeGenPGO::emitCounterRegionMapping(const Decl *D) {
 
   std::string CoverageMapping;
   llvm::raw_string_ostream OS(CoverageMapping);
-  RegionCondIDMap.reset(new llvm::DenseMap<const Stmt *, int16_t>);
+  RegionMCDCState->BranchByStmt.clear();
   CoverageMappingGen MappingGen(
       *CGM.getCoverageMapping(), CGM.getContext().getSourceManager(),
       CGM.getLangOpts(), RegionCounterMap.get(), RegionMCDCState.get());
@@ -1139,12 +1143,12 @@ void CodeGenPGO::emitMCDCTestVectorBitmapUpdate(CGBuilderTy &Builder,
 
   S = S->IgnoreParens();
 
-  auto ExprMCDCBitmapMapIterator = RegionMCDCState->BitmapMap.find(S);
-  if (ExprMCDCBitmapMapIterator == RegionMCDCState->BitmapMap.end())
+  auto DecisionStateIter = RegionMCDCState->DecisionByStmt.find(S);
+  if (DecisionStateIter == RegionMCDCState->DecisionByStmt.end())
     return;
 
-  // Extract the ID of the global bitmap associated with this expression.
-  unsigned MCDCTestVectorBitmapID = ExprMCDCBitmapMapIterator->second;
+  // Extract the offset of the global bitmap associated with this expression.
+  unsigned MCDCTestVectorBitmapOffset = DecisionStateIter->second.BitmapIdx;
   auto *I8PtrTy = llvm::PointerType::getUnqual(CGM.getLLVMContext());
 
   // Emit intrinsic responsible for updating the global bitmap corresponding to
@@ -1155,7 +1159,7 @@ void CodeGenPGO::emitMCDCTestVectorBitmapUpdate(CGBuilderTy &Builder,
   llvm::Value *Args[5] = {llvm::ConstantExpr::getBitCast(FuncNameVar, I8PtrTy),
                           Builder.getInt64(FunctionHash),
                           Builder.getInt32(RegionMCDCState->BitmapBytes),
-                          Builder.getInt32(MCDCTestVectorBitmapID),
+                          Builder.getInt32(MCDCTestVectorBitmapOffset),
                           MCDCCondBitmapAddr.getPointer()};
   Builder.CreateCall(
       CGM.getIntrinsic(llvm::Intrinsic::instrprof_mcdc_tvbitmap_update), Args);
@@ -1168,7 +1172,7 @@ void CodeGenPGO::emitMCDCCondBitmapReset(CGBuilderTy &Builder, const Expr *S,
 
   S = S->IgnoreParens();
 
-  if (!RegionMCDCState->BitmapMap.contains(S))
+  if (!RegionMCDCState->DecisionByStmt.contains(S))
     return;
 
   // Emit intrinsic that resets a dedicated temporary value on the stack to 0.
@@ -1190,13 +1194,13 @@ void CodeGenPGO::emitMCDCCondBitmapUpdate(CGBuilderTy &Builder, const Expr *S,
   // also make debugging a bit easier.
   S = CodeGenFunction::stripCond(S);
 
-  auto ExprMCDCConditionIDMapIterator = RegionMCDCState->CondIDMap.find(S);
-  if (ExprMCDCConditionIDMapIterator == RegionMCDCState->CondIDMap.end())
+  auto BranchStateIter = RegionMCDCState->BranchByStmt.find(S);
+  if (BranchStateIter == RegionMCDCState->BranchByStmt.end())
     return;
 
   // Extract the ID of the condition we are setting in the bitmap.
-  auto CondID = ExprMCDCConditionIDMapIterator->second;
-  assert(CondID >= 0 && "Condition has no ID!");
+  const auto &Branch = BranchStateIter->second;
+  assert(Branch.ID >= 0 && "Condition has no ID!");
 
   auto *I8PtrTy = llvm::PointerType::getUnqual(CGM.getLLVMContext());
 
@@ -1205,7 +1209,7 @@ void CodeGenPGO::emitMCDCCondBitmapUpdate(CGBuilderTy &Builder, const Expr *S,
   // the resulting value is used to update the boolean expression's bitmap.
   llvm::Value *Args[5] = {llvm::ConstantExpr::getBitCast(FuncNameVar, I8PtrTy),
                           Builder.getInt64(FunctionHash),
-                          Builder.getInt32(CondID),
+                          Builder.getInt32(Branch.ID),
                           MCDCCondBitmapAddr.getPointer(), Val};
   Builder.CreateCall(
       CGM.getIntrinsic(llvm::Intrinsic::instrprof_mcdc_condbitmap_update),

diff --git a/clang/lib/CodeGen/CodeGenPGO.h b/clang/lib/CodeGen/CodeGenPGO.h
@@ -36,8 +36,6 @@ class CodeGenPGO {
   unsigned NumRegionCounters;
   uint64_t FunctionHash;
   std::unique_ptr<llvm::DenseMap<const Stmt *, unsigned>> RegionCounterMap;
-  std::unique_ptr<llvm::DenseMap<const Stmt *, unsigned>> RegionMCDCBitmapMap;
-  std::unique_ptr<llvm::DenseMap<const Stmt *, int16_t>> RegionCondIDMap;
   std::unique_ptr<llvm::DenseMap<const Stmt *, uint64_t>> StmtCountMap;
   std::unique_ptr<llvm::InstrProfRecord> ProfRecord;
   std::unique_ptr<MCDC::State> RegionMCDCState;

diff --git a/clang/lib/CodeGen/CoverageMappingGen.cpp b/clang/lib/CodeGen/CoverageMappingGen.cpp
@@ -684,7 +684,6 @@ struct MCDCCoverageBuilder {
 
   llvm::SmallVector<mcdc::ConditionIDs> DecisionStack;
   MCDC::State &MCDCState;
-  llvm::DenseMap<const Stmt *, mcdc::ConditionID> &CondIDs;
   mcdc::ConditionID NextID = 0;
   bool NotMapped = false;
 
@@ -699,8 +698,8 @@ struct MCDCCoverageBuilder {
 
 public:
   MCDCCoverageBuilder(CodeGenModule &CGM, MCDC::State &MCDCState)
-      : CGM(CGM), DecisionStack(1, DecisionStackSentinel), MCDCState(MCDCState),
-        CondIDs(MCDCState.CondIDMap) {}
+      : CGM(CGM), DecisionStack(1, DecisionStackSentinel),
+        MCDCState(MCDCState) {}
 
   /// Return whether the build of the control flow map is at the top-level
   /// (root) of a logical operator nest in a boolean expression prior to the
@@ -714,16 +713,16 @@ struct MCDCCoverageBuilder {
 
   /// Set the given condition's ID.
   void setCondID(const Expr *Cond, mcdc::ConditionID ID) {
-    CondIDs[CodeGenFunction::stripCond(Cond)] = ID;
+    MCDCState.BranchByStmt[CodeGenFunction::stripCond(Cond)].ID = ID;
   }
 
   /// Return the ID of a given condition.
   mcdc::ConditionID getCondID(const Expr *Cond) const {
-    auto I = CondIDs.find(CodeGenFunction::stripCond(Cond));
-    if (I == CondIDs.end())
+    auto I = MCDCState.BranchByStmt.find(CodeGenFunction::stripCond(Cond));
+    if (I == MCDCState.BranchByStmt.end())
       return -1;
     else
-      return I->second;
+      return I->second.ID;
   }
 
   /// Return the LHS Decision ([0,0] if not set).
@@ -738,7 +737,7 @@ struct MCDCCoverageBuilder {
 
     // If binary expression is disqualified, don't do mapping.
     if (!isBuilding() &&
-        !MCDCState.BitmapMap.contains(CodeGenFunction::stripCond(E)))
+        !MCDCState.DecisionByStmt.contains(CodeGenFunction::stripCond(E)))
       NotMapped = true;
 
     // Don't go any further if we don't need to map condition IDs.
@@ -750,7 +749,7 @@ struct MCDCCoverageBuilder {
     // If the operator itself has an assigned ID, this means it represents a
     // larger subtree.  In this case, assign that ID to its LHS node.  Its RHS
     // will receive a new ID below. Otherwise, assign ID+1 to LHS.
-    if (CondIDs.contains(CodeGenFunction::stripCond(E)))
+    if (MCDCState.BranchByStmt.contains(CodeGenFunction::stripCond(E)))
       setCondID(E->getLHS(), getCondID(E));
     else
       setCondID(E->getLHS(), NextID++);
@@ -853,8 +852,6 @@ struct CounterCoverageMappingBuilder
     return Counter::getCounter(CounterMap[S]);
   }
 
-  unsigned getRegionBitmap(const Stmt *S) { return MCDCState.BitmapMap[S]; }
-
   /// Push a region onto the stack.
   ///
   /// Returns the index on the stack where the region was pushed. This can be
@@ -887,12 +884,11 @@ struct CounterCoverageMappingBuilder
     return RegionStack.size() - 1;
   }
 
-  size_t pushRegion(unsigned BitmapIdx, uint16_t Conditions,
+  size_t pushRegion(const mcdc::DecisionParameters &DecisionParams,
                     std::optional<SourceLocation> StartLoc = std::nullopt,
                     std::optional<SourceLocation> EndLoc = std::nullopt) {
 
-    RegionStack.emplace_back(mcdc::DecisionParameters{BitmapIdx, Conditions},
-                             StartLoc, EndLoc);
+    RegionStack.emplace_back(DecisionParams, StartLoc, EndLoc);
 
     return RegionStack.size() - 1;
   }
@@ -1058,8 +1054,9 @@ struct CounterCoverageMappingBuilder
   /// Create a Decision Region with a BitmapIdx and number of Conditions. This
   /// type of region "contains" branch regions, one for each of the conditions.
   /// The visualization tool will group everything together.
-  void createDecisionRegion(const Expr *C, unsigned BitmapIdx, unsigned Conds) {
-    popRegions(pushRegion(BitmapIdx, Conds, getStart(C), getEnd(C)));
+  void createDecisionRegion(const Expr *C,
+                            const mcdc::DecisionParameters &DecisionParams) {
+    popRegions(pushRegion(DecisionParams, getStart(C), getEnd(C)));
   }
 
   /// Create a Branch Region around a SwitchCase for code coverage
@@ -1311,7 +1308,7 @@ struct CounterCoverageMappingBuilder
       return;
     assert(SpellingRegion(SM, NewStartLoc, EndLoc).isInSourceOrder());
     handleFileExit(NewStartLoc);
-    size_t Index = pushRegion({}, NewStartLoc, EndLoc);
+    size_t Index = pushRegion(Counter{}, NewStartLoc, EndLoc);
     getRegion().setSkipped(true);
     handleFileExit(EndLoc);
     popRegions(Index);
@@ -1961,6 +1958,20 @@ struct CounterCoverageMappingBuilder
                        subtractCounters(ParentCount, TrueCount));
   }
 
+  void createDecision(const BinaryOperator *E) {
+    unsigned NumConds = MCDCBuilder.getTotalConditionsAndReset(E);
+    if (NumConds == 0)
+      return;
+
+    auto DecisionParams = mcdc::DecisionParameters{
+        MCDCState.DecisionByStmt[E].BitmapIdx,
+        NumConds,
+    };
+
+    // Create MCDC Decision Region.
+    createDecisionRegion(E, DecisionParams);
+  }
+
   void VisitBinLAnd(const BinaryOperator *E) {
     bool IsRootNode = MCDCBuilder.isIdle();
 
@@ -1981,11 +1992,6 @@ struct CounterCoverageMappingBuilder
     // Track RHS True/False Decision.
     const auto DecisionRHS = MCDCBuilder.back();
 
-    // Create MCDC Decision Region if at top-level (root).
-    unsigned NumConds = 0;
-    if (IsRootNode && (NumConds = MCDCBuilder.getTotalConditionsAndReset(E)))
-      createDecisionRegion(E, getRegionBitmap(E), NumConds);
-
     // Extract the RHS's Execution Counter.
     Counter RHSExecCnt = getRegionCounter(E);
 
@@ -2002,6 +2008,10 @@ struct CounterCoverageMappingBuilder
     // Create Branch Region around RHS condition.
     createBranchRegion(E->getRHS(), RHSTrueCnt,
                        subtractCounters(RHSExecCnt, RHSTrueCnt), DecisionRHS);
+
+    // Create MCDC Decision Region if at top-level (root).
+    if (IsRootNode)
+      createDecision(E);
   }
 
   // Determine whether the right side of OR operation need to be visited.
@@ -2034,11 +2044,6 @@ struct CounterCoverageMappingBuilder
     // Track RHS True/False Decision.
     const auto DecisionRHS = MCDCBuilder.back();
 
-    // Create MCDC Decision Region if at top-level (root).
-    unsigned NumConds = 0;
-    if (IsRootNode && (NumConds = MCDCBuilder.getTotalConditionsAndReset(E)))
-      createDecisionRegion(E, getRegionBitmap(E), NumConds);
-
     // Extract the RHS's Execution Counter.
     Counter RHSExecCnt = getRegionCounter(E);
 
@@ -2059,6 +2064,10 @@ struct CounterCoverageMappingBuilder
     // Create Branch Region around RHS condition.
     createBranchRegion(E->getRHS(), subtractCounters(RHSExecCnt, RHSFalseCnt),
                        RHSFalseCnt, DecisionRHS);
+
+    // Create MCDC Decision Region if at top-level (root).
+    if (IsRootNode)
+      createDecision(E);
   }
 
   void VisitLambdaExpr(const LambdaExpr *LE) {

diff --git a/clang/lib/CodeGen/MCDCState.h b/clang/lib/CodeGen/MCDCState.h
@@ -27,8 +27,18 @@ using namespace llvm::coverage::mcdc;
 /// Per-Function MC/DC state
 struct State {
   unsigned BitmapBytes = 0;
-  llvm::DenseMap<const Stmt *, unsigned> BitmapMap;
-  llvm::DenseMap<const Stmt *, ConditionID> CondIDMap;
+
+  struct Decision {
+    unsigned BitmapIdx;
+  };
+
+  llvm::DenseMap<const Stmt *, Decision> DecisionByStmt;
+
+  struct Branch {
+    ConditionID ID;
+  };
+
+  llvm::DenseMap<const Stmt *, Branch> BranchByStmt;
 };
 
