diff --git a/clang/lib/CodeGen/CGAtomic.cpp b/clang/lib/CodeGen/CGAtomic.cpp
@@ -727,7 +727,7 @@ static void EmitAtomicOp(CodeGenFunction &CGF, AtomicExpr *E, Address Dest,
 
   llvm::Value *LoadVal1 = CGF.Builder.CreateLoad(Val1);
   llvm::AtomicRMWInst *RMWI =
-      CGF.Builder.CreateAtomicRMW(Op, Ptr, LoadVal1, Order, Scope);
+      CGF.emitAtomicRMWInst(Op, Ptr, LoadVal1, Order, Scope);
   RMWI->setVolatile(E->isVolatile());
 
   // For __atomic_*_fetch operations, perform the operation again to
@@ -2034,6 +2034,17 @@ std::pair<RValue, llvm::Value *> CodeGenFunction::EmitAtomicCompareExchange(
                                            IsWeak);
 }
 
+llvm::AtomicRMWInst *
+CodeGenFunction::emitAtomicRMWInst(llvm::AtomicRMWInst::BinOp Op, Address Addr,
+                                   llvm::Value *Val, llvm::AtomicOrdering Order,
+                                   llvm::SyncScope::ID SSID) {
+
+  llvm::AtomicRMWInst *RMW =
+      Builder.CreateAtomicRMW(Op, Addr, Val, Order, SSID);
+  getTargetHooks().setTargetAtomicMetadata(*this, *RMW);
+  return RMW;
+}
+
 void CodeGenFunction::EmitAtomicUpdate(
     LValue LVal, llvm::AtomicOrdering AO,
     const llvm::function_ref<RValue(RValue)> &UpdateOp, bool IsVolatile) {

diff --git a/clang/lib/CodeGen/CGExprScalar.cpp b/clang/lib/CodeGen/CGExprScalar.cpp
@@ -2841,9 +2841,10 @@ ScalarExprEmitter::EmitScalarPrePostIncDec(const UnaryOperator *E, LValue LV,
           isInc ? llvm::Instruction::FAdd : llvm::Instruction::FSub;
       llvm::Value *amt = llvm::ConstantFP::get(
           VMContext, llvm::APFloat(static_cast<float>(1.0)));
-      llvm::Value *old =
-          Builder.CreateAtomicRMW(aop, LV.getAddress(), amt,
-                                  llvm::AtomicOrdering::SequentiallyConsistent);
+      llvm::AtomicRMWInst *old =
+          CGF.emitAtomicRMWInst(aop, LV.getAddress(), amt,
+                                llvm::AtomicOrdering::SequentiallyConsistent);
+
       return isPre ? Builder.CreateBinOp(op, old, amt) : old;
     }
     value = EmitLoadOfLValue(LV, E->getExprLoc());
@@ -3583,9 +3584,9 @@ LValue ScalarExprEmitter::EmitCompoundAssignLValue(
             EmitScalarConversion(OpInfo.RHS, E->getRHS()->getType(), LHSTy,
                                  E->getExprLoc()),
             LHSTy);
-        Value *OldVal = Builder.CreateAtomicRMW(
-            AtomicOp, LHSLV.getAddress(), Amt,
-            llvm::AtomicOrdering::SequentiallyConsistent);
+
+        llvm::AtomicRMWInst *OldVal =
+            CGF.emitAtomicRMWInst(AtomicOp, LHSLV.getAddress(), Amt);
 
         // Since operation is atomic, the result type is guaranteed to be the
         // same as the input in LLVM terms.

diff --git a/clang/lib/CodeGen/CGStmtOpenMP.cpp b/clang/lib/CodeGen/CGStmtOpenMP.cpp
@@ -6326,8 +6326,8 @@ static std::pair<bool, RValue> emitOMPAtomicRMW(CodeGenFunction &CGF, LValue X,
       UpdateVal = CGF.Builder.CreateCast(llvm::Instruction::CastOps::UIToFP, IC,
                                          X.getAddress().getElementType());
   }
-  llvm::Value *Res =
-      CGF.Builder.CreateAtomicRMW(RMWOp, X.getAddress(), UpdateVal, AO);
+  llvm::AtomicRMWInst *Res =
+      CGF.emitAtomicRMWInst(RMWOp, X.getAddress(), UpdateVal, AO);
   return std::make_pair(true, RValue::get(Res));
 }
 

diff --git a/clang/lib/CodeGen/CodeGenFunction.h b/clang/lib/CodeGen/CodeGenFunction.h
@@ -4160,6 +4160,13 @@ class CodeGenFunction : public CodeGenTypeCache {
           llvm::AtomicOrdering::SequentiallyConsistent,
       bool IsWeak = false, AggValueSlot Slot = AggValueSlot::ignored());
 
+  /// Emit an atomicrmw instruction, and applying relevant metadata when
+  /// applicable.
+  llvm::AtomicRMWInst *emitAtomicRMWInst(
+      llvm::AtomicRMWInst::BinOp Op, Address Addr, llvm::Value *Val,
+      llvm::AtomicOrdering Order = llvm::AtomicOrdering::SequentiallyConsistent,
+      llvm::SyncScope::ID SSID = llvm::SyncScope::System);
+
   void EmitAtomicUpdate(LValue LVal, llvm::AtomicOrdering AO,
                         const llvm::function_ref<RValue(RValue)> &UpdateOp,
                         bool IsVolatile);

diff --git a/clang/lib/CodeGen/TargetInfo.h b/clang/lib/CodeGen/TargetInfo.h
@@ -334,6 +334,10 @@ class TargetCodeGenInfo {
                                                  llvm::AtomicOrdering Ordering,
                                                  llvm::LLVMContext &Ctx) const;
 
+  /// Allow the target to apply other metadata to an atomic instruction
+  virtual void setTargetAtomicMetadata(CodeGenFunction &CGF,
+                                       llvm::AtomicRMWInst &RMW) const {}
+
   /// Interface class for filling custom fields of a block literal for OpenCL.
   class TargetOpenCLBlockHelper {
   public:

diff --git a/clang/lib/CodeGen/Targets/AMDGPU.cpp b/clang/lib/CodeGen/Targets/AMDGPU.cpp
@@ -311,6 +311,8 @@ class AMDGPUTargetCodeGenInfo : public TargetCodeGenInfo {
                                          SyncScope Scope,
                                          llvm::AtomicOrdering Ordering,
                                          llvm::LLVMContext &Ctx) const override;
+  void setTargetAtomicMetadata(CodeGenFunction &CGF,
+                               llvm::AtomicRMWInst &RMW) const override;
   llvm::Value *createEnqueuedBlockKernel(CodeGenFunction &CGF,
                                          llvm::Function *BlockInvokeFunc,
                                          llvm::Type *BlockTy) const override;
@@ -546,6 +548,23 @@ AMDGPUTargetCodeGenInfo::getLLVMSyncScopeID(const LangOptions &LangOpts,
   return Ctx.getOrInsertSyncScopeID(Name);
 }
 
+void AMDGPUTargetCodeGenInfo::setTargetAtomicMetadata(
+    CodeGenFunction &CGF, llvm::AtomicRMWInst &RMW) const {
+  if (!CGF.getTarget().allowAMDGPUUnsafeFPAtomics())
+    return;
+
+  // TODO: Introduce new, more controlled options that also work for integers,
+  // and deprecate allowAMDGPUUnsafeFPAtomics.
+  llvm::AtomicRMWInst::BinOp RMWOp = RMW.getOperation();
+  if (llvm::AtomicRMWInst::isFPOperation(RMWOp)) {
+    llvm::MDNode *Empty = llvm::MDNode::get(CGF.getLLVMContext(), {});
+    RMW.setMetadata("amdgpu.no.fine.grained.memory", Empty);
+
+    if (RMWOp == llvm::AtomicRMWInst::FAdd && RMW.getType()->isFloatTy())
+      RMW.setMetadata("amdgpu.ignore.denormal.mode", Empty);
+  }
+}
+
 bool AMDGPUTargetCodeGenInfo::shouldEmitStaticExternCAliases() const {
   return false;
 }

diff --git a/clang/lib/Driver/ToolChains/AMDGPU.cpp b/clang/lib/Driver/ToolChains/AMDGPU.cpp
@@ -620,8 +620,10 @@ void amdgpu::Linker::ConstructJob(Compilation &C, const JobAction &JA,
                                   const char *LinkingOutput) const {
   std::string Linker = getToolChain().GetLinkerPath();
   ArgStringList CmdArgs;
-  CmdArgs.push_back("--no-undefined");
-  CmdArgs.push_back("-shared");
+  if (!Args.hasArg(options::OPT_r)) {
+    CmdArgs.push_back("--no-undefined");
+    CmdArgs.push_back("-shared");
+  }
 
   addLinkerCompressDebugSectionsOption(getToolChain(), Args, CmdArgs);
   Args.AddAllArgs(CmdArgs, options::OPT_L);

diff --git a/clang/lib/Driver/ToolChains/Clang.cpp b/clang/lib/Driver/ToolChains/Clang.cpp
@@ -3384,9 +3384,28 @@ static void RenderFloatingPointOptions(const ToolChain &TC, const Driver &D,
     CmdArgs.push_back("-ffast-math");
 
   // Handle __FINITE_MATH_ONLY__ similarly.
-  if (!HonorINFs && !HonorNaNs)
+  // The -ffinite-math-only is added to CmdArgs when !HonorINFs && !HonorNaNs.
+  // Otherwise process the Xclang arguments to determine if -menable-no-infs and
+  // -menable-no-nans are set by the user.
+  bool shouldAddFiniteMathOnly = false;
+  if (!HonorINFs && !HonorNaNs) {
+    shouldAddFiniteMathOnly = true;
+  } else {
+    bool InfValues = true;
+    bool NanValues = true;
+    for (const auto *Arg : Args.filtered(options::OPT_Xclang)) {
+      StringRef ArgValue = Arg->getValue();
+      if (ArgValue == "-menable-no-nans")
+        NanValues = false;
+      else if (ArgValue == "-menable-no-infs")
+        InfValues = false;
+    }
+    if (!NanValues && !InfValues)
+      shouldAddFiniteMathOnly = true;
+  }
+  if (shouldAddFiniteMathOnly) {
     CmdArgs.push_back("-ffinite-math-only");
-
+  }
   if (const Arg *A = Args.getLastArg(options::OPT_mfpmath_EQ)) {
     CmdArgs.push_back("-mfpmath");
     CmdArgs.push_back(A->getValue());
@@ -3755,6 +3774,11 @@ static void RenderOpenCLOptions(const ArgList &Args, ArgStringList &CmdArgs,
     CmdArgs.push_back(Args.MakeArgString(CLExtStr));
   }
 