 } // namespace clang::CodeGen::MCDC

diff --git a/clang/lib/CodeGen/Targets/AArch64.cpp b/clang/lib/CodeGen/Targets/AArch64.cpp
@@ -8,6 +8,7 @@
 
 #include "ABIInfoImpl.h"
 #include "TargetInfo.h"
+#include "clang/Basic/DiagnosticFrontend.h"
 
 using namespace clang;
 using namespace clang::CodeGen;
@@ -155,6 +156,11 @@ class AArch64TargetCodeGenInfo : public TargetCodeGenInfo {
     }
     return TargetCodeGenInfo::isScalarizableAsmOperand(CGF, Ty);
   }
+
+  void checkFunctionCallABI(CodeGenModule &CGM, SourceLocation CallLoc,
+                            const FunctionDecl *Caller,
+                            const FunctionDecl *Callee,
+                            const CallArgList &Args) const override;
 };
 
 class WindowsAArch64TargetCodeGenInfo : public AArch64TargetCodeGenInfo {
@@ -814,6 +820,43 @@ Address AArch64ABIInfo::EmitMSVAArg(CodeGenFunction &CGF, Address VAListAddr,
                           /*allowHigherAlign*/ false);
 }
 
+static bool isStreaming(const FunctionDecl *F) {
+  if (F->hasAttr<ArmLocallyStreamingAttr>())
+    return true;
+  if (const auto *T = F->getType()->getAs<FunctionProtoType>())
+    return T->getAArch64SMEAttributes() & FunctionType::SME_PStateSMEnabledMask;
+  return false;
+}
+
+static bool isStreamingCompatible(const FunctionDecl *F) {
+  if (const auto *T = F->getType()->getAs<FunctionProtoType>())
+    return T->getAArch64SMEAttributes() &
+           FunctionType::SME_PStateSMCompatibleMask;
+  return false;
+}
+
+void AArch64TargetCodeGenInfo::checkFunctionCallABI(
+    CodeGenModule &CGM, SourceLocation CallLoc, const FunctionDecl *Caller,
+    const FunctionDecl *Callee, const CallArgList &Args) const {
+  if (!Caller || !Callee || !Callee->hasAttr<AlwaysInlineAttr>())
+    return;
+
+  bool CallerIsStreaming = isStreaming(Caller);
+  bool CalleeIsStreaming = isStreaming(Callee);
+  bool CallerIsStreamingCompatible = isStreamingCompatible(Caller);
+  bool CalleeIsStreamingCompatible = isStreamingCompatible(Callee);
+
+  if (!CalleeIsStreamingCompatible &&
+      (CallerIsStreaming != CalleeIsStreaming || CallerIsStreamingCompatible))
+    CGM.getDiags().Report(CallLoc,
+                          diag::err_function_always_inline_attribute_mismatch)
+        << Caller->getDeclName() << Callee->getDeclName() << "streaming";
+  if (auto *NewAttr = Callee->getAttr<ArmNewAttr>())
+    if (NewAttr->isNewZA())
+      CGM.getDiags().Report(CallLoc, diag::err_function_always_inline_new_za)
+          << Callee->getDeclName();
+}
+
 std::unique_ptr<TargetCodeGenInfo>
 CodeGen::createAArch64TargetCodeGenInfo(CodeGenModule &CGM,
                                         AArch64ABIKind Kind) {

diff --git a/clang/lib/CodeGen/Targets/NVPTX.cpp b/clang/lib/CodeGen/Targets/NVPTX.cpp
@@ -47,6 +47,10 @@ class NVPTXTargetCodeGenInfo : public TargetCodeGenInfo {
                            CodeGen::CodeGenModule &M) const override;
   bool shouldEmitStaticExternCAliases() const override;
 
+  llvm::Constant *getNullPointer(const CodeGen::CodeGenModule &CGM,
+                                 llvm::PointerType *T,
+                                 QualType QT) const override;
+
   llvm::Type *getCUDADeviceBuiltinSurfaceDeviceType() const override {
     // On the device side, surface reference is represented as an object handle
     // in 64-bit integer.
@@ -285,6 +289,20 @@ void NVPTXTargetCodeGenInfo::addNVVMMetadata(llvm::GlobalValue *GV,
 bool NVPTXTargetCodeGenInfo::shouldEmitStaticExternCAliases() const {
   return false;
 }
+
+llvm::Constant *
+NVPTXTargetCodeGenInfo::getNullPointer(const CodeGen::CodeGenModule &CGM,
+                                       llvm::PointerType *PT,
+                                       QualType QT) const {
+  auto &Ctx = CGM.getContext();
+  if (PT->getAddressSpace() != Ctx.getTargetAddressSpace(LangAS::opencl_local))
+    return llvm::ConstantPointerNull::get(PT);
+
+  auto NPT = llvm::PointerType::get(
+      PT->getContext(), Ctx.getTargetAddressSpace(LangAS::opencl_generic));
+  return llvm::ConstantExpr::getAddrSpaceCast(
+      llvm::ConstantPointerNull::get(NPT), PT);
+}
 }
 
 void CodeGenModule::handleCUDALaunchBoundsAttr(llvm::Function *F,

diff --git a/clang/lib/Driver/OffloadBundler.cpp b/clang/lib/Driver/OffloadBundler.cpp
@@ -588,8 +588,15 @@ class ObjectFileHandler final : public FileHandler {
     StringRef Content = *ContentOrErr;
 
     // Copy fat object contents to the output when extracting host bundle.
-    if (Content.size() == 1u && Content.front() == 0)
-      Content = StringRef(Input.getBufferStart(), Input.getBufferSize());
+    std::string ModifiedContent;
+    if (Content.size() == 1u && Content.front() == 0) {
+      auto HostBundleOrErr = getHostBundle();
+      if (!HostBundleOrErr)
+        return HostBundleOrErr.takeError();
+
+      ModifiedContent = std::move(*HostBundleOrErr);
+      Content = ModifiedContent;
+    }
 
     OS.write(Content.data(), Content.size());
     return Error::success();
@@ -692,6 +699,35 @@ class ObjectFileHandler final : public FileHandler {
     }
     return Error::success();
   }
+
+  Expected<std::string> getHostBundle() {
+    TempFileHandlerRAII TempFiles;
+
+    auto ModifiedObjPathOrErr = TempFiles.Create(std::nullopt);
+    if (!ModifiedObjPathOrErr)
+      return ModifiedObjPathOrErr.takeError();
+    StringRef ModifiedObjPath = *ModifiedObjPathOrErr;
+
+    BumpPtrAllocator Alloc;
+    StringSaver SS{Alloc};
+    SmallVector<StringRef, 16> ObjcopyArgs{"llvm-objcopy"};
+
+    ObjcopyArgs.push_back("--regex");
+    ObjcopyArgs.push_back("--remove-section=__CLANG_OFFLOAD_BUNDLE__.*");
+    ObjcopyArgs.push_back("--");
+    ObjcopyArgs.push_back(BundlerConfig.InputFileNames.front());
+    ObjcopyArgs.push_back(ModifiedObjPath);
+
+    if (Error Err = executeObjcopy(BundlerConfig.ObjcopyPath, ObjcopyArgs))
+      return std::move(Err);
+
+    auto BufOrErr = MemoryBuffer::getFile(ModifiedObjPath);
+    if (!BufOrErr)
+      return createStringError(BufOrErr.getError(),
+                               "Failed to read back the modified object file");
+
+    return BufOrErr->get()->getBuffer().str();
+  }
 };
 
 /// Handler for text files. The bundled file will have the following format.

diff --git a/clang/lib/Driver/ToolChain.cpp b/clang/lib/Driver/ToolChain.cpp
@@ -77,10 +77,19 @@ static ToolChain::RTTIMode CalculateRTTIMode(const ArgList &Args,
   return NoRTTI ? ToolChain::RM_Disabled : ToolChain::RM_Enabled;
 }
 
+static ToolChain::ExceptionsMode CalculateExceptionsMode(const ArgList &Args) {
+  if (Args.hasFlag(options::OPT_fexceptions, options::OPT_fno_exceptions,
+                   true)) {
+    return ToolChain::EM_Enabled;
+  }
+  return ToolChain::EM_Disabled;
+}
+
 ToolChain::ToolChain(const Driver &D, const llvm::Triple &T,
                      const ArgList &Args)
     : D(D), Triple(T), Args(Args), CachedRTTIArg(GetRTTIArgument(Args)),
-      CachedRTTIMode(CalculateRTTIMode(Args, Triple, CachedRTTIArg)) {
+      CachedRTTIMode(CalculateRTTIMode(Args, Triple, CachedRTTIArg)),
+      CachedExceptionsMode(CalculateExceptionsMode(Args)) {
   auto addIfExists = [this](path_list &List, const std::string &Path) {
     if (getVFS().exists(Path))
       List.push_back(Path);
@@ -191,7 +200,12 @@ static void getAArch64MultilibFlags(const Driver &D,
   for (const auto &Ext : AArch64::Extensions)
     if (FeatureSet.contains(Ext.NegFeature))
       MArch.push_back(("no" + Ext.Name).str());
-  MArch.insert(MArch.begin(), ("-march=" + Triple.getArchName()).str());
+  StringRef ArchName;
+  for (const auto &ArchInfo : AArch64::ArchInfos)
+    if (FeatureSet.contains(ArchInfo->ArchFeature))
+      ArchName = ArchInfo->Name;
+  assert(!ArchName.empty() && "at least one architecture should be found");
+  MArch.insert(MArch.begin(), ("-march=" + ArchName).str());
   Result.push_back(llvm::join(MArch, "+"));
 }
 
@@ -264,6 +278,18 @@ ToolChain::getMultilibFlags(const llvm::opt::ArgList &Args) const {
     break;
   }
 
+  // Include fno-exceptions and fno-rtti
+  // to improve multilib selection
+  if (getRTTIMode() == ToolChain::RTTIMode::RM_Disabled)
+    Result.push_back("-fno-rtti");
+  else
+    Result.push_back("-frtti");
+
+  if (getExceptionsMode() == ToolChain::ExceptionsMode::EM_Disabled)
+    Result.push_back("-fno-exceptions");
+  else
+    Result.push_back("-fexceptions");
+
   // Sort and remove duplicates.
   std::sort(Result.begin(), Result.end());
   Result.erase(std::unique(Result.begin(), Result.end()), Result.end());
@@ -655,19 +681,29 @@ std::string ToolChain::getCompilerRT(const ArgList &Args, StringRef Component,
   // Check for runtime files in the new layout without the architecture first.
   std::string CRTBasename =
       buildCompilerRTBasename(Args, Component, Type, /*AddArch=*/false);
+  SmallString<128> Path;
   for (const auto &LibPath : getLibraryPaths()) {
     SmallString<128> P(LibPath);
     llvm::sys::path::append(P, CRTBasename);
     if (getVFS().exists(P))
       return std::string(P);
+    if (Path.empty())
+      Path = P;
   }
+  if (getTriple().isOSAIX())
+    Path.clear();
 
-  // Fall back to the old expected compiler-rt name if the new one does not
-  // exist.
+  // Check the filename for the old layout if the new one does not exist.
   CRTBasename =
       buildCompilerRTBasename(Args, Component, Type, /*AddArch=*/true);
-  SmallString<128> Path(getCompilerRTPath());
-  llvm::sys::path::append(Path, CRTBasename);
+  SmallString<128> OldPath(getCompilerRTPath());
+  llvm::sys::path::append(OldPath, CRTBasename);
+  if (Path.empty() || getVFS().exists(OldPath))
+    return std::string(OldPath);
+
+  // If none is found, use a file name from the new layout, which may get
+  // printed in an error message, aiding users in knowing what Clang is
+  // looking for.
   return std::string(Path);
 }
 

diff --git a/clang/lib/Driver/ToolChains/AMDGPU.cpp b/clang/lib/Driver/ToolChains/AMDGPU.cpp
@@ -625,6 +625,7 @@ void amdgpu::Linker::ConstructJob(Compilation &C, const JobAction &JA,
 
   addLinkerCompressDebugSectionsOption(getToolChain(), Args, CmdArgs);
   Args.AddAllArgs(CmdArgs, options::OPT_L);
+  getToolChain().AddFilePathLibArgs(Args, CmdArgs);
   AddLinkerInputs(getToolChain(), Inputs, Args, CmdArgs, JA);
   if (C.getDriver().isUsingLTO())
     addLTOOptions(getToolChain(), Args, CmdArgs, Output, Inputs[0],

diff --git a/clang/lib/Driver/ToolChains/Arch/ARM.cpp b/clang/lib/Driver/ToolChains/Arch/ARM.cpp
@@ -890,25 +890,25 @@ llvm::ARM::FPUKind arm::getARMTargetFeatures(const Driver &D,
     // SCTLR.U bit, which is architecture-specific. We assume ARMv6
     // Darwin and NetBSD targets support unaligned accesses, and others don't.
     //
-    // ARMv7 always has SCTLR.U set to 1, but it has a new SCTLR.A bit
-    // which raises an alignment fault on unaligned accesses. Linux
-    // defaults this bit to 0 and handles it as a system-wide (not
-    // per-process) setting. It is therefore safe to assume that ARMv7+
-    // Linux targets support unaligned accesses. The same goes for NaCl
-    // and Windows.
+    // ARMv7 always has SCTLR.U set to 1, but it has a new SCTLR.A bit which
+    // raises an alignment fault on unaligned accesses. Assume ARMv7+ supports
+    // unaligned accesses, except ARMv6-M, and ARMv8-M without the Main
+    // Extension. This aligns with the default behavior of ARM's downstream
+    // versions of GCC and Clang.
     //
-    // The above behavior is consistent with GCC.
+    // Users can change the default behavior via -m[no-]unaliged-access.
     int VersionNum = getARMSubArchVersionNumber(Triple);
     if (Triple.isOSDarwin() || Triple.isOSNetBSD()) {
       if (VersionNum < 6 ||
           Triple.getSubArch() == llvm::Triple::SubArchType::ARMSubArch_v6m)
         Features.push_back("+strict-align");
-    } else if (Triple.isOSLinux() || Triple.isOSNaCl() ||
-               Triple.isOSWindows()) {
-      if (VersionNum < 7)
-        Features.push_back("+strict-align");
-    } else
+    } else if (VersionNum < 7 ||
+               Triple.getSubArch() ==
+                   llvm::Triple::SubArchType::ARMSubArch_v6m ||
+               Triple.getSubArch() ==
+                   llvm::Triple::SubArchType::ARMSubArch_v8m_baseline) {
       Features.push_back("+strict-align");
+    }
   }
 