+  if (Args.hasArg(options::OPT_cl_finite_math_only)) {
+    CmdArgs.push_back("-menable-no-infs");
+    CmdArgs.push_back("-menable-no-nans");
+  }
+
   for (const auto &Arg : ForwardedArguments)
     if (const auto *A = Args.getLastArg(Arg))
       CmdArgs.push_back(Args.MakeArgString(A->getOption().getPrefixedName()));

diff --git a/clang/lib/Driver/ToolChains/PS4CPU.cpp b/clang/lib/Driver/ToolChains/PS4CPU.cpp
@@ -266,6 +266,10 @@ void tools::PS5cpu::Linker::ConstructJob(Compilation &C, const JobAction &JA,
   if (UseJMC)
     AddLTOFlag("-enable-jmc-instrument");
 
+  if (Args.hasFlag(options::OPT_fstack_size_section,
+                   options::OPT_fno_stack_size_section, false))
+    AddLTOFlag("-stack-size-section");
+
   if (Arg *A = Args.getLastArg(options::OPT_fcrash_diagnostics_dir))
     AddLTOFlag(Twine("-crash-diagnostics-dir=") + A->getValue());
 

diff --git a/clang/lib/Frontend/InitPreprocessor.cpp b/clang/lib/Frontend/InitPreprocessor.cpp
@@ -1318,7 +1318,7 @@ static void InitializePredefinedMacros(const TargetInfo &TI,
   if (!LangOpts.MathErrno)
     Builder.defineMacro("__NO_MATH_ERRNO__");
 
-  if (LangOpts.FastMath || LangOpts.FiniteMathOnly)
+  if (LangOpts.FastMath || (LangOpts.NoHonorInfs && LangOpts.NoHonorNaNs))
     Builder.defineMacro("__FINITE_MATH_ONLY__", "1");
   else
     Builder.defineMacro("__FINITE_MATH_ONLY__", "0");

diff --git a/clang/lib/Frontend/PrintPreprocessedOutput.cpp b/clang/lib/Frontend/PrintPreprocessedOutput.cpp
@@ -980,7 +980,6 @@ static void PrintPreprocessedTokens(Preprocessor &PP, Token &Tok,
         *Callbacks->OS << static_cast<unsigned>(*Iter);
         PrintComma = true;
       }
-      IsStartOfLine = true;
     } else if (Tok.isAnnotation()) {
       // Ignore annotation tokens created by pragmas - the pragmas themselves
       // will be reproduced in the preprocessed output.

diff --git a/clang/lib/Lex/PPMacroExpansion.cpp b/clang/lib/Lex/PPMacroExpansion.cpp
@@ -1604,6 +1604,16 @@ static bool isTargetVariantEnvironment(const TargetInfo &TI,
   return false;
 }
 
+#if defined(__sun__) && defined(__svr4__)
+// GCC mangles std::tm as tm for binary compatibility on Solaris (Issue
+// #33114).  We need to match this to allow the std::put_time calls to link
+// (PR #99075).
+asm("_ZNKSt8time_putIcSt19ostreambuf_iteratorIcSt11char_traitsIcEEE3putES3_"
+    "RSt8ios_basecPKSt2tmPKcSB_ = "
+    "_ZNKSt8time_putIcSt19ostreambuf_iteratorIcSt11char_traitsIcEEE3putES3_"
+    "RSt8ios_basecPK2tmPKcSB_");
+#endif
+
 /// ExpandBuiltinMacro - If an identifier token is read that is to be expanded
 /// as a builtin macro, handle it and return the next token as 'Tok'.
 void Preprocessor::ExpandBuiltinMacro(Token &Tok) {

diff --git a/clang/lib/Sema/SemaChecking.cpp b/clang/lib/Sema/SemaChecking.cpp
@@ -1477,6 +1477,18 @@ static bool BuiltinSEHScopeCheck(Sema &SemaRef, CallExpr *TheCall,
   return false;
 }
 
+// In OpenCL, __builtin_alloca_* should return a pointer to address space
+// that corresponds to the stack address space i.e private address space.
+static void builtinAllocaAddrSpace(Sema &S, CallExpr *TheCall) {
+  QualType RT = TheCall->getType();
+  assert((RT->isPointerType() && !(RT->getPointeeType().hasAddressSpace())) &&
+         "__builtin_alloca has invalid address space");
+
+  RT = RT->getPointeeType();
+  RT = S.Context.getAddrSpaceQualType(RT, LangAS::opencl_private);
+  TheCall->setType(S.Context.getPointerType(RT));
+}
+
 namespace {
 enum PointerAuthOpKind {
   PAO_Strip,
@@ -2214,6 +2226,9 @@ Sema::CheckBuiltinFunctionCall(FunctionDecl *FDecl, unsigned BuiltinID,
   case Builtin::BI__builtin_alloca_uninitialized:
     Diag(TheCall->getBeginLoc(), diag::warn_alloca)
         << TheCall->getDirectCallee();
+    if (getLangOpts().OpenCL) {
+      builtinAllocaAddrSpace(*this, TheCall);
+    }
     break;
   case Builtin::BI__arithmetic_fence:
     if (BuiltinArithmeticFence(TheCall))

diff --git a/clang/lib/Sema/SemaDecl.cpp b/clang/lib/Sema/SemaDecl.cpp
@@ -11009,6 +11009,9 @@ static bool AttrCompatibleWithMultiVersion(attr::Kind Kind,
   switch (Kind) {
   default:
     return false;
+  case attr::ArmLocallyStreaming:
+    return MVKind == MultiVersionKind::TargetVersion ||
+           MVKind == MultiVersionKind::TargetClones;
   case attr::Used:
     return MVKind == MultiVersionKind::Target;
   case attr::NonNull:
@@ -11145,7 +11148,21 @@ bool Sema::areMultiversionVariantFunctionsCompatible(
     FunctionType::ExtInfo OldTypeInfo = OldType->getExtInfo();
     FunctionType::ExtInfo NewTypeInfo = NewType->getExtInfo();
 
-    if (OldTypeInfo.getCC() != NewTypeInfo.getCC())
+    const auto *OldFPT = OldFD->getType()->getAs<FunctionProtoType>();
+    const auto *NewFPT = NewFD->getType()->getAs<FunctionProtoType>();
+
+    bool ArmStreamingCCMismatched = false;
+    if (OldFPT && NewFPT) {
+      unsigned Diff =
+          OldFPT->getAArch64SMEAttributes() ^ NewFPT->getAArch64SMEAttributes();
+      // Arm-streaming, arm-streaming-compatible and non-streaming versions
+      // cannot be mixed.
+      if (Diff & (FunctionType::SME_PStateSMEnabledMask |
+                  FunctionType::SME_PStateSMCompatibleMask))
+        ArmStreamingCCMismatched = true;
+    }
+
+    if (OldTypeInfo.getCC() != NewTypeInfo.getCC() || ArmStreamingCCMismatched)
       return Diag(DiffDiagIDAt.first, DiffDiagIDAt.second) << CallingConv;
 
     QualType OldReturnType = OldType->getReturnType();
@@ -11165,9 +11182,8 @@ bool Sema::areMultiversionVariantFunctionsCompatible(
     if (!CLinkageMayDiffer && OldFD->isExternC() != NewFD->isExternC())
       return Diag(DiffDiagIDAt.first, DiffDiagIDAt.second) << LanguageLinkage;
 
-    if (CheckEquivalentExceptionSpec(
-            OldFD->getType()->getAs<FunctionProtoType>(), OldFD->getLocation(),
-            NewFD->getType()->getAs<FunctionProtoType>(), NewFD->getLocation()))
+    if (CheckEquivalentExceptionSpec(OldFPT, OldFD->getLocation(), NewFPT,
+                                     NewFD->getLocation()))
       return true;
   }
   return false;

diff --git a/clang/lib/Sema/SemaDeclAttr.cpp b/clang/lib/Sema/SemaDeclAttr.cpp
@@ -3034,9 +3034,6 @@ bool Sema::checkTargetVersionAttr(SourceLocation LiteralLoc, Decl *D,
       return Diag(LiteralLoc, diag::warn_unsupported_target_attribute)
              << Unsupported << None << CurFeature << TargetVersion;
   }
-  if (IsArmStreamingFunction(cast<FunctionDecl>(D),
-                             /*IncludeLocallyStreaming=*/false))
-    return Diag(LiteralLoc, diag::err_sme_streaming_cannot_be_multiversioned);
   return false;
 }
 