   // llvm does not support reserving registers in general. There is support

diff --git a/clang/lib/Driver/ToolChains/Clang.cpp b/clang/lib/Driver/ToolChains/Clang.cpp
@@ -1111,8 +1111,8 @@ void Clang::AddPreprocessingOptions(Compilation &C, const JobAction &JA,
         C.getActiveOffloadKinds() == Action::OFK_None) {
       SmallString<128> P(llvm::sys::path::parent_path(D.InstalledDir));
       llvm::sys::path::append(P, "include");
-      llvm::sys::path::append(P, "gpu-none-llvm");
-      CmdArgs.push_back("-c-isystem");
+      llvm::sys::path::append(P, getToolChain().getTripleString());
+      CmdArgs.push_back("-internal-isystem");
       CmdArgs.push_back(Args.MakeArgString(P));
     } else if (C.getActiveOffloadKinds() == Action::OFK_OpenMP) {
       // TODO: CUDA / HIP include their own headers for some common functions

diff --git a/clang/lib/Driver/ToolChains/CommonArgs.cpp b/clang/lib/Driver/ToolChains/CommonArgs.cpp
@@ -1087,10 +1087,41 @@ static void addOpenMPDeviceLibC(const ToolChain &TC, const ArgList &Args,
                           "llvm-libc-decls");
   bool HasLibC = llvm::sys::fs::exists(LibCDecls) &&
                  llvm::sys::fs::is_directory(LibCDecls);
-  if (Args.hasFlag(options::OPT_gpulibc, options::OPT_nogpulibc, HasLibC)) {
-    CmdArgs.push_back("-lcgpu");
-    CmdArgs.push_back("-lmgpu");
+  if (!Args.hasFlag(options::OPT_gpulibc, options::OPT_nogpulibc, HasLibC))
+    return;
+
+  // We don't have access to the offloading toolchains here, so determine from
+  // the arguments if we have any active NVPTX or AMDGPU toolchains.
+  llvm::DenseSet<const char *> Libraries;
+  if (const Arg *Targets = Args.getLastArg(options::OPT_fopenmp_targets_EQ)) {
+    if (llvm::any_of(Targets->getValues(),
+                     [](auto S) { return llvm::Triple(S).isAMDGPU(); })) {
+      Libraries.insert("-lcgpu-amdgpu");
+      Libraries.insert("-lmgpu-amdgpu");
+    }
+    if (llvm::any_of(Targets->getValues(),
+                     [](auto S) { return llvm::Triple(S).isNVPTX(); })) {
+      Libraries.insert("-lcgpu-nvptx");
+      Libraries.insert("-lmgpu-nvptx");
+    }
   }
+
+  for (StringRef Arch : Args.getAllArgValues(options::OPT_offload_arch_EQ)) {
+    if (llvm::any_of(llvm::split(Arch, ","), [](StringRef Str) {
+          return IsAMDGpuArch(StringToCudaArch(Str));
+        })) {
+      Libraries.insert("-lcgpu-amdgpu");
+      Libraries.insert("-lmgpu-amdgpu");
+    }
+    if (llvm::any_of(llvm::split(Arch, ","), [](StringRef Str) {
+          return IsNVIDIAGpuArch(StringToCudaArch(Str));
+        })) {
+      Libraries.insert("-lcgpu-nvptx");
+      Libraries.insert("-lmgpu-nvptx");
+    }
+  }
+
+  llvm::append_range(CmdArgs, Libraries);
 }
 
 void tools::addOpenMPRuntimeLibraryPath(const ToolChain &TC,

diff --git a/clang/lib/Driver/ToolChains/Cuda.cpp b/clang/lib/Driver/ToolChains/Cuda.cpp
@@ -609,6 +609,10 @@ void NVPTX::Linker::ConstructJob(Compilation &C, const JobAction &JA,
   // Add paths specified in LIBRARY_PATH environment variable as -L options.
   addDirectoryList(Args, CmdArgs, "-L", "LIBRARY_PATH");
 
+  // Add standard library search paths passed on the command line.
+  Args.AddAllArgs(CmdArgs, options::OPT_L);
+  getToolChain().AddFilePathLibArgs(Args, CmdArgs);
+
   // Add paths for the default clang library path.
   SmallString<256> DefaultLibPath =
       llvm::sys::path::parent_path(TC.getDriver().Dir);
@@ -623,35 +627,34 @@ void NVPTX::Linker::ConstructJob(Compilation &C, const JobAction &JA,
       continue;
     }
 
-    // Currently, we only pass the input files to the linker, we do not pass
-    // any libraries that may be valid only for the host.
-    if (!II.isFilename())
-      continue;
-
     // The 'nvlink' application performs RDC-mode linking when given a '.o'
     // file and device linking when given a '.cubin' file. We always want to
     // perform device linking, so just rename any '.o' files.
     // FIXME: This should hopefully be removed if NVIDIA updates their tooling.
-    auto InputFile = getToolChain().getInputFilename(II);
-    if (llvm::sys::path::extension(InputFile) != ".cubin") {
-      // If there are no actions above this one then this is direct input and we
-      // can copy it. Otherwise the input is internal so a `.cubin` file should
-      // exist.
-      if (II.getAction() && II.getAction()->getInputs().size() == 0) {
-        const char *CubinF =
-            Args.MakeArgString(getToolChain().getDriver().GetTemporaryPath(
-                llvm::sys::path::stem(InputFile), "cubin"));
-        if (llvm::sys::fs::copy_file(InputFile, C.addTempFile(CubinF)))
-          continue;
+    if (II.isFilename()) {
+      auto InputFile = getToolChain().getInputFilename(II);
+      if (llvm::sys::path::extension(InputFile) != ".cubin") {
+        // If there are no actions above this one then this is direct input and
+        // we can copy it. Otherwise the input is internal so a `.cubin` file
+        // should exist.
+        if (II.getAction() && II.getAction()->getInputs().size() == 0) {
+          const char *CubinF =
+              Args.MakeArgString(getToolChain().getDriver().GetTemporaryPath(
+                  llvm::sys::path::stem(InputFile), "cubin"));
+          if (llvm::sys::fs::copy_file(InputFile, C.addTempFile(CubinF)))
+            continue;
 
-        CmdArgs.push_back(CubinF);
+          CmdArgs.push_back(CubinF);
+        } else {
+          SmallString<256> Filename(InputFile);
+          llvm::sys::path::replace_extension(Filename, "cubin");
+          CmdArgs.push_back(Args.MakeArgString(Filename));
+        }
       } else {
-        SmallString<256> Filename(InputFile);
-        llvm::sys::path::replace_extension(Filename, "cubin");
-        CmdArgs.push_back(Args.MakeArgString(Filename));
+        CmdArgs.push_back(Args.MakeArgString(InputFile));
       }
-    } else {
-      CmdArgs.push_back(Args.MakeArgString(InputFile));
+    } else if (!II.isNothing()) {
+      II.getInputArg().renderAsInput(Args, CmdArgs);
     }
   }
 

diff --git a/clang/lib/Driver/ToolChains/HIPUtility.cpp b/clang/lib/Driver/ToolChains/HIPUtility.cpp
@@ -9,13 +9,24 @@
 #include "HIPUtility.h"
 #include "CommonArgs.h"
 #include "clang/Driver/Compilation.h"
+#include "clang/Driver/Options.h"
+#include "llvm/ADT/StringExtras.h"
 #include "llvm/ADT/StringRef.h"
+#include "llvm/Object/Archive.h"
+#include "llvm/Object/ObjectFile.h"
+#include "llvm/Support/MD5.h"
+#include "llvm/Support/MemoryBuffer.h"
 #include "llvm/Support/Path.h"
+#include "llvm/Support/raw_ostream.h"
 #include "llvm/TargetParser/Triple.h"
+#include <deque>
+#include <set>
 
+using namespace clang;
 using namespace clang::driver;
 using namespace clang::driver::tools;
 using namespace llvm::opt;
+using llvm::dyn_cast;
 
 #if defined(_WIN32) || defined(_WIN64)
 #define NULL_FILE "nul"
@@ -36,6 +47,169 @@ static std::string normalizeForBundler(const llvm::Triple &T,
                      : T.normalize();
 }
 
+// Collect undefined __hip_fatbin* and __hip_gpubin_handle* symbols from all
+// input object or archive files.
+class HIPUndefinedFatBinSymbols {
+public:
+  HIPUndefinedFatBinSymbols(const Compilation &C)
+      : C(C), DiagID(C.getDriver().getDiags().getCustomDiagID(
+                  DiagnosticsEngine::Error,
+                  "Error collecting HIP undefined fatbin symbols: %0")),
+        Quiet(C.getArgs().hasArg(options::OPT__HASH_HASH_HASH)),
+        Verbose(C.getArgs().hasArg(options::OPT_v)) {
+    populateSymbols();
+    if (Verbose) {
+      for (auto Name : FatBinSymbols)
+        llvm::errs() << "Found undefined HIP fatbin symbol: " << Name << "\n";
+      for (auto Name : GPUBinHandleSymbols)
+        llvm::errs() << "Found undefined HIP gpubin handle symbol: " << Name
+                     << "\n";
+    }
+  }
+
+  const std::set<std::string> &getFatBinSymbols() const {
+    return FatBinSymbols;
+  }
+
+  const std::set<std::string> &getGPUBinHandleSymbols() const {
+    return GPUBinHandleSymbols;
+  }
+
+private:
+  const Compilation &C;
+  unsigned DiagID;
+  bool Quiet;
+  bool Verbose;
+  std::set<std::string> FatBinSymbols;
+  std::set<std::string> GPUBinHandleSymbols;
+  std::set<std::string> DefinedFatBinSymbols;
+  std::set<std::string> DefinedGPUBinHandleSymbols;
+  const std::string FatBinPrefix = "__hip_fatbin";
+  const std::string GPUBinHandlePrefix = "__hip_gpubin_handle";
+
+  void populateSymbols() {
+    std::deque<const Action *> WorkList;
+    std::set<const Action *> Visited;
+
+    for (const auto &Action : C.getActions())
+      WorkList.push_back(Action);
+
+    while (!WorkList.empty()) {
+      const Action *CurrentAction = WorkList.front();
+      WorkList.pop_front();
+
+      if (!CurrentAction || !Visited.insert(CurrentAction).second)
+        continue;
+
+      if (const auto *IA = dyn_cast<InputAction>(CurrentAction)) {
+        std::string ID = IA->getId().str();
+        if (!ID.empty()) {
+          ID = llvm::utohexstr(llvm::MD5Hash(ID), /*LowerCase=*/true);
+          FatBinSymbols.insert(Twine(FatBinPrefix + "_" + ID).str());
+          GPUBinHandleSymbols.insert(
+              Twine(GPUBinHandlePrefix + "_" + ID).str());
+          continue;
+        }
+        if (IA->getInputArg().getNumValues() == 0)
+          continue;
+        const char *Filename = IA->getInputArg().getValue();
+        if (!Filename)
+          continue;
+        auto BufferOrErr = llvm::MemoryBuffer::getFile(Filename);
+        // Input action could be options to linker, therefore, ignore it
+        // if cannot read it. If it turns out to be a file that cannot be read,
+        // the error will be caught by the linker.
+        if (!BufferOrErr)
+          continue;
+
+        processInput(BufferOrErr.get()->getMemBufferRef());
+      } else
+        WorkList.insert(WorkList.end(), CurrentAction->getInputs().begin(),
+                        CurrentAction->getInputs().end());
+    }
+  }
+
+  void processInput(const llvm::MemoryBufferRef &Buffer) {
+    // Try processing as object file first.
+    auto ObjFileOrErr = llvm::object::ObjectFile::createObjectFile(Buffer);
+    if (ObjFileOrErr) {
+      processSymbols(**ObjFileOrErr);
+      return;
+    }
+
+    // Then try processing as archive files.
+    llvm::consumeError(ObjFileOrErr.takeError());
+    auto ArchiveOrErr = llvm::object::Archive::create(Buffer);
+    if (ArchiveOrErr) {
+      llvm::Error Err = llvm::Error::success();
+      llvm::object::Archive &Archive = *ArchiveOrErr.get();
+      for (auto &Child : Archive.children(Err)) {
+        auto ChildBufOrErr = Child.getMemoryBufferRef();
+        if (ChildBufOrErr)
+          processInput(*ChildBufOrErr);
+        else
+          errorHandler(ChildBufOrErr.takeError());
+      }
+
+      if (Err)
+        errorHandler(std::move(Err));
+      return;
+    }
+
+    // Ignore other files.
+    llvm::consumeError(ArchiveOrErr.takeError());
+  }
+
+  void processSymbols(const llvm::object::ObjectFile &Obj) {
+    for (const auto &Symbol : Obj.symbols()) {
+      auto FlagOrErr = Symbol.getFlags();
+      if (!FlagOrErr) {
+        errorHandler(FlagOrErr.takeError());
+        continue;
+      }
+
+      auto NameOrErr = Symbol.getName();
+      if (!NameOrErr) {
+        errorHandler(NameOrErr.takeError());
+        continue;
+      }
+      llvm::StringRef Name = *NameOrErr;
+
+      bool isUndefined =
+          FlagOrErr.get() & llvm::object::SymbolRef::SF_Undefined;
+      bool isFatBinSymbol = Name.starts_with(FatBinPrefix);
+      bool isGPUBinHandleSymbol = Name.starts_with(GPUBinHandlePrefix);
+
+      // Handling for defined symbols
+      if (!isUndefined) {
+        if (isFatBinSymbol) {
+          DefinedFatBinSymbols.insert(Name.str());
+          FatBinSymbols.erase(Name.str());
+        } else if (isGPUBinHandleSymbol) {
+          DefinedGPUBinHandleSymbols.insert(Name.str());
+          GPUBinHandleSymbols.erase(Name.str());
+        }
+        continue;
+      }
+
+      // Add undefined symbols if they are not in the defined sets
+      if (isFatBinSymbol &&
+          DefinedFatBinSymbols.find(Name.str()) == DefinedFatBinSymbols.end())
+        FatBinSymbols.insert(Name.str());
+      else if (isGPUBinHandleSymbol &&
+               DefinedGPUBinHandleSymbols.find(Name.str()) ==
+                   DefinedGPUBinHandleSymbols.end())
+        GPUBinHandleSymbols.insert(Name.str());
+    }
+  }
+
+  void errorHandler(llvm::Error Err) {
+    if (Quiet)
+      return;
+    C.getDriver().Diag(DiagID) << llvm::toString(std::move(Err));
+  }
+};
+
 // Construct a clang-offload-bundler command to bundle code objects for
 // different devices into a HIP fat binary.
 void HIP::constructHIPFatbinCommand(Compilation &C, const JobAction &JA,
@@ -130,26 +304,84 @@ void HIP::constructGenerateObjFileFromHIPFatBinary(
   auto HostTriple =
       C.getSingleOffloadToolChain<Action::OFK_Host>()->getTriple();
 
+  HIPUndefinedFatBinSymbols Symbols(C);
+
+  std::string PrimaryHipFatbinSymbol;
+  std::string PrimaryGpuBinHandleSymbol;
+  bool FoundPrimaryHipFatbinSymbol = false;
+  bool FoundPrimaryGpuBinHandleSymbol = false;
+
+  std::vector<std::string> AliasHipFatbinSymbols;
+  std::vector<std::string> AliasGpuBinHandleSymbols;
+
+  // Iterate through symbols to find the primary ones and collect others for
+  // aliasing
+  for (const auto &Symbol : Symbols.getFatBinSymbols()) {
+    if (!FoundPrimaryHipFatbinSymbol) {
+      PrimaryHipFatbinSymbol = Symbol;
+      FoundPrimaryHipFatbinSymbol = true;
+    } else
+      AliasHipFatbinSymbols.push_back(Symbol);
+  }
+
+  for (const auto &Symbol : Symbols.getGPUBinHandleSymbols()) {
+    if (!FoundPrimaryGpuBinHandleSymbol) {
+      PrimaryGpuBinHandleSymbol = Symbol;
+      FoundPrimaryGpuBinHandleSymbol = true;
+    } else
+      AliasGpuBinHandleSymbols.push_back(Symbol);
+  }
+
   // Add MC directives to embed target binaries. We ensure that each
   // section and image is 16-byte aligned. This is not mandatory, but
   // increases the likelihood of data to be aligned with a cache block
   // in several main host machines.
   ObjStream << "#       HIP Object Generator\n";
   ObjStream << "# *** Automatically generated by Clang ***\n";
-  if (HostTriple.isWindowsMSVCEnvironment()) {
-    ObjStream << "  .section .hip_fatbin, \"dw\"\n";
-  } else {
-    ObjStream << "  .protected __hip_fatbin\n";
-    ObjStream << "  .type __hip_fatbin,@object\n";
-    ObjStream << "  .section .hip_fatbin,\"a\",@progbits\n";
+  if (FoundPrimaryGpuBinHandleSymbol) {
+    // Define the first gpubin handle symbol
+    if (HostTriple.isWindowsMSVCEnvironment())
+      ObjStream << "  .section .hip_gpubin_handle,\"dw\"\n";
+    else {
+      ObjStream << "  .protected " << PrimaryGpuBinHandleSymbol << "\n";
+      ObjStream << "  .type " << PrimaryGpuBinHandleSymbol << ",@object\n";
+      ObjStream << "  .section .hip_gpubin_handle,\"aw\"\n";
+    }
+    ObjStream << "  .globl " << PrimaryGpuBinHandleSymbol << "\n";
+    ObjStream << "  .p2align 3\n"; // Align 8
+    ObjStream << PrimaryGpuBinHandleSymbol << ":\n";
+    ObjStream << "  .zero 8\n"; // Size 8
+
+    // Generate alias directives for other gpubin handle symbols
+    for (const auto &AliasSymbol : AliasGpuBinHandleSymbols) {
+      ObjStream << "  .globl " << AliasSymbol << "\n";
+      ObjStream << "  .set " << AliasSymbol << "," << PrimaryGpuBinHandleSymbol
+                << "\n";
+    }
+  }
+  if (FoundPrimaryHipFatbinSymbol) {
+    // Define the first fatbin symbol
+    if (HostTriple.isWindowsMSVCEnvironment())
+      ObjStream << "  .section .hip_fatbin,\"dw\"\n";
+    else {
+      ObjStream << "  .protected " << PrimaryHipFatbinSymbol << "\n";
+      ObjStream << "  .type " << PrimaryHipFatbinSymbol << ",@object\n";
+      ObjStream << "  .section .hip_fatbin,\"a\",@progbits\n";
+    }
+    ObjStream << "  .globl " << PrimaryHipFatbinSymbol << "\n";
+    ObjStream << "  .p2align " << llvm::Log2(llvm::Align(HIPCodeObjectAlign))
+              << "\n";
+    // Generate alias directives for other fatbin symbols
+    for (const auto &AliasSymbol : AliasHipFatbinSymbols) {
+      ObjStream << "  .globl " << AliasSymbol << "\n";
+      ObjStream << "  .set " << AliasSymbol << "," << PrimaryHipFatbinSymbol
+                << "\n";
+    }
+    ObjStream << PrimaryHipFatbinSymbol << ":\n";
+    ObjStream << "  .incbin ";
+    llvm::sys::printArg(ObjStream, BundleFile, /*Quote=*/true);
+    ObjStream << "\n";
   }
-  ObjStream << "  .globl __hip_fatbin\n";
-  ObjStream << "  .p2align " << llvm::Log2(llvm::Align(HIPCodeObjectAlign))
-            << "\n";
-  ObjStream << "__hip_fatbin:\n";
-  ObjStream << "  .incbin ";
-  llvm::sys::printArg(ObjStream, BundleFile, /*Quote=*/true);
-  ObjStream << "\n";
   if (HostTriple.isOSLinux() && HostTriple.isOSBinFormatELF())
     ObjStream << "  .section .note.GNU-stack, \"\", @progbits\n";
   ObjStream.flush();

diff --git a/clang/lib/Driver/ToolChains/RISCVToolchain.cpp b/clang/lib/Driver/ToolChains/RISCVToolchain.cpp
@@ -86,6 +86,11 @@ RISCVToolChain::GetUnwindLibType(const llvm::opt::ArgList &Args) const {
   return ToolChain::UNW_None;
 }
 
+ToolChain::UnwindTableLevel RISCVToolChain::getDefaultUnwindTableLevel(
+    const llvm::opt::ArgList &Args) const {
+  return UnwindTableLevel::None;
+}
+
 void RISCVToolChain::addClangTargetOptions(
     const llvm::opt::ArgList &DriverArgs,
     llvm::opt::ArgStringList &CC1Args,

diff --git a/clang/lib/Driver/ToolChains/RISCVToolchain.h b/clang/lib/Driver/ToolChains/RISCVToolchain.h
@@ -28,6 +28,8 @@ class LLVM_LIBRARY_VISIBILITY RISCVToolChain : public Generic_ELF {
   RuntimeLibType GetDefaultRuntimeLibType() const override;
   UnwindLibType
   GetUnwindLibType(const llvm::opt::ArgList &Args) const override;
+  UnwindTableLevel
+  getDefaultUnwindTableLevel(const llvm::opt::ArgList &Args) const override;
   void
   AddClangSystemIncludeArgs(const llvm::opt::ArgList &DriverArgs,
                             llvm::opt::ArgStringList &CC1Args) const override;

diff --git a/clang/lib/Format/Format.cpp b/clang/lib/Format/Format.cpp
@@ -915,6 +915,8 @@ template <> struct MappingTraits<FormatStyle> {
     IO.mapOptional("AlignConsecutiveMacros", Style.AlignConsecutiveMacros);
     IO.mapOptional("AlignConsecutiveShortCaseStatements",
                    Style.AlignConsecutiveShortCaseStatements);
+    IO.mapOptional("AlignConsecutiveTableGenCondOperatorColons",
+                   Style.AlignConsecutiveTableGenCondOperatorColons);
     IO.mapOptional("AlignEscapedNewlines", Style.AlignEscapedNewlines);
     IO.mapOptional("AlignOperands", Style.AlignOperands);
     IO.mapOptional("AlignTrailingComments", Style.AlignTrailingComments);
@@ -1420,6 +1422,7 @@ FormatStyle getLLVMStyle(FormatStyle::LanguageKind Language) {
   LLVMStyle.AlignConsecutiveDeclarations = {};
   LLVMStyle.AlignConsecutiveMacros = {};
   LLVMStyle.AlignConsecutiveShortCaseStatements = {};
+  LLVMStyle.AlignConsecutiveTableGenCondOperatorColons = {};
   LLVMStyle.AlignEscapedNewlines = FormatStyle::ENAS_Right;
   LLVMStyle.AlignOperands = FormatStyle::OAS_Align;
   LLVMStyle.AlignTrailingComments = {};
@@ -1591,7 +1594,7 @@ FormatStyle getLLVMStyle(FormatStyle::LanguageKind Language) {
   LLVMStyle.PenaltyBreakScopeResolution = 500;
   LLVMStyle.PenaltyBreakString = 1000;
   LLVMStyle.PenaltyBreakTemplateDeclaration = prec::Relational;
-  LLVMStyle.PenaltyExcessCharacter = 1000000;
+  LLVMStyle.PenaltyExcessCharacter = 1'000'000;
   LLVMStyle.PenaltyIndentedWhitespace = 0;
   LLVMStyle.PenaltyReturnTypeOnItsOwnLine = 60;
 
@@ -1914,9 +1917,12 @@ FormatStyle getClangFormatStyle() {
   FormatStyle Style = getLLVMStyle();
   Style.InsertBraces = true;
   Style.InsertNewlineAtEOF = true;
+  Style.IntegerLiteralSeparator.Decimal = 3;
+  Style.IntegerLiteralSeparator.DecimalMinDigits = 5;
   Style.LineEnding = FormatStyle::LE_LF;
   Style.RemoveBracesLLVM = true;
   Style.RemoveParentheses = FormatStyle::RPS_ReturnStatement;
+  Style.RemoveSemicolon = true;
   return Style;
 }
 

diff --git a/clang/lib/Format/FormatToken.cpp b/clang/lib/Format/FormatToken.cpp
@@ -137,7 +137,7 @@ unsigned CommaSeparatedList::formatAfterToken(LineState &State,
   // bin-packed. Add a severe penalty to this so that column layouts are
   // preferred if possible.
   if (!Format)
-    return 10000;
+    return 10'000;
 
   // Format the entire list.
   unsigned Penalty = 0;

diff --git a/clang/lib/Format/MacroCallReconstructor.cpp b/clang/lib/Format/MacroCallReconstructor.cpp
@@ -33,7 +33,7 @@ void forEachToken(const UnwrappedLine &Line, const T &Call,
                   FormatToken *Parent = nullptr) {
   bool First = true;
   for (const auto &N : Line.Tokens) {
-    Call(N.Tok, Parent, First);
+    Call(N.Tok, Parent, First, Line.Level);
     First = false;
     for (const auto &Child : N.Children)
       forEachToken(Child, Call, N.Tok);
@@ -44,26 +44,25 @@ MacroCallReconstructor::MacroCallReconstructor(
     unsigned Level,
     const llvm::DenseMap<FormatToken *, std::unique_ptr<UnwrappedLine>>
         &ActiveExpansions)
-    : Level(Level), IdToReconstructed(ActiveExpansions) {
+    : Result(Level), IdToReconstructed(ActiveExpansions) {
   Result.Tokens.push_back(std::make_unique<LineNode>());
   ActiveReconstructedLines.push_back(&Result);
 }
 
 void MacroCallReconstructor::addLine(const UnwrappedLine &Line) {
   assert(State != Finalized);
   LLVM_DEBUG(llvm::dbgs() << "MCR: new line...\n");
-  forEachToken(Line, [&](FormatToken *Token, FormatToken *Parent, bool First) {
-    add(Token, Parent, First);
-  });
+  forEachToken(Line, [&](FormatToken *Token, FormatToken *Parent, bool First,
+                         unsigned Level) { add(Token, Parent, First, Level); });
   assert(InProgress || finished());
 }
 