@@ -3133,10 +3130,6 @@ bool Sema::checkTargetClonesAttrString(
           HasNotDefault = true;
         }
       }
-      if (IsArmStreamingFunction(cast<FunctionDecl>(D),
-                                 /*IncludeLocallyStreaming=*/false))
-        return Diag(LiteralLoc,
-                    diag::err_sme_streaming_cannot_be_multiversioned);
     } else {
       // Other targets ( currently X86 )
       if (Cur.starts_with("arch=")) {

diff --git a/clang/lib/Sema/SemaDeclCXX.cpp b/clang/lib/Sema/SemaDeclCXX.cpp
@@ -10385,7 +10385,7 @@ void Sema::checkIncorrectVTablePointerAuthenticationAttribute(
   while (1) {
     assert(PrimaryBase);
     const CXXRecordDecl *Base = nullptr;
-    for (auto BasePtr : PrimaryBase->bases()) {
+    for (const CXXBaseSpecifier &BasePtr : PrimaryBase->bases()) {
       if (!BasePtr.getType()->getAsCXXRecordDecl()->isDynamicClass())
         continue;
       Base = BasePtr.getType()->getAsCXXRecordDecl();

diff --git a/clang/lib/Sema/SemaExprCXX.cpp b/clang/lib/Sema/SemaExprCXX.cpp
@@ -6030,6 +6030,32 @@ static bool EvaluateBinaryTypeTrait(Sema &Self, TypeTrait BTT, const TypeSourceI
     return cast<CXXRecordDecl>(rhsRecord->getDecl())
       ->isDerivedFrom(cast<CXXRecordDecl>(lhsRecord->getDecl()));
   }
+  case BTT_IsVirtualBaseOf: {
+    const RecordType *BaseRecord = LhsT->getAs<RecordType>();
+    const RecordType *DerivedRecord = RhsT->getAs<RecordType>();
+
+    if (!BaseRecord || !DerivedRecord) {
+      DiagnoseVLAInCXXTypeTrait(Self, Lhs,
+                                tok::kw___builtin_is_virtual_base_of);
+      DiagnoseVLAInCXXTypeTrait(Self, Rhs,
+                                tok::kw___builtin_is_virtual_base_of);
+      return false;
+    }
+
+    if (BaseRecord->isUnionType() || DerivedRecord->isUnionType())
+      return false;
+
+    if (!BaseRecord->isStructureOrClassType() ||
+        !DerivedRecord->isStructureOrClassType())
+      return false;
+
+    if (Self.RequireCompleteType(Rhs->getTypeLoc().getBeginLoc(), RhsT,
+                                 diag::err_incomplete_type))
+      return false;
+
+    return cast<CXXRecordDecl>(DerivedRecord->getDecl())
+        ->isVirtuallyDerivedFrom(cast<CXXRecordDecl>(BaseRecord->getDecl()));
+  }
   case BTT_IsSame:
     return Self.Context.hasSameType(LhsT, RhsT);
   case BTT_TypeCompatible: {

diff --git a/clang/lib/Sema/SemaTemplateDeduction.cpp b/clang/lib/Sema/SemaTemplateDeduction.cpp
@@ -951,9 +951,11 @@ class PackDeductionScope {
 
     // Skip over the pack elements that were expanded into separate arguments.
     // If we partially expanded, this is the number of partial arguments.
+    // FIXME: `&& FixedNumExpansions` is a workaround for UB described in
+    // https://github.com/llvm/llvm-project/issues/100095
     if (IsPartiallyExpanded)
       PackElements += NumPartialPackArgs;
-    else if (IsExpanded)
+    else if (IsExpanded && FixedNumExpansions)
       PackElements += *FixedNumExpansions;
 
     for (auto &Pack : Packs) {

diff --git a/clang/lib/StaticAnalyzer/Checkers/MmapWriteExecChecker.cpp b/clang/lib/StaticAnalyzer/Checkers/MmapWriteExecChecker.cpp
@@ -21,49 +21,56 @@
 #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/CheckerHelpers.h"
 
 using namespace clang;
 using namespace ento;
 
 namespace {
-class MmapWriteExecChecker : public Checker<check::PreCall> {
+class MmapWriteExecChecker
+    : public Checker<check::ASTDecl<TranslationUnitDecl>, check::PreCall> {
   CallDescription MmapFn{CDM::CLibrary, {"mmap"}, 6};
   CallDescription MprotectFn{CDM::CLibrary, {"mprotect"}, 3};
-  static int ProtWrite;
-  static int ProtExec;
-  static int ProtRead;
   const BugType BT{this, "W^X check fails, Write Exec prot flags set",
                    "Security"};
 
+  // Default values are used if definition of the flags is not found.
+  mutable int ProtRead = 0x01;
+  mutable int ProtWrite = 0x02;
+  mutable int ProtExec = 0x04;
+
 public:
+  void checkASTDecl(const TranslationUnitDecl *TU, AnalysisManager &Mgr,
+                    BugReporter &BR) const;
   void checkPreCall(const CallEvent &Call, CheckerContext &C) const;
-  int ProtExecOv;
-  int ProtReadOv;
 };
 }
 
-int MmapWriteExecChecker::ProtWrite = 0x02;
-int MmapWriteExecChecker::ProtExec  = 0x04;
-int MmapWriteExecChecker::ProtRead  = 0x01;
+void MmapWriteExecChecker::checkASTDecl(const TranslationUnitDecl *TU,
+                                        AnalysisManager &Mgr,
+                                        BugReporter &BR) const {
+  Preprocessor &PP = Mgr.getPreprocessor();
+  const std::optional<int> FoundProtRead = tryExpandAsInteger("PROT_READ", PP);
+  const std::optional<int> FoundProtWrite =
+      tryExpandAsInteger("PROT_WRITE", PP);
+  const std::optional<int> FoundProtExec = tryExpandAsInteger("PROT_EXEC", PP);
+  if (FoundProtRead && FoundProtWrite && FoundProtExec) {
+    ProtRead = *FoundProtRead;
+    ProtWrite = *FoundProtWrite;
+    ProtExec = *FoundProtExec;
+  }
+}
 
 void MmapWriteExecChecker::checkPreCall(const CallEvent &Call,
-                                         CheckerContext &C) const {
+                                        CheckerContext &C) const {
   if (matchesAny(Call, MmapFn, MprotectFn)) {
     SVal ProtVal = Call.getArgSVal(2);
     auto ProtLoc = ProtVal.getAs<nonloc::ConcreteInt>();
     if (!ProtLoc)
       return;
     int64_t Prot = ProtLoc->getValue().getSExtValue();
-    if (ProtExecOv != ProtExec)
-      ProtExec = ProtExecOv;
-    if (ProtReadOv != ProtRead)
-      ProtRead = ProtReadOv;
-
-    // Wrong settings
-    if (ProtRead == ProtExec)
-      return;
 
-    if ((Prot & (ProtWrite | ProtExec)) == (ProtWrite | ProtExec)) {
+    if ((Prot & ProtWrite) && (Prot & ProtExec)) {
       ExplodedNode *N = C.generateNonFatalErrorNode();
       if (!N)
         return;
@@ -80,17 +87,10 @@ void MmapWriteExecChecker::checkPreCall(const CallEvent &Call,
   }
 }
 
-void ento::registerMmapWriteExecChecker(CheckerManager &mgr) {
-  MmapWriteExecChecker *Mwec =
-      mgr.registerChecker<MmapWriteExecChecker>();
-  Mwec->ProtExecOv =
-    mgr.getAnalyzerOptions()
-      .getCheckerIntegerOption(Mwec, "MmapProtExec");
-  Mwec->ProtReadOv =
-    mgr.getAnalyzerOptions()
-      .getCheckerIntegerOption(Mwec, "MmapProtRead");
+void ento::registerMmapWriteExecChecker(CheckerManager &Mgr) {
+  Mgr.registerChecker<MmapWriteExecChecker>();
 }
 
-bool ento::shouldRegisterMmapWriteExecChecker(const CheckerManager &mgr) {
+bool ento::shouldRegisterMmapWriteExecChecker(const CheckerManager &) {
   return true;
 }
diff --git a/clang/test/AST/Interp/builtins.cpp b/clang/test/AST/Interp/builtins.cpp
@@ -31,3 +31,8 @@ constexpr bool assume() {
   return true;
 }
 static_assert(assume(), "");
+
+void test_builtin_os_log(void *buf, int i, const char *data) {
+  constexpr int len = __builtin_os_log_format_buffer_size("%d %{public}s %{private}.16P", i, data, data);
+  static_assert(len > 0, "Expect len > 0");
+}
diff --git a/clang/test/Analysis/analyzer-config.c b/clang/test/Analysis/analyzer-config.c
@@ -9,8 +9,6 @@
 // CHECK-NEXT: alpha.clone.CloneChecker:ReportNormalClones = true
 // CHECK-NEXT: alpha.cplusplus.STLAlgorithmModeling:AggressiveStdFindModeling = false
 // CHECK-NEXT: alpha.osx.cocoa.DirectIvarAssignment:AnnotatedFunctions = false
-// CHECK-NEXT: alpha.security.MmapWriteExec:MmapProtExec = 0x04
-// CHECK-NEXT: alpha.security.MmapWriteExec:MmapProtRead = 0x01
 // CHECK-NEXT: alpha.security.taint.TaintPropagation:Config = ""
 // CHECK-NEXT: apply-fixits = false
 // CHECK-NEXT: assume-controlled-environment = false

diff --git a/clang/test/Analysis/mmap-writeexec.c b/clang/test/Analysis/mmap-writeexec.c
@@ -1,13 +1,14 @@
-// RUN: %clang_analyze_cc1 -triple i686-unknown-linux -analyzer-checker=alpha.security.MmapWriteExec -analyzer-config alpha.security.MmapWriteExec:MmapProtExec=1 -analyzer-config alpha.security.MmapWriteExec:MmapProtRead=4 -DUSE_ALTERNATIVE_PROT_EXEC_DEFINITION -verify %s
+// RUN: %clang_analyze_cc1 -triple i686-unknown-linux -analyzer-checker=alpha.security.MmapWriteExec -DUSE_ALTERNATIVE_PROT_EXEC_DEFINITION -verify %s
 // RUN: %clang_analyze_cc1 -triple x86_64-unknown-apple-darwin10 -analyzer-checker=alpha.security.MmapWriteExec -verify %s
 