 UnwrappedLine MacroCallReconstructor::takeResult() && {
   finalize();
   assert(Result.Tokens.size() == 1 &&
          Result.Tokens.front()->Children.size() == 1);
-  UnwrappedLine Final =
-      createUnwrappedLine(*Result.Tokens.front()->Children.front(), Level);
+  UnwrappedLine Final = createUnwrappedLine(
+      *Result.Tokens.front()->Children.front(), Result.Level);
   assert(!Final.Tokens.empty());
   return Final;
 }
@@ -72,7 +71,8 @@ UnwrappedLine MacroCallReconstructor::takeResult() && {
 // ExpandedParent in the incoming unwrapped line. \p First specifies whether it
 // is the first token in a given unwrapped line.
 void MacroCallReconstructor::add(FormatToken *Token,
-                                 FormatToken *ExpandedParent, bool First) {
+                                 FormatToken *ExpandedParent, bool First,
+                                 unsigned Level) {
   LLVM_DEBUG(
       llvm::dbgs() << "MCR: Token: " << Token->TokenText << ", Parent: "
                    << (ExpandedParent ? ExpandedParent->TokenText : "<null>")
@@ -102,7 +102,7 @@ void MacroCallReconstructor::add(FormatToken *Token,
       First = true;
   }
 
-  prepareParent(ExpandedParent, First);
+  prepareParent(ExpandedParent, First, Level);
 
   if (Token->MacroCtx) {
     // If this token was generated by a macro call, add the reconstructed
@@ -129,7 +129,7 @@ void MacroCallReconstructor::add(FormatToken *Token,
 // is the parent of ActiveReconstructedLines.back() in the reconstructed
 // unwrapped line.
 void MacroCallReconstructor::prepareParent(FormatToken *ExpandedParent,
-                                           bool NewLine) {
+                                           bool NewLine, unsigned Level) {
   LLVM_DEBUG({
     llvm::dbgs() << "ParentMap:\n";
     debugParentMap();
@@ -172,7 +172,7 @@ void MacroCallReconstructor::prepareParent(FormatToken *ExpandedParent,
     }
     assert(!ActiveReconstructedLines.empty());
     ActiveReconstructedLines.back()->Tokens.back()->Children.push_back(
-        std::make_unique<ReconstructedLine>());
+        std::make_unique<ReconstructedLine>(Level));
     ActiveReconstructedLines.push_back(
         &*ActiveReconstructedLines.back()->Tokens.back()->Children.back());
   } else if (parentLine().Tokens.back()->Tok != Parent) {
@@ -424,7 +424,8 @@ bool MacroCallReconstructor::processNextReconstructed() {
       SpelledParentToReconstructedParent[MacroCallStructure.back()
                                              .ParentLastToken] = Token;
       appendToken(Token);
-      prepareParent(Token, /*NewLine=*/true);
+      prepareParent(Token, /*NewLine=*/true,
+                    MacroCallStructure.back().Line->Level);
       Token->MacroParent = true;
       return false;
     }
@@ -435,7 +436,8 @@ bool MacroCallReconstructor::processNextReconstructed() {
             [MacroCallStructure.back().Line->Tokens.back()->Tok] = Token;
         Token->MacroParent = true;
         appendToken(Token, MacroCallStructure.back().Line);
-        prepareParent(Token, /*NewLine=*/true);
+        prepareParent(Token, /*NewLine=*/true,
+                      MacroCallStructure.back().Line->Level);
         return true;
       }
       if (Token->is(tok::r_paren)) {
@@ -509,16 +511,36 @@ MacroCallReconstructor::createUnwrappedLine(const ReconstructedLine &Line,
   for (const auto &N : Line.Tokens) {
     Result.Tokens.push_back(N->Tok);
     UnwrappedLineNode &Current = Result.Tokens.back();
-    for (const auto &Child : N->Children) {
-      if (Child->Tokens.empty())
-        continue;
-      Current.Children.push_back(createUnwrappedLine(*Child, Level + 1));
-    }
-    if (Current.Children.size() == 1 &&
-        Current.Tok->isOneOf(tok::l_paren, tok::comma)) {
-      Result.Tokens.splice(Result.Tokens.end(),
-                           Current.Children.front().Tokens);
-      Current.Children.clear();
+    auto NumChildren =
+        std::count_if(N->Children.begin(), N->Children.end(),
+                      [](const auto &Child) { return !Child->Tokens.empty(); });
+    if (NumChildren == 1 && Current.Tok->isOneOf(tok::l_paren, tok::comma)) {
+      // If we only have one child, and the child is due to a macro expansion
+      // (either attached to a left parenthesis or comma), merge the child into
+      // the current line to prevent forced breaks for macro arguments.
+      auto *Child = std::find_if(
+          N->Children.begin(), N->Children.end(),
+          [](const auto &Child) { return !Child->Tokens.empty(); });
+      auto Line = createUnwrappedLine(**Child, Level);
+      Result.Tokens.splice(Result.Tokens.end(), Line.Tokens);
+    } else if (NumChildren > 0) {
+      // When there are multiple children with different indent, make sure that
+      // we indent them:
+      // 1. One level below the current line's level.
+      // 2. At the correct level relative to each other.
+      unsigned MinChildLevel =
+          std::min_element(N->Children.begin(), N->Children.end(),
+                           [](const auto &E1, const auto &E2) {
+                             return E1->Level < E2->Level;
+                           })
+              ->get()
+              ->Level;
+      for (const auto &Child : N->Children) {
+        if (Child->Tokens.empty())
+          continue;
+        Current.Children.push_back(createUnwrappedLine(
+            *Child, Level + 1 + (Child->Level - MinChildLevel)));
+      }
     }
   }
   return Result;

diff --git a/clang/lib/Format/Macros.h b/clang/lib/Format/Macros.h
@@ -231,8 +231,9 @@ class MacroCallReconstructor {
   UnwrappedLine takeResult() &&;
 
 private:
-  void add(FormatToken *Token, FormatToken *ExpandedParent, bool First);
-  void prepareParent(FormatToken *ExpandedParent, bool First);
+  void add(FormatToken *Token, FormatToken *ExpandedParent, bool First,
+           unsigned Level);
+  void prepareParent(FormatToken *ExpandedParent, bool First, unsigned Level);
   FormatToken *getParentInResult(FormatToken *Parent);
   void reconstruct(FormatToken *Token);
   void startReconstruction(FormatToken *Token);
@@ -272,6 +273,8 @@ class MacroCallReconstructor {
   // FIXME: Investigate changing UnwrappedLine to a pointer type and using it
   // instead of rolling our own type.
   struct ReconstructedLine {
+    explicit ReconstructedLine(unsigned Level) : Level(Level) {}
+    unsigned Level;
     llvm::SmallVector<std::unique_ptr<LineNode>> Tokens;
   };
 
@@ -373,9 +376,6 @@ class MacroCallReconstructor {
   // \- )
   llvm::SmallVector<MacroCallState> MacroCallStructure;
 
-  // Level the generated UnwrappedLine will be at.
-  const unsigned Level;
-
   // Maps from identifier of the macro call to an unwrapped line containing
   // all tokens of the macro call.
   const llvm::DenseMap<FormatToken *, std::unique_ptr<UnwrappedLine>>

diff --git a/clang/lib/Format/TokenAnnotator.cpp b/clang/lib/Format/TokenAnnotator.cpp
@@ -3817,7 +3817,7 @@ void TokenAnnotator::calculateFormattingInformation(AnnotatedLine &Line) const {
         do {
           Tok = Tok->Next;
         } while (Tok && Tok->isNot(TT_OverloadedOperatorLParen));
-        if (!Tok)
+        if (!Tok || !Tok->MatchingParen)
           break;
         const auto *LeftParen = Tok;
         for (Tok = Tok->Next; Tok && Tok != LeftParen->MatchingParen;

diff --git a/clang/lib/Format/UnwrappedLineFormatter.cpp b/clang/lib/Format/UnwrappedLineFormatter.cpp
@@ -1221,7 +1221,7 @@ class OptimizingLineFormatter : public LineFormatter {
     // While not empty, take first element and follow edges.
     while (!Queue.empty()) {
       // Quit if we still haven't found a solution by now.
-      if (Count > 25000000)
+      if (Count > 25'000'000)
         return 0;
 
       Penalty = Queue.top().first.first;
@@ -1235,7 +1235,7 @@ class OptimizingLineFormatter : public LineFormatter {
 
       // Cut off the analysis of certain solutions if the analysis gets too
       // complex. See description of IgnoreStackForComparison.
-      if (Count > 50000)
+      if (Count > 50'000)
         Node->State.IgnoreStackForComparison = true;
 
       if (!Seen.insert(&Node->State).second) {

diff --git a/clang/lib/Format/UnwrappedLineParser.cpp b/clang/lib/Format/UnwrappedLineParser.cpp
@@ -90,6 +90,12 @@ class ScopedDeclarationState {
 
 } // end anonymous namespace
 
+std::ostream &operator<<(std::ostream &Stream, const UnwrappedLine &Line) {
+  llvm::raw_os_ostream OS(Stream);
+  printLine(OS, Line);
+  return Stream;
+}
+
 class ScopedLineState {
 public:
   ScopedLineState(UnwrappedLineParser &Parser,

diff --git a/clang/lib/Format/UnwrappedLineParser.h b/clang/lib/Format/UnwrappedLineParser.h
@@ -420,6 +420,8 @@ struct UnwrappedLineNode {
   SmallVector<UnwrappedLine, 0> Children;
 };
 
+std::ostream &operator<<(std::ostream &Stream, const UnwrappedLine &Line);
+
 } // end namespace format
 } // end namespace clang
 

diff --git a/clang/lib/Format/WhitespaceManager.cpp b/clang/lib/Format/WhitespaceManager.cpp
@@ -111,6 +111,8 @@ const tooling::Replacements &WhitespaceManager::generateReplacements() {
   alignConsecutiveDeclarations();
   alignConsecutiveBitFields();
   alignConsecutiveAssignments();
+  if (Style.isTableGen())
+    alignConsecutiveTableGenCondOperatorColons();
   alignChainedConditionals();
   alignTrailingComments();
   alignEscapedNewlines();
@@ -849,7 +851,12 @@ void WhitespaceManager::alignConsecutiveAssignments() {
 }
 
 void WhitespaceManager::alignConsecutiveBitFields() {
-  if (!Style.AlignConsecutiveBitFields.Enabled)
+  alignConsecutiveColons(Style.AlignConsecutiveBitFields, TT_BitFieldColon);
+}
+
+void WhitespaceManager::alignConsecutiveColons(
+    const FormatStyle::AlignConsecutiveStyle &AlignStyle, TokenType Type) {
+  if (!AlignStyle.Enabled)
     return;
 
   AlignTokens(
@@ -863,9 +870,9 @@ void WhitespaceManager::alignConsecutiveBitFields() {
         if (&C != &Changes.back() && (&C + 1)->NewlinesBefore > 0)
           return false;
 
-        return C.Tok->is(TT_BitFieldColon);
+        return C.Tok->is(Type);
       },
-      Changes, /*StartAt=*/0, Style.AlignConsecutiveBitFields);
+      Changes, /*StartAt=*/0, AlignStyle);
 }
 
 void WhitespaceManager::alignConsecutiveShortCaseStatements() {
@@ -972,6 +979,11 @@ void WhitespaceManager::alignConsecutiveShortCaseStatements() {
                              Changes);
 }
 
+void WhitespaceManager::alignConsecutiveTableGenCondOperatorColons() {
+  alignConsecutiveColons(Style.AlignConsecutiveTableGenCondOperatorColons,
+                         TT_TableGenCondOperatorColon);
+}
+
 void WhitespaceManager::alignConsecutiveDeclarations() {
   if (!Style.AlignConsecutiveDeclarations.Enabled)
     return;

diff --git a/clang/lib/Format/WhitespaceManager.h b/clang/lib/Format/WhitespaceManager.h
@@ -226,6 +226,11 @@ class WhitespaceManager {
   /// Align consecutive bitfields over all \c Changes.
   void alignConsecutiveBitFields();
 
+  /// Align consecutive colon. For bitfields, TableGen DAGArgs and defintions.
+  void
+  alignConsecutiveColons(const FormatStyle::AlignConsecutiveStyle &AlignStyle,
+                         TokenType Type);
+
   /// Align consecutive declarations over all \c Changes.
   void alignConsecutiveDeclarations();
 
@@ -235,6 +240,9 @@ class WhitespaceManager {
   /// Align consecutive short case statements over all \c Changes.
   void alignConsecutiveShortCaseStatements();
 
+  /// Align consecutive TableGen cond operator colon over all \c Changes.
+  void alignConsecutiveTableGenCondOperatorColons();
+
   /// Align trailing comments over all \c Changes.
   void alignTrailingComments();
 

diff --git a/clang/lib/Frontend/ASTUnit.cpp b/clang/lib/Frontend/ASTUnit.cpp
@@ -540,7 +540,17 @@ class ASTInfoCollector : public ASTReaderListener {
     if (InitializedLanguage)
       return false;
 
+    // FIXME: We did similar things in ReadHeaderSearchOptions too. But such
+    // style is not scaling. Probably we need to invite some mechanism to
+    // handle such patterns generally.
+    auto PICLevel = LangOpt.PICLevel;
+    auto PIE = LangOpt.PIE;
+
     LangOpt = LangOpts;
+
+    LangOpt.PICLevel = PICLevel;
+    LangOpt.PIE = PIE;
+
     InitializedLanguage = true;
 
     updated();
@@ -790,7 +800,8 @@ std::unique_ptr<ASTUnit> ASTUnit::LoadFromASTFile(
     const std::string &Filename, const PCHContainerReader &PCHContainerRdr,
     WhatToLoad ToLoad, IntrusiveRefCntPtr<DiagnosticsEngine> Diags,
     const FileSystemOptions &FileSystemOpts,
-    std::shared_ptr<HeaderSearchOptions> HSOpts, bool OnlyLocalDecls,
+    std::shared_ptr<HeaderSearchOptions> HSOpts,
+    std::shared_ptr<LangOptions> LangOpts, bool OnlyLocalDecls,
     CaptureDiagsKind CaptureDiagnostics, bool AllowASTWithCompilerErrors,
     bool UserFilesAreVolatile, IntrusiveRefCntPtr<llvm::vfs::FileSystem> VFS) {
   std::unique_ptr<ASTUnit> AST(new ASTUnit(true));
@@ -804,7 +815,7 @@ std::unique_ptr<ASTUnit> ASTUnit::LoadFromASTFile(
 
   ConfigureDiags(Diags, *AST, CaptureDiagnostics);
 
-  AST->LangOpts = std::make_shared<LangOptions>();
+  AST->LangOpts = LangOpts ? LangOpts : std::make_shared<LangOptions>();
   AST->OnlyLocalDecls = OnlyLocalDecls;
   AST->CaptureDiagnostics = CaptureDiagnostics;
   AST->Diagnostics = Diags;

diff --git a/clang/lib/Frontend/FrontendAction.cpp b/clang/lib/Frontend/FrontendAction.cpp
@@ -689,7 +689,7 @@ bool FrontendAction::BeginSourceFile(CompilerInstance &CI,
     std::unique_ptr<ASTUnit> AST = ASTUnit::LoadFromASTFile(
         std::string(InputFile), CI.getPCHContainerReader(),
         ASTUnit::LoadEverything, Diags, CI.getFileSystemOpts(),
-        CI.getHeaderSearchOptsPtr());
+        CI.getHeaderSearchOptsPtr(), CI.getLangOptsPtr());
 
     if (!AST)
       return false;

diff --git a/clang/lib/Headers/arm_acle.h b/clang/lib/Headers/arm_acle.h
@@ -313,7 +313,7 @@ __qdbl(int32_t __t) {
 }
 #endif
 
-/* 8.4.3 Accumultating multiplications */
+/* 8.4.3 Accumulating multiplications */
 #if defined(__ARM_FEATURE_DSP) && __ARM_FEATURE_DSP
 static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
 __smlabb(int32_t __a, int32_t __b, int32_t __c) {
@@ -545,7 +545,7 @@ __usub16(uint16x2_t __a, uint16x2_t __b) {
 }
 #endif
 
-/* 8.5.10 Parallel 16-bit multiplications */
+/* 8.5.10 Parallel 16-bit multiplication */
 #if defined(__ARM_FEATURE_SIMD32) && __ARM_FEATURE_SIMD32
 static __inline__ int32_t __attribute__((__always_inline__, __nodebug__))
 __smlad(int16x2_t __a, int16x2_t __b, int32_t __c) {
@@ -748,7 +748,7 @@ __arm_st64bv0(void *__addr, data512_t __value) {
 #define __arm_wsrf(sysreg, v) __arm_wsr(sysreg, __builtin_bit_cast(uint32_t, v))
 #define __arm_wsrf64(sysreg, v) __arm_wsr64(sysreg, __builtin_bit_cast(uint64_t, v))
 
-/* 10.3 Memory Tagging Extensions (MTE) Intrinsics */
+/* 10.3 MTE intrinsics */
 #if defined(__ARM_64BIT_STATE) && __ARM_64BIT_STATE
 #define __arm_mte_create_random_tag(__ptr, __mask)  __builtin_arm_irg(__ptr, __mask)
 #define __arm_mte_increment_tag(__ptr, __tag_offset)  __builtin_arm_addg(__ptr, __tag_offset)
@@ -757,7 +757,7 @@ __arm_st64bv0(void *__addr, data512_t __value) {
 #define __arm_mte_set_tag(__ptr) __builtin_arm_stg(__ptr)
 #define __arm_mte_ptrdiff(__ptra, __ptrb) __builtin_arm_subp(__ptra, __ptrb)
 
-/* 18 Memory Operations Intrinsics */
+/* 18 memcpy family of operations intrinsics - MOPS */
 #define __arm_mops_memset_tag(__tagged_address, __value, __size)    \
   __builtin_arm_mops_memset_tag(__tagged_address, __value, __size)
 #endif

diff --git a/clang/lib/Headers/cpuid.h b/clang/lib/Headers/cpuid.h
@@ -219,6 +219,7 @@
 #define bit_PREFETCHI     0x00004000
 #define bit_USERMSR       0x00008000
 #define bit_AVX10         0x00080000
+#define bit_APXF          0x00200000
 
 /* Features in %eax for leaf 13 sub-leaf 1 */
 #define bit_XSAVEOPT    0x00000001

diff --git a/clang/lib/Headers/hlsl/hlsl_intrinsics.h b/clang/lib/Headers/hlsl/hlsl_intrinsics.h
@@ -179,6 +179,104 @@ double3 cos(double3);
 _HLSL_BUILTIN_ALIAS(__builtin_elementwise_cos)
 double4 cos(double4);
 
+//===----------------------------------------------------------------------===//
+// dot product builtins
+//===----------------------------------------------------------------------===//
+
+/// \fn K dot(T X, T Y)
+/// \brief Return the dot product (a scalar value) of \a X and \a Y.
+/// \param X The X input value.
+/// \param Y The Y input value.
+
+_HLSL_16BIT_AVAILABILITY(shadermodel, 6.2)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+half dot(half, half);
+_HLSL_16BIT_AVAILABILITY(shadermodel, 6.2)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+half dot(half2, half2);
+_HLSL_16BIT_AVAILABILITY(shadermodel, 6.2)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+half dot(half3, half3);
+_HLSL_16BIT_AVAILABILITY(shadermodel, 6.2)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+half dot(half4, half4);
+
+#ifdef __HLSL_ENABLE_16_BIT
+_HLSL_AVAILABILITY(shadermodel, 6.2)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+int16_t dot(int16_t, int16_t);
+_HLSL_AVAILABILITY(shadermodel, 6.2)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+int16_t dot(int16_t2, int16_t2);
+_HLSL_AVAILABILITY(shadermodel, 6.2)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+int16_t dot(int16_t3, int16_t3);
+_HLSL_AVAILABILITY(shadermodel, 6.2)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+int16_t dot(int16_t4, int16_t4);
+
+_HLSL_AVAILABILITY(shadermodel, 6.2)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+uint16_t dot(uint16_t, uint16_t);
+_HLSL_AVAILABILITY(shadermodel, 6.2)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+uint16_t dot(uint16_t2, uint16_t2);
+_HLSL_AVAILABILITY(shadermodel, 6.2)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+uint16_t dot(uint16_t3, uint16_t3);
+_HLSL_AVAILABILITY(shadermodel, 6.2)
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+uint16_t dot(uint16_t4, uint16_t4);
+#endif
+
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+float dot(float, float);
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+float dot(float2, float2);
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+float dot(float3, float3);
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+float dot(float4, float4);
+
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+double dot(double, double);
+
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+int dot(int, int);
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+int dot(int2, int2);
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+int dot(int3, int3);
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+int dot(int4, int4);
+
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+uint dot(uint, uint);
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+uint dot(uint2, uint2);
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+uint dot(uint3, uint3);
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+uint dot(uint4, uint4);
+
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+int64_t dot(int64_t, int64_t);
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+int64_t dot(int64_t2, int64_t2);
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+int64_t dot(int64_t3, int64_t3);
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+int64_t dot(int64_t4, int64_t4);
+
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+uint64_t dot(uint64_t, uint64_t);
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+uint64_t dot(uint64_t2, uint64_t2);
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+uint64_t dot(uint64_t3, uint64_t3);
+_HLSL_BUILTIN_ALIAS(__builtin_hlsl_dot)
+uint64_t dot(uint64_t4, uint64_t4);
+
 //===----------------------------------------------------------------------===//
 // floor builtins
 //===----------------------------------------------------------------------===//

diff --git a/clang/lib/Parse/ParseExprCXX.cpp b/clang/lib/Parse/ParseExprCXX.cpp
@@ -1385,6 +1385,16 @@ ExprResult Parser::ParseLambdaExpressionAfterIntroducer(
       Diag(RAngleLoc,
            diag::err_lambda_template_parameter_list_empty);
     } else {
+      // We increase the template depth before recursing into a requires-clause.
+      //
+      // This depth is used for setting up a LambdaScopeInfo (in
+      // Sema::RecordParsingTemplateParameterDepth), which is used later when
+      // inventing template parameters in InventTemplateParameter.
+      //
+      // This way, abbreviated generic lambdas could have different template
+      // depths, avoiding substitution into the wrong template parameters during
+      // constraint satisfaction check.
+      ++CurTemplateDepthTracker;
       ExprResult RequiresClause;
       if (TryConsumeToken(tok::kw_requires)) {
         RequiresClause =
@@ -1396,7 +1406,6 @@ ExprResult Parser::ParseLambdaExpressionAfterIntroducer(
 
       Actions.ActOnLambdaExplicitTemplateParameterList(
           Intro, LAngleLoc, TemplateParams, RAngleLoc, RequiresClause);
-      ++CurTemplateDepthTracker;
     }
   }
 

diff --git a/clang/lib/Parse/ParseOpenACC.cpp b/clang/lib/Parse/ParseOpenACC.cpp
@@ -560,6 +560,21 @@ bool doesDirectiveHaveAssociatedStmt(OpenACCDirectiveKind DirKind) {
   llvm_unreachable("Unhandled directive->assoc stmt");
 }
 
+unsigned getOpenACCScopeFlags(OpenACCDirectiveKind DirKind) {
+  switch (DirKind) {
+  case OpenACCDirectiveKind::Parallel:
+    // Mark this as a BreakScope/ContinueScope as well as a compute construct
+    // so that we can diagnose trying to 'break'/'continue' inside of one.
+    return Scope::BreakScope | Scope::ContinueScope |
+           Scope::OpenACCComputeConstructScope;
+  case OpenACCDirectiveKind::Invalid:
+    llvm_unreachable("Shouldn't be creating a scope for an invalid construct");
+  default:
+    break;
+  }
+  return 0;
+}
+
 } // namespace
 
 // OpenACC 3.3, section 1.7:
@@ -1228,6 +1243,8 @@ StmtResult Parser::ParseOpenACCDirectiveStmt() {
 
   if (doesDirectiveHaveAssociatedStmt(DirInfo.DirKind)) {
     ParsingOpenACCDirectiveRAII DirScope(*this, /*Value=*/false);
+    ParseScope ACCScope(this, getOpenACCScopeFlags(DirInfo.DirKind));
+
     AssocStmt = getActions().ActOnOpenACCAssociatedStmt(DirInfo.DirKind,
                                                         ParseStatement());
   }

diff --git a/clang/lib/Sema/CMakeLists.txt b/clang/lib/Sema/CMakeLists.txt
@@ -27,6 +27,7 @@ add_clang_library(clangSema
   Sema.cpp
   SemaAccess.cpp
   SemaAttr.cpp
+  SemaAPINotes.cpp
   SemaAvailability.cpp
   SemaCXXScopeSpec.cpp
   SemaCast.cpp

diff --git a/clang/lib/Sema/Scope.cpp b/clang/lib/Sema/Scope.cpp
@@ -225,6 +225,7 @@ void Scope::dumpImpl(raw_ostream &OS) const {
       {CompoundStmtScope, "CompoundStmtScope"},
       {ClassInheritanceScope, "ClassInheritanceScope"},
       {CatchScope, "CatchScope"},
+      {OpenACCComputeConstructScope, "OpenACCComputeConstructScope"},
   };
 
   for (auto Info : FlagInfo) {

diff --git a/clang/lib/Sema/SemaAPINotes.cpp b/clang/lib/Sema/SemaAPINotes.cpp
diff --git a/clang/lib/Sema/SemaChecking.cpp b/clang/lib/Sema/SemaChecking.cpp
@@ -2120,10 +2120,11 @@ bool Sema::CheckTSBuiltinFunctionCall(const TargetInfo &TI, unsigned BuiltinID,
 // not a valid type, emit an error message and return true. Otherwise return
 // false.
 static bool checkMathBuiltinElementType(Sema &S, SourceLocation Loc,
-                                        QualType Ty) {
-  if (!Ty->getAs<VectorType>() && !ConstantMatrixType::isValidElementType(Ty)) {
+                                        QualType ArgTy, int ArgIndex) {
+  if (!ArgTy->getAs<VectorType>() &&
+      !ConstantMatrixType::isValidElementType(ArgTy)) {
     return S.Diag(Loc, diag::err_builtin_invalid_arg_type)
-           << 1 << /* vector, integer or float ty*/ 0 << Ty;
+           << ArgIndex << /* vector, integer or float ty*/ 0 << ArgTy;
   }
 
   return false;
@@ -2165,7 +2166,10 @@ static bool SemaBuiltinCpu(Sema &S, const TargetInfo &TI, CallExpr *TheCall,
     return S.Diag(TheCall->getBeginLoc(), diag::err_builtin_target_unsupported)
            << SourceRange(TheCall->getBeginLoc(), TheCall->getEndLoc());
   if (!IsCPUSupports && !TheTI->supportsCpuIs())
-    return S.Diag(TheCall->getBeginLoc(), diag::err_builtin_target_unsupported)
+    return S.Diag(TheCall->getBeginLoc(),
+                  TI.getTriple().isOSAIX()
+                      ? diag::err_builtin_aix_os_unsupported
+                      : diag::err_builtin_target_unsupported)
            << SourceRange(TheCall->getBeginLoc(), TheCall->getEndLoc());
 
   Expr *Arg = TheCall->getArg(0)->IgnoreParenImpCasts();
@@ -2958,6 +2962,9 @@ Sema::CheckBuiltinFunctionCall(FunctionDecl *FDecl, unsigned BuiltinID,
   }
   }
 
+  if (getLangOpts().HLSL && CheckHLSLBuiltinFunctionCall(BuiltinID, TheCall))
+    return ExprError();
+
   // Since the target specific builtins for each arch overlap, only check those
   // of the arch we are compiling for.
   if (Context.BuiltinInfo.isTSBuiltin(BuiltinID)) {
@@ -5158,6 +5165,70 @@ bool Sema::CheckPPCMMAType(QualType Type, SourceLocation TypeLoc) {
   return false;
 }
 