-#define PROT_WRITE  0x02
 #ifndef USE_ALTERNATIVE_PROT_EXEC_DEFINITION
-#define PROT_EXEC   0x04
-#define PROT_READ   0x01
-#else
 #define PROT_EXEC   0x01
+#define PROT_WRITE  0x02
 #define PROT_READ   0x04
+#else
+#define PROT_EXEC   0x08
+#define PROT_WRITE  0x04
+#define PROT_READ   0x02
 #endif
 #define MAP_PRIVATE 0x0002
 #define MAP_ANON    0x1000

diff --git a/clang/test/CodeGen/AMDGPU/amdgpu-atomic-float.c b/clang/test/CodeGen/AMDGPU/amdgpu-atomic-float.c
diff --git a/clang/test/CodeGen/aarch64-fmv-streaming.c b/clang/test/CodeGen/aarch64-fmv-streaming.c
@@ -0,0 +1,107 @@
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sme -emit-llvm -o - %s | FileCheck %s
+
+
+// CHECK-LABEL: define {{[^@]+}}@n_callee._Msve
+// CHECK-SAME: () #[[ATTR0:[0-9]+]] {
+//
+// CHECK-LABEL: define {{[^@]+}}@n_callee._Msimd
+// CHECK-SAME: () #[[ATTR1:[0-9]+]] {
+//
+__arm_locally_streaming __attribute__((target_clones("sve", "simd"))) void n_callee(void) {}
+// CHECK-LABEL: define {{[^@]+}}@n_callee._Msme2
+// CHECK-SAME: () #[[ATTR2:[0-9]+]] {
+//
+__attribute__((target_version("sme2"))) void n_callee(void) {}
+// CHECK-LABEL: define {{[^@]+}}@n_callee.default
+// CHECK-SAME: () #[[ATTR3:[0-9]+]] {
+//
+__attribute__((target_version("default"))) void n_callee(void) {}
+
+
+// CHECK-LABEL: define {{[^@]+}}@s_callee._Msve
+// CHECK-SAME: () #[[ATTR4:[0-9]+]] {
+//
+// CHECK-LABEL: define {{[^@]+}}@s_callee._Msimd
+// CHECK-SAME: () #[[ATTR5:[0-9]+]] {
+//
+__attribute__((target_clones("sve", "simd"))) void s_callee(void) __arm_streaming {}
+// CHECK-LABEL: define {{[^@]+}}@s_callee._Msme2
+// CHECK-SAME: () #[[ATTR6:[0-9]+]] {
+//
+__arm_locally_streaming __attribute__((target_version("sme2"))) void s_callee(void) __arm_streaming {}
+// CHECK-LABEL: define {{[^@]+}}@s_callee.default
+// CHECK-SAME: () #[[ATTR7:[0-9]+]] {
+//
+__attribute__((target_version("default"))) void s_callee(void) __arm_streaming {}
+
+
+// CHECK-LABEL: define {{[^@]+}}@sc_callee._Msve
+// CHECK-SAME: () #[[ATTR8:[0-9]+]] {
+//
+// CHECK-LABEL: define {{[^@]+}}@sc_callee._Msimd
+// CHECK-SAME: () #[[ATTR9:[0-9]+]] {
+//
+__attribute__((target_clones("sve", "simd"))) void sc_callee(void) __arm_streaming_compatible {}
+// CHECK-LABEL: define {{[^@]+}}@sc_callee._Msme2
+// CHECK-SAME: () #[[ATTR10:[0-9]+]] {
+//
+__arm_locally_streaming __attribute__((target_version("sme2"))) void sc_callee(void) __arm_streaming_compatible {}
+// CHECK-LABEL: define {{[^@]+}}@sc_callee.default
+// CHECK-SAME: () #[[ATTR11:[0-9]+]] {
+//
+__attribute__((target_version("default"))) void sc_callee(void) __arm_streaming_compatible {}
+
+
+// CHECK-LABEL: define {{[^@]+}}@n_caller
+// CHECK-SAME: () #[[ATTR3:[0-9]+]] {
+// CHECK:    call void @n_callee()
+// CHECK:    call void @s_callee() #[[ATTR12:[0-9]+]]
+// CHECK:    call void @sc_callee() #[[ATTR13:[0-9]+]]
+//
+void n_caller(void) {
+  n_callee();
+  s_callee();
+  sc_callee();
+}
+
+
+// CHECK-LABEL: define {{[^@]+}}@s_caller
+// CHECK-SAME: () #[[ATTR7:[0-9]+]] {
+// CHECK:    call void @n_callee()
+// CHECK:    call void @s_callee() #[[ATTR12]]
+// CHECK:    call void @sc_callee() #[[ATTR13]]
+//
+void s_caller(void) __arm_streaming {
+  n_callee();
+  s_callee();
+  sc_callee();
+}
+
+
+// CHECK-LABEL: define {{[^@]+}}@sc_caller
+// CHECK-SAME: () #[[ATTR11:[0-9]+]] {
+// CHECK:    call void @n_callee()
+// CHECK:    call void @s_callee() #[[ATTR12]]
+// CHECK:    call void @sc_callee() #[[ATTR13]]
+//
+void sc_caller(void) __arm_streaming_compatible {
+  n_callee();
+  s_callee();
+  sc_callee();
+}
+
+
+// CHECK: attributes #[[ATTR0:[0-9]+]] = {{.*}} "aarch64_pstate_sm_body"
+// CHECK: attributes #[[ATTR1:[0-9]+]] = {{.*}} "aarch64_pstate_sm_body"
+// CHECK: attributes #[[ATTR2:[0-9]+]] = {{.*}}
+// CHECK: attributes #[[ATTR3]] = {{.*}}
+// CHECK: attributes #[[ATTR4:[0-9]+]] = {{.*}} "aarch64_pstate_sm_enabled"
+// CHECK: attributes #[[ATTR5:[0-9]+]] = {{.*}} "aarch64_pstate_sm_enabled"
+// CHECK: attributes #[[ATTR6:[0-9]+]] = {{.*}} "aarch64_pstate_sm_body" "aarch64_pstate_sm_enabled"
+// CHECK: attributes #[[ATTR7]] = {{.*}} "aarch64_pstate_sm_enabled"
+// CHECK: attributes #[[ATTR8:[0-9]+]] = {{.*}} "aarch64_pstate_sm_compatible"
+// CHECK: attributes #[[ATTR9:[0-9]+]] = {{.*}} "aarch64_pstate_sm_compatible"
+// CHECK: attributes #[[ATTR10]] = {{.*}} "aarch64_pstate_sm_body" "aarch64_pstate_sm_compatible"
+// CHECK: attributes #[[ATTR11]] = {{.*}} "aarch64_pstate_sm_compatible"
+// CHECK: attributes #[[ATTR12]] = {{.*}} "aarch64_pstate_sm_enabled"
+// CHECK: attributes #[[ATTR13]] = {{.*}} "aarch64_pstate_sm_compatible"
diff --git a/clang/test/CodeGen/finite-math.c b/clang/test/CodeGen/finite-math.c
@@ -1,4 +1,4 @@
-// RUN: %clang_cc1 -ffinite-math-only -emit-llvm -o - %s | FileCheck %s -check-prefix=CHECK -check-prefix=FINITE
+// RUN: %clang_cc1 -menable-no-infs -menable-no-nans -emit-llvm -o - %s | FileCheck %s -check-prefix=CHECK -check-prefix=FINITE
 // RUN: %clang_cc1 -fno-signed-zeros -emit-llvm -o - %s | FileCheck %s -check-prefix=CHECK  -check-prefix=NSZ
 // RUN: %clang_cc1 -freciprocal-math -emit-llvm -o - %s | FileCheck %s -check-prefix=CHECK  -check-prefix=RECIP
 // RUN: %clang_cc1 -mreassociate -emit-llvm -o - %s | FileCheck %s -check-prefix=CHECK  -check-prefix=REASSOC

diff --git a/clang/test/CodeGen/fp-floatcontrol-stack.cpp b/clang/test/CodeGen/fp-floatcontrol-stack.cpp
@@ -1,7 +1,7 @@
 // RUN: %clang_cc1 -triple x86_64-linux-gnu -ffp-contract=on -DDEFAULT=1 -emit-llvm -o - %s | FileCheck --check-prefix=CHECK-DDEFAULT %s
 // RUN: %clang_cc1 -triple x86_64-linux-gnu -ffp-contract=on -DEBSTRICT=1 -ffp-exception-behavior=strict -emit-llvm -o - %s | FileCheck --check-prefix=CHECK-DEBSTRICT %s
 // RUN: %clang_cc1 -triple x86_64-linux-gnu -DFAST=1 -ffast-math -ffp-contract=fast -emit-llvm -o - %s | FileCheck --check-prefix=CHECK-FAST %s
-// RUN: %clang_cc1 -triple x86_64-linux-gnu -ffp-contract=on -DNOHONOR=1 -menable-no-infs -menable-no-nans -emit-llvm -o - %s | FileCheck --check-prefix=CHECK-NOHONOR %s
+// RUN: %clang_cc1 -triple x86_64-linux-gnu -ffp-contract=on -DNOHONOR=1 -ffinite-math-only -menable-no-infs -menable-no-nans -emit-llvm -o - %s | FileCheck --check-prefix=CHECK-NOHONOR %s
 
 #define FUN(n) \
   (float z) { return n * z + n; }

diff --git a/clang/test/CodeGen/fp-options-to-fast-math-flags.c b/clang/test/CodeGen/fp-options-to-fast-math-flags.c
@@ -1,7 +1,7 @@
 // RUN: %clang_cc1 -triple x86_64-unknown-unknown -emit-llvm -o - %s | FileCheck -check-prefix CHECK-PRECISE %s
 // RUN: %clang_cc1 -triple x86_64-unknown-unknown -menable-no-nans -emit-llvm -o - %s | FileCheck -check-prefix CHECK-NO-NANS %s
 // RUN: %clang_cc1 -triple x86_64-unknown-unknown -menable-no-infs -emit-llvm -o - %s | FileCheck -check-prefix CHECK-NO-INFS %s
-// RUN: %clang_cc1 -triple x86_64-unknown-unknown -ffinite-math-only -emit-llvm -o - %s | FileCheck -check-prefix CHECK-FINITE %s
+// RUN: %clang_cc1 -triple x86_64-unknown-unknown -menable-no-infs -menable-no-nans -emit-llvm -o - %s | FileCheck -check-prefix CHECK-FINITE %s
 // RUN: %clang_cc1 -triple x86_64-unknown-unknown -fno-signed-zeros -emit-llvm -o - %s | FileCheck -check-prefix CHECK-NO-SIGNED-ZEROS %s
 // RUN: %clang_cc1 -triple x86_64-unknown-unknown -mreassociate -emit-llvm -o - %s | FileCheck -check-prefix CHECK-REASSOC %s
 // RUN: %clang_cc1 -triple x86_64-unknown-unknown -freciprocal-math -emit-llvm -o - %s | FileCheck -check-prefix CHECK-RECIP %s

diff --git a/clang/test/CodeGen/nofpclass.c b/clang/test/CodeGen/nofpclass.c
@@ -1,7 +1,7 @@
 // NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --check-attributes --version 2
 // REQUIRES: x86-registered-target
-// RUN: %clang_cc1 -triple x86_64-unknown-unknown -target-feature +avx -fenable-matrix -ffinite-math-only -emit-llvm -o - %s | FileCheck -check-prefixes=CFINITEONLY %s
-// RUN: %clang_cc1 -x cl -triple x86_64-unknown-unknown -target-feature +avx -fenable-matrix -cl-finite-math-only -emit-llvm -o - %s | FileCheck -check-prefixes=CLFINITEONLY %s
+// RUN: %clang_cc1 -triple x86_64-unknown-unknown -target-feature +avx -fenable-matrix -menable-no-infs -menable-no-nans -emit-llvm -o - %s | FileCheck -check-prefixes=CFINITEONLY %s
+// RUN: %clang_cc1 -x cl -triple x86_64-unknown-unknown -target-feature +avx -fenable-matrix -menable-no-nans -menable-no-infs -emit-llvm -o - %s | FileCheck -check-prefixes=CLFINITEONLY %s
 
 // RUN: %clang_cc1 -triple x86_64-unknown-unknown -target-feature +avx -fenable-matrix -menable-no-nans -emit-llvm -o - %s | FileCheck -check-prefixes=NONANS %s
 // RUN: %clang_cc1 -triple x86_64-unknown-unknown -target-feature +avx -fenable-matrix -menable-no-infs -emit-llvm -o - %s | FileCheck -check-prefixes=NOINFS %s

diff --git a/clang/test/CodeGenCUDA/amdgpu-atomic-ops.cu b/clang/test/CodeGenCUDA/amdgpu-atomic-ops.cu
@@ -1,6 +1,10 @@
 // RUN: %clang_cc1 -x hip %s -emit-llvm -o - -triple=amdgcn-amd-amdhsa \
 // RUN:   -fcuda-is-device -target-cpu gfx906 -fnative-half-type \
-// RUN:   -fnative-half-arguments-and-returns | FileCheck %s
+// RUN:   -fnative-half-arguments-and-returns | FileCheck -check-prefixes=CHECK,SAFEIR %s
+
+// RUN: %clang_cc1 -x hip %s -emit-llvm -o - -triple=amdgcn-amd-amdhsa \
+// RUN:   -fcuda-is-device -target-cpu gfx906 -fnative-half-type \
+// RUN:   -fnative-half-arguments-and-returns -munsafe-fp-atomics | FileCheck -check-prefixes=CHECK,UNSAFEIR %s
 
 // RUN: %clang_cc1 -x hip %s -O3 -S -o - -triple=amdgcn-amd-amdhsa \
 // RUN:   -fcuda-is-device -target-cpu gfx1100 -fnative-half-type \
@@ -18,24 +22,38 @@
 
 __global__ void ffp1(float *p) {
   // CHECK-LABEL: @_Z4ffp1Pf
-  // CHECK: atomicrmw fadd ptr {{.*}} monotonic
-  // CHECK: atomicrmw fmax ptr {{.*}} monotonic
-  // CHECK: atomicrmw fmin ptr {{.*}} monotonic
-  // CHECK: atomicrmw fmax ptr {{.*}} syncscope("agent-one-as") monotonic
-  // CHECK: atomicrmw fmin ptr {{.*}} syncscope("workgroup-one-as") monotonic
+  // SAFEIR: atomicrmw fadd ptr {{.*}} monotonic, align 4{{$}}
+  // SAFEIR: atomicrmw fsub ptr {{.*}} monotonic, align 4{{$}}
+  // SAFEIR: atomicrmw fmax ptr {{.*}} monotonic, align 4{{$}}
+  // SAFEIR: atomicrmw fmin ptr {{.*}} monotonic, align 4{{$}}
+  // SAFEIR: atomicrmw fmax ptr {{.*}} syncscope("agent-one-as") monotonic, align 4{{$}}
+  // SAFEIR: atomicrmw fmin ptr {{.*}} syncscope("workgroup-one-as") monotonic, align 4{{$}}
+
+  // UNSAFEIR: atomicrmw fadd ptr {{.*}} monotonic, align 4, !amdgpu.no.fine.grained.memory !{{[0-9]+}}, !amdgpu.ignore.denormal.mode !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fsub ptr {{.*}} monotonic, align 4, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fmax ptr {{.*}} monotonic, align 4, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fmin ptr {{.*}} monotonic, align 4, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fmax ptr {{.*}} syncscope("agent-one-as") monotonic, align 4, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fmin ptr {{.*}} syncscope("workgroup-one-as") monotonic, align 4, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+
   // SAFE: _Z4ffp1Pf
   // SAFE: global_atomic_cmpswap
   // SAFE: global_atomic_cmpswap
   // SAFE: global_atomic_cmpswap
   // SAFE: global_atomic_cmpswap
   // SAFE: global_atomic_cmpswap
+  // SAFE: global_atomic_cmpswap
+
   // UNSAFE: _Z4ffp1Pf
   // UNSAFE: global_atomic_add_f32
   // UNSAFE: global_atomic_cmpswap
   // UNSAFE: global_atomic_cmpswap
   // UNSAFE: global_atomic_cmpswap
   // UNSAFE: global_atomic_cmpswap
+  // UNSAFE: global_atomic_cmpswap
+
   __atomic_fetch_add(p, 1.0f, memory_order_relaxed);
+  __atomic_fetch_sub(p, 1.0f, memory_order_relaxed);
   __atomic_fetch_max(p, 1.0f, memory_order_relaxed);
   __atomic_fetch_min(p, 1.0f, memory_order_relaxed);
   __hip_atomic_fetch_max(p, 1.0f, memory_order_relaxed, __HIP_MEMORY_SCOPE_AGENT);
@@ -44,23 +62,36 @@ __global__ void ffp1(float *p) {
 
 __global__ void ffp2(double *p) {
   // CHECK-LABEL: @_Z4ffp2Pd
-  // CHECK: atomicrmw fsub ptr {{.*}} monotonic
-  // CHECK: atomicrmw fmax ptr {{.*}} monotonic
-  // CHECK: atomicrmw fmin ptr {{.*}} monotonic
-  // CHECK: atomicrmw fmax ptr {{.*}} syncscope("agent-one-as") monotonic
-  // CHECK: atomicrmw fmin ptr {{.*}} syncscope("workgroup-one-as") monotonic
+  // SAFEIR: atomicrmw fadd ptr {{.*}} monotonic, align 8{{$}}
+  // SAFEIR: atomicrmw fsub ptr {{.*}} monotonic, align 8{{$}}
+  // SAFEIR: atomicrmw fmax ptr {{.*}} monotonic, align 8{{$}}
+  // SAFEIR: atomicrmw fmin ptr {{.*}} monotonic, align 8{{$}}
+  // SAFEIR: atomicrmw fmax ptr {{.*}} syncscope("agent-one-as") monotonic, align 8{{$}}
+  // SAFEIR: atomicrmw fmin ptr {{.*}} syncscope("workgroup-one-as") monotonic, align 8{{$}}
+
+  // UNSAFEIR: atomicrmw fadd ptr {{.*}} monotonic, align 8, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fsub ptr {{.*}} monotonic, align 8, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fmax ptr {{.*}} monotonic, align 8, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fmin ptr {{.*}} monotonic, align 8, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fmax ptr {{.*}} syncscope("agent-one-as") monotonic, align 8, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fmin ptr {{.*}} syncscope("workgroup-one-as") monotonic, align 8, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+
   // SAFE-LABEL: @_Z4ffp2Pd
   // SAFE: global_atomic_cmpswap_b64
   // SAFE: global_atomic_cmpswap_b64
   // SAFE: global_atomic_cmpswap_b64
   // SAFE: global_atomic_cmpswap_b64
   // SAFE: global_atomic_cmpswap_b64
+  // SAFE: global_atomic_cmpswap_b64
+
   // UNSAFE-LABEL: @_Z4ffp2Pd
+  // UNSAFE: global_atomic_add_f64
   // UNSAFE: global_atomic_cmpswap_x2
   // UNSAFE: global_atomic_cmpswap_x2
   // UNSAFE: global_atomic_cmpswap_x2
   // UNSAFE: global_atomic_max_f64
   // UNSAFE: global_atomic_min_f64
+  __atomic_fetch_add(p, 1.0, memory_order_relaxed);
   __atomic_fetch_sub(p, 1.0, memory_order_relaxed);
   __atomic_fetch_max(p, 1.0, memory_order_relaxed);
   __atomic_fetch_min(p, 1.0, memory_order_relaxed);
@@ -71,11 +102,20 @@ __global__ void ffp2(double *p) {
 // long double is the same as double for amdgcn.
 __global__ void ffp3(long double *p) {
   // CHECK-LABEL: @_Z4ffp3Pe
-  // CHECK: atomicrmw fsub ptr {{.*}} monotonic
-  // CHECK: atomicrmw fmax ptr {{.*}} monotonic
-  // CHECK: atomicrmw fmin ptr {{.*}} monotonic
-  // CHECK: atomicrmw fmax ptr {{.*}} syncscope("agent-one-as") monotonic
-  // CHECK: atomicrmw fmin ptr {{.*}} syncscope("workgroup-one-as") monotonic
+  // SAFEIR: atomicrmw fadd ptr {{.*}} monotonic, align 8{{$}}
+  // SAFEIR: atomicrmw fsub ptr {{.*}} monotonic, align 8{{$}}
+  // SAFEIR: atomicrmw fmax ptr {{.*}} monotonic, align 8{{$}}
+  // SAFEIR: atomicrmw fmin ptr {{.*}} monotonic, align 8{{$}}
+  // SAFEIR: atomicrmw fmax ptr {{.*}} syncscope("agent-one-as") monotonic, align 8{{$}}
+  // SAFEIR: atomicrmw fmin ptr {{.*}} syncscope("workgroup-one-as") monotonic, align 8{{$}}
+
+  // UNSAFEIR: atomicrmw fadd ptr {{.*}} monotonic, align 8, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fsub ptr {{.*}} monotonic, align 8, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fmax ptr {{.*}} monotonic, align 8, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fmin ptr {{.*}} monotonic, align 8, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fmax ptr {{.*}} syncscope("agent-one-as") monotonic, align 8, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fmin ptr {{.*}} syncscope("workgroup-one-as") monotonic, align 8, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+
   // SAFE-LABEL: @_Z4ffp3Pe
   // SAFE: global_atomic_cmpswap_b64
   // SAFE: global_atomic_cmpswap_b64
@@ -88,6 +128,7 @@ __global__ void ffp3(long double *p) {
   // UNSAFE: global_atomic_cmpswap_x2
   // UNSAFE: global_atomic_max_f64
   // UNSAFE: global_atomic_min_f64
+  __atomic_fetch_add(p, 1.0L, memory_order_relaxed);
   __atomic_fetch_sub(p, 1.0L, memory_order_relaxed);
   __atomic_fetch_max(p, 1.0L, memory_order_relaxed);
   __atomic_fetch_min(p, 1.0L, memory_order_relaxed);
@@ -98,37 +139,52 @@ __global__ void ffp3(long double *p) {
 __device__ double ffp4(double *p, float f) {
   // CHECK-LABEL: @_Z4ffp4Pdf
   // CHECK: fpext float {{.*}} to double
-  // CHECK: atomicrmw fsub ptr {{.*}} monotonic
+  // SAFEIR: atomicrmw fsub ptr {{.*}} monotonic, align 8{{$}}
+  // UNSAFEIR: atomicrmw fsub ptr {{.*}} monotonic, align 8, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
   return __atomic_fetch_sub(p, f, memory_order_relaxed);
 }
 