+// Helper function for CheckHLSLBuiltinFunctionCall
+bool CheckVectorElementCallArgs(Sema *S, CallExpr *TheCall) {
+  assert(TheCall->getNumArgs() > 1);
+  ExprResult A = TheCall->getArg(0);
+  ExprResult B = TheCall->getArg(1);
+  QualType ArgTyA = A.get()->getType();
+  QualType ArgTyB = B.get()->getType();
+  auto *VecTyA = ArgTyA->getAs<VectorType>();
+  auto *VecTyB = ArgTyB->getAs<VectorType>();
+  SourceLocation BuiltinLoc = TheCall->getBeginLoc();
+  if (VecTyA == nullptr && VecTyB == nullptr)
+    return false;
+
+  if (VecTyA && VecTyB) {
+    bool retValue = false;
+    if (VecTyA->getElementType() != VecTyB->getElementType()) {
+      // Note: type promotion is intended to be handeled via the intrinsics
+      //  and not the builtin itself.
+      S->Diag(TheCall->getBeginLoc(), diag::err_vec_builtin_incompatible_vector)
+          << TheCall->getDirectCallee()
+          << SourceRange(A.get()->getBeginLoc(), B.get()->getEndLoc());
+      retValue = true;
+    }
+    if (VecTyA->getNumElements() != VecTyB->getNumElements()) {
+      // if we get here a HLSLVectorTruncation is needed.
+      S->Diag(BuiltinLoc, diag::err_vec_builtin_incompatible_vector)
+          << TheCall->getDirectCallee()
+          << SourceRange(TheCall->getArg(0)->getBeginLoc(),
+                         TheCall->getArg(1)->getEndLoc());
+      retValue = true;
+    }
+
+    if (retValue)
+      TheCall->setType(VecTyA->getElementType());
+
+    return retValue;
+  }
+
+  // Note: if we get here one of the args is a scalar which
+  // requires a VectorSplat on Arg0 or Arg1
+  S->Diag(BuiltinLoc, diag::err_vec_builtin_non_vector)
+      << TheCall->getDirectCallee()
+      << SourceRange(TheCall->getArg(0)->getBeginLoc(),
+                     TheCall->getArg(1)->getEndLoc());
+  return true;
+}
+
+// Note: returning true in this case results in CheckBuiltinFunctionCall
+// returning an ExprError
+bool Sema::CheckHLSLBuiltinFunctionCall(unsigned BuiltinID, CallExpr *TheCall) {
+  switch (BuiltinID) {
+  case Builtin::BI__builtin_hlsl_dot: {
+    if (checkArgCount(*this, TheCall, 2))
+      return true;
+    if (CheckVectorElementCallArgs(this, TheCall))
+      return true;
+    if (SemaBuiltinVectorToScalarMath(TheCall))
+      return true;
+    break;
+  }
+  }
+  return false;
+}
+
 bool Sema::CheckAMDGCNBuiltinFunctionCall(unsigned BuiltinID,
                                           CallExpr *TheCall) {
   // position of memory order and scope arguments in the builtin
@@ -19033,6 +19104,9 @@ static bool isLayoutCompatible(ASTContext &C, FieldDecl *Field1,
       return false;
   }
 
+  if (Field1->hasAttr<clang::NoUniqueAddressAttr>() ||
+      Field2->hasAttr<clang::NoUniqueAddressAttr>())
+    return false;
   return true;
 }
 
@@ -19588,23 +19662,43 @@ bool Sema::PrepareBuiltinElementwiseMathOneArgCall(CallExpr *TheCall) {
   TheCall->setArg(0, A.get());
   QualType TyA = A.get()->getType();
 
-  if (checkMathBuiltinElementType(*this, A.get()->getBeginLoc(), TyA))
+  if (checkMathBuiltinElementType(*this, A.get()->getBeginLoc(), TyA, 1))
     return true;
 
   TheCall->setType(TyA);
   return false;
 }
 
 bool Sema::SemaBuiltinElementwiseMath(CallExpr *TheCall) {
+  QualType Res;
+  if (SemaBuiltinVectorMath(TheCall, Res))
+    return true;
+  TheCall->setType(Res);
+  return false;
+}
+
+bool Sema::SemaBuiltinVectorToScalarMath(CallExpr *TheCall) {
+  QualType Res;
+  if (SemaBuiltinVectorMath(TheCall, Res))
+    return true;
+
+  if (auto *VecTy0 = Res->getAs<VectorType>())
+    TheCall->setType(VecTy0->getElementType());
+  else
+    TheCall->setType(Res);
+
+  return false;
+}
+
+bool Sema::SemaBuiltinVectorMath(CallExpr *TheCall, QualType &Res) {
   if (checkArgCount(*this, TheCall, 2))
     return true;
 
   ExprResult A = TheCall->getArg(0);
   ExprResult B = TheCall->getArg(1);
   // Do standard promotions between the two arguments, returning their common
   // type.
-  QualType Res =
-      UsualArithmeticConversions(A, B, TheCall->getExprLoc(), ACK_Comparison);
+  Res = UsualArithmeticConversions(A, B, TheCall->getExprLoc(), ACK_Comparison);
   if (A.isInvalid() || B.isInvalid())
     return true;
 
@@ -19616,12 +19710,11 @@ bool Sema::SemaBuiltinElementwiseMath(CallExpr *TheCall) {
                 diag::err_typecheck_call_different_arg_types)
            << TyA << TyB;
 
-  if (checkMathBuiltinElementType(*this, A.get()->getBeginLoc(), TyA))
+  if (checkMathBuiltinElementType(*this, A.get()->getBeginLoc(), TyA, 1))
     return true;
 
   TheCall->setArg(0, A.get());
   TheCall->setArg(1, B.get());
-  TheCall->setType(Res);
   return false;
 }
 

diff --git a/clang/lib/Sema/SemaDecl.cpp b/clang/lib/Sema/SemaDecl.cpp
@@ -16424,6 +16424,7 @@ void Sema::ActOnFinishDelayedAttribute(Scope *S, Decl *D,
   if (TemplateDecl *TD = dyn_cast<TemplateDecl>(D))
     D = TD->getTemplatedDecl();
   ProcessDeclAttributeList(S, D, Attrs);
+  ProcessAPINotes(D);
 
   if (CXXMethodDecl *Method = dyn_cast_or_null<CXXMethodDecl>(D))
     if (Method->isStatic())
@@ -19683,6 +19684,7 @@ void Sema::ActOnFields(Scope *S, SourceLocation RecLoc, Decl *EnclosingDecl,
       CDecl->setIvarRBraceLoc(RBrac);
     }
   }
+  ProcessAPINotes(Record);
 }
 
 /// Determine whether the given integral value is representable within
@@ -19997,6 +19999,7 @@ Decl *Sema::ActOnEnumConstant(Scope *S, Decl *theEnumDecl, Decl *lastEnumConst,
   // Process attributes.
   ProcessDeclAttributeList(S, New, Attrs);
   AddPragmaAttributes(S, New);
+  ProcessAPINotes(New);
 
   // Register this decl in the current scope stack.
   New->setAccess(TheEnumDecl->getAccess());
@@ -20195,6 +20198,7 @@ void Sema::ActOnEnumBody(SourceLocation EnumLoc, SourceRange BraceRange,
   QualType EnumType = Context.getTypeDeclType(Enum);
 
   ProcessDeclAttributeList(S, Enum, Attrs);
+  ProcessAPINotes(Enum);
 
   if (Enum->isDependentType()) {
     for (unsigned i = 0, e = Elements.size(); i != e; ++i) {

diff --git a/clang/lib/Sema/SemaDeclAttr.cpp b/clang/lib/Sema/SemaDeclAttr.cpp
@@ -10146,6 +10146,9 @@ void Sema::ProcessDeclAttributes(Scope *S, Decl *D, const Declarator &PD) {
 
   // Apply additional attributes specified by '#pragma clang attribute'.
   AddPragmaAttributes(S, D);
+
+  // Look for API notes that map to attributes.
+  ProcessAPINotes(D);
 }
 
 /// Is the given declaration allowed to use a forbidden type?

diff --git a/clang/lib/Sema/SemaDeclCXX.cpp b/clang/lib/Sema/SemaDeclCXX.cpp
@@ -11724,6 +11724,7 @@ Decl *Sema::ActOnStartNamespaceDef(Scope *NamespcScope,
 
   ProcessDeclAttributeList(DeclRegionScope, Namespc, AttrList);
   AddPragmaAttributes(DeclRegionScope, Namespc);
+  ProcessAPINotes(Namespc);
 
   // FIXME: Should we be merging attributes?
   if (const VisibilityAttr *Attr = Namespc->getAttr<VisibilityAttr>())
@@ -12270,8 +12271,10 @@ Decl *Sema::ActOnUsingDirective(Scope *S, SourceLocation UsingLoc,
     Diag(IdentLoc, diag::err_expected_namespace_name) << SS.getRange();
   }
 
-  if (UDir)
+  if (UDir) {
     ProcessDeclAttributeList(S, UDir, AttrList);
+    ProcessAPINotes(UDir);
+  }
 
   return UDir;
 }
@@ -13563,6 +13566,7 @@ Decl *Sema::ActOnAliasDeclaration(Scope *S, AccessSpecifier AS,
 
   ProcessDeclAttributeList(S, NewTD, AttrList);
   AddPragmaAttributes(S, NewTD);
+  ProcessAPINotes(NewTD);
 
   CheckTypedefForVariablyModifiedType(S, NewTD);
   Invalid |= NewTD->isInvalidDecl();

diff --git a/clang/lib/Sema/SemaDeclObjC.cpp b/clang/lib/Sema/SemaDeclObjC.cpp
@@ -1072,6 +1072,7 @@ ObjCInterfaceDecl *Sema::ActOnStartClassInterface(
 
   ProcessDeclAttributeList(TUScope, IDecl, AttrList);
   AddPragmaAttributes(TUScope, IDecl);
+  ProcessAPINotes(IDecl);
 
   // Merge attributes from previous declarations.
   if (PrevIDecl)
@@ -1273,6 +1274,7 @@ ObjCProtocolDecl *Sema::ActOnStartProtocolInterface(
 
   ProcessDeclAttributeList(TUScope, PDecl, AttrList);
   AddPragmaAttributes(TUScope, PDecl);
+  ProcessAPINotes(PDecl);
 
   // Merge attributes from previous declarations.
   if (PrevDecl)
@@ -4807,6 +4809,7 @@ Decl *Sema::ActOnMethodDeclaration(
     // Apply the attributes to the parameter.
     ProcessDeclAttributeList(TUScope, Param, ArgInfo[i].ArgAttrs);
     AddPragmaAttributes(TUScope, Param);
+    ProcessAPINotes(Param);
 
     if (Param->hasAttr<BlocksAttr>()) {
       Diag(Param->getLocation(), diag::err_block_on_nonlocal);
@@ -4837,6 +4840,7 @@ Decl *Sema::ActOnMethodDeclaration(
 
   ProcessDeclAttributeList(TUScope, ObjCMethod, AttrList);
   AddPragmaAttributes(TUScope, ObjCMethod);
+  ProcessAPINotes(ObjCMethod);
 
   // Add the method now.
   const ObjCMethodDecl *PrevMethod = nullptr;

diff --git a/clang/lib/Sema/SemaObjCProperty.cpp b/clang/lib/Sema/SemaObjCProperty.cpp
@@ -2509,6 +2509,8 @@ void Sema::ProcessPropertyDecl(ObjCPropertyDecl *property) {
       GetterMethod->addAttr(SectionAttr::CreateImplicit(
           Context, SA->getName(), Loc, SectionAttr::GNU_section));
 
+    ProcessAPINotes(GetterMethod);
+
     if (getLangOpts().ObjCAutoRefCount)
       CheckARCMethodDecl(GetterMethod);
   } else
@@ -2578,6 +2580,9 @@ void Sema::ProcessPropertyDecl(ObjCPropertyDecl *property) {
       if (const SectionAttr *SA = property->getAttr<SectionAttr>())
         SetterMethod->addAttr(SectionAttr::CreateImplicit(
             Context, SA->getName(), Loc, SectionAttr::GNU_section));
+
+      ProcessAPINotes(SetterMethod);
+
       // It's possible for the user to have set a very odd custom
       // setter selector that causes it to have a method family.
       if (getLangOpts().ObjCAutoRefCount)

diff --git a/clang/lib/Sema/SemaStmt.cpp b/clang/lib/Sema/SemaStmt.cpp
@@ -3356,6 +3356,15 @@ Sema::ActOnContinueStmt(SourceLocation ContinueLoc, Scope *CurScope) {
     // initialization of that variable.
     return StmtError(Diag(ContinueLoc, diag::err_continue_from_cond_var_init));
   }
+
+  // A 'continue' that would normally have execution continue on a block outside
+  // of a compute construct counts as 'branching out of' the compute construct,
+  // so diagnose here.
+  if (S->isOpenACCComputeConstructScope())
+    return StmtError(
+        Diag(ContinueLoc, diag::err_acc_branch_in_out_compute_construct)
+        << /*branch*/ 0 << /*out of */ 0);
+
   CheckJumpOutOfSEHFinally(*this, ContinueLoc, *S);
 
   return new (Context) ContinueStmt(ContinueLoc);
@@ -3371,6 +3380,21 @@ Sema::ActOnBreakStmt(SourceLocation BreakLoc, Scope *CurScope) {
   if (S->isOpenMPLoopScope())
     return StmtError(Diag(BreakLoc, diag::err_omp_loop_cannot_use_stmt)
                      << "break");
+
+  // OpenACC doesn't allow 'break'ing from a compute construct, so diagnose if
+  // we are trying to do so.  This can come in 2 flavors: 1-the break'able thing
+  // (besides the compute construct) 'contains' the compute construct, at which
+  // point the 'break' scope will be the compute construct.  Else it could be a
+  // loop of some sort that has a direct parent of the compute construct.
+  // However, a 'break' in a 'switch' marked as a compute construct doesn't
+  // count as 'branch out of' the compute construct.
+  if (S->isOpenACCComputeConstructScope() ||
+      (S->isLoopScope() && S->getParent() &&
+       S->getParent()->isOpenACCComputeConstructScope()))
+    return StmtError(
+        Diag(BreakLoc, diag::err_acc_branch_in_out_compute_construct)
+        << /*branch*/ 0 << /*out of */ 0);
+
   CheckJumpOutOfSEHFinally(*this, BreakLoc, *S);
 
   return new (Context) BreakStmt(BreakLoc);
@@ -3925,6 +3949,12 @@ Sema::ActOnReturnStmt(SourceLocation ReturnLoc, Expr *RetValExp,
       RetValExp, nullptr, /*RecoverUncorrectedTypos=*/true);
   if (RetVal.isInvalid())
     return StmtError();
+
+  if (getCurScope()->isInOpenACCComputeConstructScope())
+    return StmtError(
+        Diag(ReturnLoc, diag::err_acc_branch_in_out_compute_construct)
+        << /*return*/ 1 << /*out of */ 0);
+
   StmtResult R =
       BuildReturnStmt(ReturnLoc, RetVal.get(), /*AllowRecovery=*/true);
   if (R.isInvalid() || ExprEvalContexts.back().isDiscardedStatementContext())

diff --git a/clang/lib/Sema/SemaTemplate.cpp b/clang/lib/Sema/SemaTemplate.cpp
@@ -2158,6 +2158,7 @@ DeclResult Sema::CheckClassTemplate(
     NewClass->startDefinition();
 