 __device__ double ffp5(double *p, int i) {
   // CHECK-LABEL: @_Z4ffp5Pdi
   // CHECK: sitofp i32 {{.*}} to double
-  // CHECK: atomicrmw fsub ptr {{.*}} monotonic
+  // SAFEIR: atomicrmw fsub ptr {{.*}} monotonic, align 8{{$}}
+  // UNSAFEIR: atomicrmw fsub ptr {{.*}} monotonic, align 8, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
   return __atomic_fetch_sub(p, i, memory_order_relaxed);
 }
 
 __global__ void ffp6(_Float16 *p) {
   // CHECK-LABEL: @_Z4ffp6PDF16
-  // CHECK: atomicrmw fadd ptr {{.*}} monotonic
-  // CHECK: atomicrmw fmax ptr {{.*}} monotonic
-  // CHECK: atomicrmw fmin ptr {{.*}} monotonic
-  // CHECK: atomicrmw fmax ptr {{.*}} syncscope("agent-one-as") monotonic
-  // CHECK: atomicrmw fmin ptr {{.*}} syncscope("workgroup-one-as") monotonic
+  // SAFEIR: atomicrmw fadd ptr {{.*}} monotonic, align 2{{$}}
+  // SAFEIR: atomicrmw fsub ptr {{.*}} monotonic, align 2{{$}}
+  // SAFEIR: atomicrmw fmax ptr {{.*}} monotonic, align 2{{$}}
+  // SAFEIR: atomicrmw fmin ptr {{.*}} monotonic, align 2{{$}}
+  // SAFEIR: atomicrmw fmax ptr {{.*}} syncscope("agent-one-as") monotonic, align 2{{$}}
+  // SAFEIR: atomicrmw fmin ptr {{.*}} syncscope("workgroup-one-as") monotonic, align 2{{$}}
+
+  // UNSAFEIR: atomicrmw fadd ptr {{.*}} monotonic, align 2, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fsub ptr {{.*}} monotonic, align 2, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fmax ptr {{.*}} monotonic, align 2, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fmin ptr {{.*}} monotonic, align 2, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fmax ptr {{.*}} syncscope("agent-one-as") monotonic, align 2, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+  // UNSAFEIR: atomicrmw fmin ptr {{.*}} syncscope("workgroup-one-as") monotonic, align 2, !amdgpu.no.fine.grained.memory !{{[0-9]+$}}
+
   // SAFE: _Z4ffp6PDF16
   // SAFE: global_atomic_cmpswap
   // SAFE: global_atomic_cmpswap
   // SAFE: global_atomic_cmpswap
   // SAFE: global_atomic_cmpswap
   // SAFE: global_atomic_cmpswap
+  // SAFE: global_atomic_cmpswap
+
   // UNSAFE: _Z4ffp6PDF16
   // UNSAFE: global_atomic_cmpswap
   // UNSAFE: global_atomic_cmpswap
   // UNSAFE: global_atomic_cmpswap
   // UNSAFE: global_atomic_cmpswap
   // UNSAFE: global_atomic_cmpswap
+  // UNSAFE: global_atomic_cmpswap
   __atomic_fetch_add(p, 1.0, memory_order_relaxed);
+  __atomic_fetch_sub(p, 1.0, memory_order_relaxed);
   __atomic_fetch_max(p, 1.0, memory_order_relaxed);
   __atomic_fetch_min(p, 1.0, memory_order_relaxed);
   __hip_atomic_fetch_max(p, 1.0f, memory_order_relaxed, __HIP_MEMORY_SCOPE_AGENT);

diff --git a/clang/test/CodeGenOpenCL/builtins-alloca.cl b/clang/test/CodeGenOpenCL/builtins-alloca.cl
@@ -0,0 +1,141 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// RUN: %clang_cc1 %s -O0 -triple amdgcn-amd-amdhsa -cl-std=CL1.2 \
+// RUN:     -emit-llvm -o - | FileCheck --check-prefixes=OPENCL %s
+// RUN: %clang_cc1 %s -O0 -triple amdgcn-amd-amdhsa -cl-std=CL2.0 \
+// RUN:     -emit-llvm -o - | FileCheck --check-prefixes=OPENCL %s
+// RUN: %clang_cc1 %s -O0 -triple amdgcn-amd-amdhsa -cl-std=CL3.0 \
+// RUN:     -emit-llvm -o - | FileCheck --check-prefixes=OPENCL %s
+// RUN: %clang_cc1 %s -O0 -triple amdgcn-amd-amdhsa -cl-std=CL3.0 -cl-ext=+__opencl_c_generic_address_space \
+// RUN:     -emit-llvm -o - | FileCheck --check-prefixes=OPENCL %s
+
+// OPENCL-LABEL: define dso_local void @test1_builtin_alloca(
+// OPENCL-SAME: i32 noundef [[N:%.*]]) #[[ATTR0:[0-9]+]] {
+// OPENCL-NEXT:  [[ENTRY:.*:]]
+// OPENCL-NEXT:    [[N_ADDR:%.*]] = alloca i32, align 4, addrspace(5)
+// OPENCL-NEXT:    [[ALLOC_PTR:%.*]] = alloca ptr addrspace(5), align 4, addrspace(5)
+// OPENCL-NEXT:    store i32 [[N]], ptr addrspace(5) [[N_ADDR]], align 4
+// OPENCL-NEXT:    [[TMP0:%.*]] = load i32, ptr addrspace(5) [[N_ADDR]], align 4
+// OPENCL-NEXT:    [[CONV:%.*]] = zext i32 [[TMP0]] to i64
+// OPENCL-NEXT:    [[MUL:%.*]] = mul i64 [[CONV]], 4
+// OPENCL-NEXT:    [[TMP1:%.*]] = alloca i8, i64 [[MUL]], align 8, addrspace(5)
+// OPENCL-NEXT:    store ptr addrspace(5) [[TMP1]], ptr addrspace(5) [[ALLOC_PTR]], align 4
+// OPENCL-NEXT:    ret void
+//
+void test1_builtin_alloca(unsigned n) {
+    __private float* alloc_ptr = (__private float*)__builtin_alloca(n*sizeof(int));
+}
+
+// OPENCL-LABEL: define dso_local void @test1_builtin_alloca_uninitialized(
+// OPENCL-SAME: i32 noundef [[N:%.*]]) #[[ATTR0]] {
+// OPENCL-NEXT:  [[ENTRY:.*:]]
+// OPENCL-NEXT:    [[N_ADDR:%.*]] = alloca i32, align 4, addrspace(5)
+// OPENCL-NEXT:    [[ALLOC_PTR_UNINITIALIZED:%.*]] = alloca ptr addrspace(5), align 4, addrspace(5)
+// OPENCL-NEXT:    store i32 [[N]], ptr addrspace(5) [[N_ADDR]], align 4
+// OPENCL-NEXT:    [[TMP0:%.*]] = load i32, ptr addrspace(5) [[N_ADDR]], align 4
+// OPENCL-NEXT:    [[CONV:%.*]] = zext i32 [[TMP0]] to i64
+// OPENCL-NEXT:    [[MUL:%.*]] = mul i64 [[CONV]], 4
+// OPENCL-NEXT:    [[TMP1:%.*]] = alloca i8, i64 [[MUL]], align 8, addrspace(5)
+// OPENCL-NEXT:    store ptr addrspace(5) [[TMP1]], ptr addrspace(5) [[ALLOC_PTR_UNINITIALIZED]], align 4
+// OPENCL-NEXT:    ret void
+//
+void test1_builtin_alloca_uninitialized(unsigned n) {
+    __private float* alloc_ptr_uninitialized = (__private float*)__builtin_alloca_uninitialized(n*sizeof(int));
+}
+
+// OPENCL-LABEL: define dso_local void @test1_builtin_alloca_with_align(
+// OPENCL-SAME: i32 noundef [[N:%.*]]) #[[ATTR0]] {
+// OPENCL-NEXT:  [[ENTRY:.*:]]
+// OPENCL-NEXT:    [[N_ADDR:%.*]] = alloca i32, align 4, addrspace(5)
+// OPENCL-NEXT:    [[ALLOC_PTR_ALIGN:%.*]] = alloca ptr addrspace(5), align 4, addrspace(5)
+// OPENCL-NEXT:    store i32 [[N]], ptr addrspace(5) [[N_ADDR]], align 4
+// OPENCL-NEXT:    [[TMP0:%.*]] = load i32, ptr addrspace(5) [[N_ADDR]], align 4
+// OPENCL-NEXT:    [[CONV:%.*]] = zext i32 [[TMP0]] to i64
+// OPENCL-NEXT:    [[MUL:%.*]] = mul i64 [[CONV]], 4
+// OPENCL-NEXT:    [[TMP1:%.*]] = alloca i8, i64 [[MUL]], align 1, addrspace(5)
+// OPENCL-NEXT:    store ptr addrspace(5) [[TMP1]], ptr addrspace(5) [[ALLOC_PTR_ALIGN]], align 4
+// OPENCL-NEXT:    ret void
+//
+void test1_builtin_alloca_with_align(unsigned n) {
+    __private float* alloc_ptr_align = (__private float*)__builtin_alloca_with_align((n*sizeof(int)), 8);
+}
+
+// OPENCL-LABEL: define dso_local void @test1_builtin_alloca_with_align_uninitialized(
+// OPENCL-SAME: i32 noundef [[N:%.*]]) #[[ATTR0]] {
+// OPENCL-NEXT:  [[ENTRY:.*:]]
+// OPENCL-NEXT:    [[N_ADDR:%.*]] = alloca i32, align 4, addrspace(5)
+// OPENCL-NEXT:    [[ALLOC_PTR_ALIGN_UNINITIALIZED:%.*]] = alloca ptr addrspace(5), align 4, addrspace(5)
+// OPENCL-NEXT:    store i32 [[N]], ptr addrspace(5) [[N_ADDR]], align 4
+// OPENCL-NEXT:    [[TMP0:%.*]] = load i32, ptr addrspace(5) [[N_ADDR]], align 4
+// OPENCL-NEXT:    [[CONV:%.*]] = zext i32 [[TMP0]] to i64
+// OPENCL-NEXT:    [[MUL:%.*]] = mul i64 [[CONV]], 4
+// OPENCL-NEXT:    [[TMP1:%.*]] = alloca i8, i64 [[MUL]], align 1, addrspace(5)
+// OPENCL-NEXT:    store ptr addrspace(5) [[TMP1]], ptr addrspace(5) [[ALLOC_PTR_ALIGN_UNINITIALIZED]], align 4
+// OPENCL-NEXT:    ret void
+//
+void test1_builtin_alloca_with_align_uninitialized(unsigned n) {
+    __private float* alloc_ptr_align_uninitialized = (__private float*)__builtin_alloca_with_align_uninitialized((n*sizeof(int)), 8);
+}
+
+// OPENCL-LABEL: define dso_local void @test2_builtin_alloca(
+// OPENCL-SAME: i32 noundef [[N:%.*]]) #[[ATTR0]] {
+// OPENCL-NEXT:  [[ENTRY:.*:]]
+// OPENCL-NEXT:    [[N_ADDR:%.*]] = alloca i32, align 4, addrspace(5)
+// OPENCL-NEXT:    [[ALLOC_PTR:%.*]] = alloca ptr addrspace(5), align 4, addrspace(5)
+// OPENCL-NEXT:    store i32 [[N]], ptr addrspace(5) [[N_ADDR]], align 4
+// OPENCL-NEXT:    [[TMP0:%.*]] = load i32, ptr addrspace(5) [[N_ADDR]], align 4
+// OPENCL-NEXT:    [[CONV:%.*]] = zext i32 [[TMP0]] to i64
+// OPENCL-NEXT:    [[TMP1:%.*]] = alloca i8, i64 [[CONV]], align 8, addrspace(5)
+// OPENCL-NEXT:    store ptr addrspace(5) [[TMP1]], ptr addrspace(5) [[ALLOC_PTR]], align 4
+// OPENCL-NEXT:    ret void
+//
+void test2_builtin_alloca(unsigned n) {
+    __private void *alloc_ptr = __builtin_alloca(n);
+}
+
+// OPENCL-LABEL: define dso_local void @test2_builtin_alloca_uninitialized(
+// OPENCL-SAME: i32 noundef [[N:%.*]]) #[[ATTR0]] {
+// OPENCL-NEXT:  [[ENTRY:.*:]]
+// OPENCL-NEXT:    [[N_ADDR:%.*]] = alloca i32, align 4, addrspace(5)
+// OPENCL-NEXT:    [[ALLOC_PTR_UNINITIALIZED:%.*]] = alloca ptr addrspace(5), align 4, addrspace(5)
+// OPENCL-NEXT:    store i32 [[N]], ptr addrspace(5) [[N_ADDR]], align 4
+// OPENCL-NEXT:    [[TMP0:%.*]] = load i32, ptr addrspace(5) [[N_ADDR]], align 4
+// OPENCL-NEXT:    [[CONV:%.*]] = zext i32 [[TMP0]] to i64
+// OPENCL-NEXT:    [[TMP1:%.*]] = alloca i8, i64 [[CONV]], align 8, addrspace(5)
+// OPENCL-NEXT:    store ptr addrspace(5) [[TMP1]], ptr addrspace(5) [[ALLOC_PTR_UNINITIALIZED]], align 4
+// OPENCL-NEXT:    ret void
+//
+void test2_builtin_alloca_uninitialized(unsigned n) {
+    __private void *alloc_ptr_uninitialized = __builtin_alloca_uninitialized(n);
+}
+
+// OPENCL-LABEL: define dso_local void @test2_builtin_alloca_with_align(
+// OPENCL-SAME: i32 noundef [[N:%.*]]) #[[ATTR0]] {
+// OPENCL-NEXT:  [[ENTRY:.*:]]
+// OPENCL-NEXT:    [[N_ADDR:%.*]] = alloca i32, align 4, addrspace(5)
+// OPENCL-NEXT:    [[ALLOC_PTR_ALIGN:%.*]] = alloca ptr addrspace(5), align 4, addrspace(5)
+// OPENCL-NEXT:    store i32 [[N]], ptr addrspace(5) [[N_ADDR]], align 4
+// OPENCL-NEXT:    [[TMP0:%.*]] = load i32, ptr addrspace(5) [[N_ADDR]], align 4
+// OPENCL-NEXT:    [[CONV:%.*]] = zext i32 [[TMP0]] to i64
+// OPENCL-NEXT:    [[TMP1:%.*]] = alloca i8, i64 [[CONV]], align 1, addrspace(5)
+// OPENCL-NEXT:    store ptr addrspace(5) [[TMP1]], ptr addrspace(5) [[ALLOC_PTR_ALIGN]], align 4
+// OPENCL-NEXT:    ret void
+//
+void test2_builtin_alloca_with_align(unsigned n) {
+    __private void *alloc_ptr_align = __builtin_alloca_with_align(n, 8);
+}
+
+// OPENCL-LABEL: define dso_local void @test2_builtin_alloca_with_align_uninitialized(
+// OPENCL-SAME: i32 noundef [[N:%.*]]) #[[ATTR0]] {
+// OPENCL-NEXT:  [[ENTRY:.*:]]
+// OPENCL-NEXT:    [[N_ADDR:%.*]] = alloca i32, align 4, addrspace(5)
+// OPENCL-NEXT:    [[ALLOC_PTR_ALIGN_UNINITIALIZED:%.*]] = alloca ptr addrspace(5), align 4, addrspace(5)
+// OPENCL-NEXT:    store i32 [[N]], ptr addrspace(5) [[N_ADDR]], align 4
+// OPENCL-NEXT:    [[TMP0:%.*]] = load i32, ptr addrspace(5) [[N_ADDR]], align 4
+// OPENCL-NEXT:    [[CONV:%.*]] = zext i32 [[TMP0]] to i64
+// OPENCL-NEXT:    [[TMP1:%.*]] = alloca i8, i64 [[CONV]], align 1, addrspace(5)
+// OPENCL-NEXT:    store ptr addrspace(5) [[TMP1]], ptr addrspace(5) [[ALLOC_PTR_ALIGN_UNINITIALIZED]], align 4
+// OPENCL-NEXT:    ret void
+//
+void test2_builtin_alloca_with_align_uninitialized(unsigned n) {
+    __private void *alloc_ptr_align_uninitialized = __builtin_alloca_with_align_uninitialized(n, 8);
+}
diff --git a/clang/test/CodeGenOpenCL/relaxed-fpmath.cl b/clang/test/CodeGenOpenCL/relaxed-fpmath.cl
@@ -1,6 +1,6 @@
 // RUN: %clang_cc1 %s -emit-llvm -o - | FileCheck %s -check-prefix=NORMAL
 // RUN: %clang_cc1 %s -emit-llvm -cl-fast-relaxed-math -o - | FileCheck %s -check-prefix=FAST
-// RUN: %clang_cc1 %s -emit-llvm -cl-finite-math-only -o - | FileCheck %s -check-prefix=FINITE
+// RUN: %clang_cc1 %s -emit-llvm -menable-no-infs -menable-no-nans -cl-finite-math-only -o - | FileCheck %s -check-prefix=FINITE
 // RUN: %clang_cc1 %s -emit-llvm -cl-unsafe-math-optimizations -o - | FileCheck %s -check-prefix=UNSAFE
 // RUN: %clang_cc1 %s -emit-llvm -cl-mad-enable -o - | FileCheck %s -check-prefix=MAD
 // RUN: %clang_cc1 %s -emit-llvm -cl-no-signed-zeros -o - | FileCheck %s -check-prefix=NOSIGNED
@@ -9,7 +9,7 @@
 // RUN: %clang_cc1 %s -DGEN_PCH=1 -finclude-default-header -triple spir-unknown-unknown -emit-pch -o %t.pch
 // RUN: %clang_cc1 %s -include-pch %t.pch -fno-validate-pch -emit-llvm -o - | FileCheck %s -check-prefix=NORMAL
 // RUN: %clang_cc1 %s -include-pch %t.pch -fno-validate-pch -emit-llvm -cl-fast-relaxed-math -o - | FileCheck %s -check-prefix=FAST
-// RUN: %clang_cc1 %s -include-pch %t.pch -fno-validate-pch -emit-llvm -cl-finite-math-only -o - | FileCheck %s -check-prefix=FINITE
+// RUN: %clang_cc1 %s -include-pch %t.pch -fno-validate-pch -emit-llvm -menable-no-infs -menable-no-nans -cl-finite-math-only -o - | FileCheck %s -check-prefix=FINITE
 // RUN: %clang_cc1 %s -include-pch %t.pch -fno-validate-pch -emit-llvm -cl-unsafe-math-optimizations -o - | FileCheck %s -check-prefix=UNSAFE
 // RUN: %clang_cc1 %s -include-pch %t.pch -fno-validate-pch -emit-llvm -cl-mad-enable -o - | FileCheck %s -check-prefix=MAD
 // RUN: %clang_cc1 %s -include-pch %t.pch -fno-validate-pch -emit-llvm -cl-no-signed-zeros -o - | FileCheck %s -check-prefix=NOSIGNED