   ProcessDeclAttributeList(S, NewClass, Attr);
+  ProcessAPINotes(NewClass);
 
   if (PrevClassTemplate)
     mergeDeclAttributes(NewClass, PrevClassTemplate->getTemplatedDecl());
@@ -4958,11 +4959,10 @@ Sema::CheckVarTemplateId(VarTemplateDecl *Template, SourceLocation TemplateLoc,
   return Decl;
 }
 
-ExprResult
-Sema::CheckVarTemplateId(const CXXScopeSpec &SS,
-                         const DeclarationNameInfo &NameInfo,
-                         VarTemplateDecl *Template, SourceLocation TemplateLoc,
-                         const TemplateArgumentListInfo *TemplateArgs) {
+ExprResult Sema::CheckVarTemplateId(
+    const CXXScopeSpec &SS, const DeclarationNameInfo &NameInfo,
+    VarTemplateDecl *Template, NamedDecl *FoundD, SourceLocation TemplateLoc,
+    const TemplateArgumentListInfo *TemplateArgs) {
 
   DeclResult Decl = CheckVarTemplateId(Template, TemplateLoc, NameInfo.getLoc(),
                                        *TemplateArgs);
@@ -4978,8 +4978,7 @@ Sema::CheckVarTemplateId(const CXXScopeSpec &SS,
                                        NameInfo.getLoc());
 
   // Build an ordinary singleton decl ref.
-  return BuildDeclarationNameExpr(SS, NameInfo, Var,
-                                  /*FoundD=*/nullptr, TemplateArgs);
+  return BuildDeclarationNameExpr(SS, NameInfo, Var, FoundD, TemplateArgs);
 }
 
 void Sema::diagnoseMissingTemplateArguments(TemplateName Name,
@@ -5066,9 +5065,9 @@ ExprResult Sema::BuildTemplateIdExpr(const CXXScopeSpec &SS,
   bool KnownDependent = false;
   // In C++1y, check variable template ids.
   if (R.getAsSingle<VarTemplateDecl>()) {
-    ExprResult Res = CheckVarTemplateId(SS, R.getLookupNameInfo(),
-                                        R.getAsSingle<VarTemplateDecl>(),
-                                        TemplateKWLoc, TemplateArgs);
+    ExprResult Res = CheckVarTemplateId(
+        SS, R.getLookupNameInfo(), R.getAsSingle<VarTemplateDecl>(),
+        R.getRepresentativeDecl(), TemplateKWLoc, TemplateArgs);
     if (Res.isInvalid() || Res.isUsable())
       return Res;
     // Result is dependent. Carry on to build an UnresolvedLookupEpxr.
@@ -9114,6 +9113,7 @@ DeclResult Sema::ActOnClassTemplateSpecialization(
   }
 
   ProcessDeclAttributeList(S, Specialization, Attr);
+  ProcessAPINotes(Specialization);
 
   // Add alignment attributes if necessary; these attributes are checked when
   // the ASTContext lays out the structure.
@@ -9709,6 +9709,40 @@ bool Sema::CheckFunctionTemplateSpecialization(
   // Ignore access information;  it doesn't figure into redeclaration checking.
   FunctionDecl *Specialization = cast<FunctionDecl>(*Result);
 
+  // C++23 [except.spec]p13:
+  //   An exception specification is considered to be needed when:
+  //   - [...]
+  //   - the exception specification is compared to that of another declaration
+  //     (e.g., an explicit specialization or an overriding virtual function);
+  //   - [...]
+  //
+  //  The exception specification of a defaulted function is evaluated as
+  //  described above only when needed; similarly, the noexcept-specifier of a
+  //  specialization of a function template or member function of a class
+  //  template is instantiated only when needed.
+  //
+  // The standard doesn't specify what the "comparison with another declaration"
+  // entails, nor the exact circumstances in which it occurs. Moreover, it does
+  // not state which properties of an explicit specialization must match the
+  // primary template.
+  //
+  // We assume that an explicit specialization must correspond with (per
+  // [basic.scope.scope]p4) and declare the same entity as (per [basic.link]p8)
+  // the declaration produced by substitution into the function template.
+  //
+  // Since the determination whether two function declarations correspond does
+  // not consider exception specification, we only need to instantiate it once
+  // we determine the primary template when comparing types per
+  // [basic.link]p11.1.
+  auto *SpecializationFPT =
+      Specialization->getType()->castAs<FunctionProtoType>();
+  // If the function has a dependent exception specification, resolve it after
+  // we have selected the primary template so we can check whether it matches.
+  if (getLangOpts().CPlusPlus17 &&
+      isUnresolvedExceptionSpec(SpecializationFPT->getExceptionSpecType()) &&
+      !ResolveExceptionSpec(FD->getLocation(), SpecializationFPT))
+    return true;
+
   FunctionTemplateSpecializationInfo *SpecInfo
     = Specialization->getTemplateSpecializationInfo();
   assert(SpecInfo && "Function template specialization info missing?");
@@ -10348,6 +10382,7 @@ DeclResult Sema::ActOnExplicitInstantiation(
 
   bool PreviouslyDLLExported = Specialization->hasAttr<DLLExportAttr>();
   ProcessDeclAttributeList(S, Specialization, Attr);
+  ProcessAPINotes(Specialization);
 
   // Add the explicit instantiation into its lexical context. However,
   // since explicit instantiations are never found by name lookup, we
@@ -10758,6 +10793,9 @@ DeclResult Sema::ActOnExplicitInstantiation(Scope *S,
       Prev->setTemplateSpecializationKind(TSK, D.getIdentifierLoc());
       // Merge attributes.
       ProcessDeclAttributeList(S, Prev, D.getDeclSpec().getAttributes());
+      if (PrevTemplate)
+        ProcessAPINotes(Prev);
+
       if (TSK == TSK_ExplicitInstantiationDefinition)
         InstantiateVariableDefinition(D.getIdentifierLoc(), Prev);
     }
@@ -10933,6 +10971,7 @@ DeclResult Sema::ActOnExplicitInstantiation(Scope *S,
   }
 
   ProcessDeclAttributeList(S, Specialization, D.getDeclSpec().getAttributes());
+  ProcessAPINotes(Specialization);
 
   // In MSVC mode, dllimported explicit instantiation definitions are treated as
   // instantiation declarations.

diff --git a/clang/lib/Sema/SemaTemplateDeduction.cpp b/clang/lib/Sema/SemaTemplateDeduction.cpp
@@ -4632,11 +4632,9 @@ TemplateDeductionResult Sema::DeduceTemplateArguments(
                                                Info.getLocation()))
     return TemplateDeductionResult::MiscellaneousDeductionFailure;
 
-  // If the function has a dependent exception specification, resolve it now,
-  // so we can check that the exception specification matches.
   auto *SpecializationFPT =
       Specialization->getType()->castAs<FunctionProtoType>();
-  if (getLangOpts().CPlusPlus17 &&
+  if (IsAddressOfFunction && getLangOpts().CPlusPlus17 &&
       isUnresolvedExceptionSpec(SpecializationFPT->getExceptionSpecType()) &&
       !ResolveExceptionSpec(Info.getLocation(), SpecializationFPT))
     return TemplateDeductionResult::MiscellaneousDeductionFailure;
@@ -4662,11 +4660,11 @@ TemplateDeductionResult Sema::DeduceTemplateArguments(
   // specialization with respect to arguments of compatible pointer to function
   // types, template argument deduction fails.
   if (!ArgFunctionType.isNull()) {
-    if (IsAddressOfFunction
-            ? !isSameOrCompatibleFunctionType(
-                  Context.getCanonicalType(SpecializationType),
-                  Context.getCanonicalType(ArgFunctionType))
-            : !Context.hasSameType(SpecializationType, ArgFunctionType)) {
+    if (IsAddressOfFunction ? !isSameOrCompatibleFunctionType(
+                                  Context.getCanonicalType(SpecializationType),
+                                  Context.getCanonicalType(ArgFunctionType))
+                            : !Context.hasSameFunctionTypeIgnoringExceptionSpec(
+                                  SpecializationType, ArgFunctionType)) {
       Info.FirstArg = TemplateArgument(SpecializationType);
       Info.SecondArg = TemplateArgument(ArgFunctionType);
       return TemplateDeductionResult::NonDeducedMismatch;

diff --git a/clang/lib/StaticAnalyzer/Checkers/ArrayBoundCheckerV2.cpp b/clang/lib/StaticAnalyzer/Checkers/ArrayBoundCheckerV2.cpp
@@ -268,6 +268,16 @@ getSimplifiedOffsets(NonLoc offset, nonloc::ConcreteInt extent,
   return std::pair<NonLoc, nonloc::ConcreteInt>(offset, extent);
 }
 
+static bool isNegative(SValBuilder &SVB, ProgramStateRef State, NonLoc Value) {
+  const llvm::APSInt *MaxV = SVB.getMaxValue(State, Value);
+  return MaxV && MaxV->isNegative();
+}
+
+static bool isUnsigned(SValBuilder &SVB, NonLoc Value) {
+  QualType T = Value.getType(SVB.getContext());
+  return T->isUnsignedIntegerType();
+}
+
 // Evaluate the comparison Value < Threshold with the help of the custom
 // simplification algorithm defined for this checker. Return a pair of states,
 // where the first one corresponds to "value below threshold" and the second
@@ -281,18 +291,32 @@ compareValueToThreshold(ProgramStateRef State, NonLoc Value, NonLoc Threshold,
   if (auto ConcreteThreshold = Threshold.getAs<nonloc::ConcreteInt>()) {
     std::tie(Value, Threshold) = getSimplifiedOffsets(Value, *ConcreteThreshold, SVB);
   }
-  if (auto ConcreteThreshold = Threshold.getAs<nonloc::ConcreteInt>()) {
-    QualType T = Value.getType(SVB.getContext());
-    if (T->isUnsignedIntegerType() && ConcreteThreshold->getValue().isNegative()) {
-      // In this case we reduced the bound check to a comparison of the form
-      //   (symbol or value with unsigned type) < (negative number)
-      // which is always false. We are handling these cases separately because
-      // evalBinOpNN can perform a signed->unsigned conversion that turns the
-      // negative number into a huge positive value and leads to wildly
-      // inaccurate conclusions.
+
+  // We want to perform a _mathematical_ comparison between the numbers `Value`
+  // and `Threshold`; but `evalBinOpNN` evaluates a C/C++ operator that may
+  // perform automatic conversions. For example the number -1 is less than the
+  // number 1000, but -1 < `1000ull` will evaluate to `false` because the `int`
+  // -1 is converted to ULONGLONG_MAX.
+  // To avoid automatic conversions, we evaluate the "obvious" cases without
+  // calling `evalBinOpNN`:
+  if (isNegative(SVB, State, Value) && isUnsigned(SVB, Threshold)) {
+    if (CheckEquality) {
+      // negative_value == unsigned_value is always false
       return {nullptr, State};
     }
+    // negative_value < unsigned_value is always false
+    return {State, nullptr};
   }
+  if (isUnsigned(SVB, Value) && isNegative(SVB, State, Threshold)) {
+    // unsigned_value == negative_value and unsigned_value < negative_value are
+    // both always false
+    return {nullptr, State};
+  }
+  // FIXME: these special cases are sufficient for handling real-world
+  // comparisons, but in theory there could be contrived situations where
+  // automatic conversion of a symbolic value (which can be negative and can be
+  // positive) leads to incorrect results.
+
   const BinaryOperatorKind OpKind = CheckEquality ? BO_EQ : BO_LT;
   auto BelowThreshold =
       SVB.evalBinOpNN(State, OpKind, Value, Threshold, SVB.getConditionType())

diff --git a/clang/lib/StaticAnalyzer/Checkers/ErrnoChecker.cpp b/clang/lib/StaticAnalyzer/Checkers/ErrnoChecker.cpp
@@ -17,7 +17,7 @@
 #include "clang/StaticAnalyzer/Checkers/BuiltinCheckerRegistration.h"
 #include "clang/StaticAnalyzer/Core/Checker.h"
 #include "clang/StaticAnalyzer/Core/CheckerManager.h"
-#include "clang/StaticAnalyzer/Core/PathSensitive/CallDescription.h"
+#include "clang/StaticAnalyzer/Core/PathSensitive/CallEvent.h"
 #include "clang/StaticAnalyzer/Core/PathSensitive/CheckerContext.h"
 #include "clang/StaticAnalyzer/Core/PathSensitive/ProgramState.h"
 #include "clang/StaticAnalyzer/Core/PathSensitive/SVals.h"