diff --git a/clang/test/Driver/amdgpu-toolchain.c b/clang/test/Driver/amdgpu-toolchain.c
@@ -28,3 +28,7 @@
 // RUN: %clang -### --target=amdgcn-amd-amdhsa -mcpu=gfx906 -nogpulib \
 // RUN:   -fuse-ld=ld %s 2>&1 | FileCheck -check-prefixes=LD %s
 // LD: ld.lld
+
+// RUN: %clang -### --target=amdgcn-amd-amdhsa -mcpu=gfx906 -nogpulib \
+// RUN:   -r %s 2>&1 | FileCheck -check-prefixes=RELO %s
+// RELO-NOT: -shared
diff --git a/clang/test/Driver/opencl.cl b/clang/test/Driver/opencl.cl
@@ -35,7 +35,7 @@
 // CHECK-OPT-DISABLE: "-cc1" {{.*}} "-cl-opt-disable"
 // CHECK-STRICT-ALIASING: "-cc1" {{.*}} "-cl-strict-aliasing"
 // CHECK-SINGLE-PRECISION-CONST: "-cc1" {{.*}} "-cl-single-precision-constant"
-// CHECK-FINITE-MATH-ONLY: "-cc1" {{.*}} "-cl-finite-math-only"
+// CHECK-FINITE-MATH-ONLY: "-cc1" {{.*}} "-menable-no-infs" "-menable-no-nans" "-cl-finite-math-only"
 // CHECK-KERNEL-ARG-INFO: "-cc1" {{.*}} "-cl-kernel-arg-info"
 // CHECK-UNSAFE-MATH-OPT: "-cc1" {{.*}} "-cl-unsafe-math-optimizations"
 // CHECK-FAST-RELAXED-MATH: "-cc1" {{.*}} "-cl-fast-relaxed-math"

diff --git a/clang/test/Driver/stack-size-section.c b/clang/test/Driver/stack-size-section.c
@@ -14,6 +14,7 @@
 
 // RUN: %clang -### --target=x86_64-linux-gnu -flto -fstack-size-section %s 2>&1 | FileCheck %s --check-prefix=LTO
 // RUN: %clang -### --target=x86_64-linux-gnu -flto -fstack-size-section -fno-stack-size-section %s 2>&1 | FileCheck %s --check-prefix=LTO-NO
+// RUN: %clang -### --target=x86_64-sie-ps5 -fstack-size-section %s 2>&1 | FileCheck %s --check-prefix=LTO
 
 // LTO: "-plugin-opt=-stack-size-section"
 // LTO-NO-NOT: "-plugin-opt=-stack-size-section"

diff --git a/clang/test/Headers/__clang_hip_cmath.hip b/clang/test/Headers/__clang_hip_cmath.hip
@@ -13,7 +13,8 @@
 // RUN:   -internal-isystem %S/../../lib/Headers/cuda_wrappers \
 // RUN:   -internal-isystem %S/Inputs/include \
 // RUN:   -triple amdgcn-amd-amdhsa -aux-triple x86_64-unknown-unknown \
-// RUN:   -target-cpu gfx906 -emit-llvm %s -fcuda-is-device -O1 -ffinite-math-only -o - \
+// RUN:   -target-cpu gfx906 -emit-llvm %s -fcuda-is-device -O1 -menable-no-infs \
+// RUN:   -menable-no-nans -o - \
 // RUN:   -D__HIPCC_RTC__ | FileCheck -check-prefix=FINITEONLY %s
 
 // DEFAULT-LABEL: @test_fma_f16(

diff --git a/clang/test/Headers/__clang_hip_math.hip b/clang/test/Headers/__clang_hip_math.hip
@@ -14,7 +14,8 @@
 // RUN:   -internal-isystem %S/../../lib/Headers/cuda_wrappers \
 // RUN:   -internal-isystem %S/Inputs/include \
 // RUN:   -triple amdgcn-amd-amdhsa -aux-triple x86_64-unknown-unknown \
-// RUN:   -target-cpu gfx906 -emit-llvm %s -fcuda-is-device -O1 -ffinite-math-only -o - \
+// RUN:   -target-cpu gfx906 -emit-llvm %s -fcuda-is-device -O1 -menable-no-infs \
+// RUN:   -menable-no-nans -o - \
 // RUN:   -D__HIPCC_RTC__ | FileCheck -check-prefixes=CHECK,FINITEONLY %s
 
 // Check that we end up with -fapprox-func set on intrinsic calls

diff --git a/clang/test/Headers/float.c b/clang/test/Headers/float.c
@@ -2,7 +2,7 @@
 // RUN: %clang_cc1 -fsyntax-only -verify -std=c99 -ffreestanding %s
 // RUN: %clang_cc1 -fsyntax-only -verify -std=c11 -ffreestanding %s
 // RUN: %clang_cc1 -fsyntax-only -verify -std=c23 -ffreestanding %s
-// RUN: %clang_cc1 -fsyntax-only -verify=finite -std=c23 -ffreestanding -ffinite-math-only %s
+// RUN: %clang_cc1 -fsyntax-only -verify=finite -std=c23 -ffreestanding -menable-no-nans -menable-no-infs %s
 // RUN: %clang_cc1 -fsyntax-only -verify -xc++ -std=c++11 -ffreestanding %s
 // RUN: %clang_cc1 -fsyntax-only -verify -xc++ -std=c++14 -ffreestanding %s
 // RUN: %clang_cc1 -fsyntax-only -verify -xc++ -std=c++17 -ffreestanding %s

diff --git a/clang/test/OpenMP/amdgpu-unsafe-fp-atomics.cpp b/clang/test/OpenMP/amdgpu-unsafe-fp-atomics.cpp
@@ -0,0 +1,59 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple amdgcn-amd-amdhsa -fopenmp-targets=amdgcn-amd-amdhsa -emit-llvm %s -fopenmp-is-target-device -o - | FileCheck -check-prefix=DEFAULT %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple amdgcn-amd-amdhsa -fopenmp-targets=amdgcn-amd-amdhsa -munsafe-fp-atomics -emit-llvm %s -fopenmp-is-target-device -o - | FileCheck -check-prefix=UNSAFE-FP-ATOMICS %s
+
+#pragma omp declare target
+
+float fv, fx;
+double dv, dx;
+
+// DEFAULT-LABEL: define hidden void @_Z15atomic_fadd_f32v(
+// DEFAULT-SAME: ) #[[ATTR0:[0-9]+]] {
+// DEFAULT-NEXT:  [[ENTRY:.*:]]
+// DEFAULT-NEXT:    [[TMP0:%.*]] = load float, ptr addrspacecast (ptr addrspace(1) @fv to ptr), align 4
+// DEFAULT-NEXT:    [[TMP1:%.*]] = atomicrmw fadd ptr addrspacecast (ptr addrspace(1) @fx to ptr), float [[TMP0]] monotonic, align 4
+// DEFAULT-NEXT:    [[ADD:%.*]] = fadd float [[TMP1]], [[TMP0]]
+// DEFAULT-NEXT:    store float [[ADD]], ptr addrspacecast (ptr addrspace(1) @fv to ptr), align 4
+// DEFAULT-NEXT:    ret void
+//
+// UNSAFE-FP-ATOMICS-LABEL: define hidden void @_Z15atomic_fadd_f32v(
+// UNSAFE-FP-ATOMICS-SAME: ) #[[ATTR0:[0-9]+]] {
+// UNSAFE-FP-ATOMICS-NEXT:  [[ENTRY:.*:]]
+// UNSAFE-FP-ATOMICS-NEXT:    [[TMP0:%.*]] = load float, ptr addrspacecast (ptr addrspace(1) @fv to ptr), align 4
+// UNSAFE-FP-ATOMICS-NEXT:    [[TMP1:%.*]] = atomicrmw fadd ptr addrspacecast (ptr addrspace(1) @fx to ptr), float [[TMP0]] monotonic, align 4, !amdgpu.no.fine.grained.memory [[META5:![0-9]+]], !amdgpu.ignore.denormal.mode [[META5]]
+// UNSAFE-FP-ATOMICS-NEXT:    [[ADD:%.*]] = fadd float [[TMP1]], [[TMP0]]
+// UNSAFE-FP-ATOMICS-NEXT:    store float [[ADD]], ptr addrspacecast (ptr addrspace(1) @fv to ptr), align 4
+// UNSAFE-FP-ATOMICS-NEXT:    ret void
+//
+void atomic_fadd_f32() {
+#pragma omp atomic capture
+  fv = fx = fx + fv;
+}
+
+// DEFAULT-LABEL: define hidden void @_Z15atomic_fadd_f64v(
+// DEFAULT-SAME: ) #[[ATTR0]] {
+// DEFAULT-NEXT:  [[ENTRY:.*:]]
+// DEFAULT-NEXT:    [[TMP0:%.*]] = load double, ptr addrspacecast (ptr addrspace(1) @dv to ptr), align 8
+// DEFAULT-NEXT:    [[TMP1:%.*]] = atomicrmw fadd ptr addrspacecast (ptr addrspace(1) @dx to ptr), double [[TMP0]] monotonic, align 8
+// DEFAULT-NEXT:    [[ADD:%.*]] = fadd double [[TMP1]], [[TMP0]]
+// DEFAULT-NEXT:    store double [[ADD]], ptr addrspacecast (ptr addrspace(1) @dv to ptr), align 8
+// DEFAULT-NEXT:    ret void
+//
+// UNSAFE-FP-ATOMICS-LABEL: define hidden void @_Z15atomic_fadd_f64v(
+// UNSAFE-FP-ATOMICS-SAME: ) #[[ATTR0]] {
+// UNSAFE-FP-ATOMICS-NEXT:  [[ENTRY:.*:]]
+// UNSAFE-FP-ATOMICS-NEXT:    [[TMP0:%.*]] = load double, ptr addrspacecast (ptr addrspace(1) @dv to ptr), align 8
+// UNSAFE-FP-ATOMICS-NEXT:    [[TMP1:%.*]] = atomicrmw fadd ptr addrspacecast (ptr addrspace(1) @dx to ptr), double [[TMP0]] monotonic, align 8, !amdgpu.no.fine.grained.memory [[META5]]
+// UNSAFE-FP-ATOMICS-NEXT:    [[ADD:%.*]] = fadd double [[TMP1]], [[TMP0]]
+// UNSAFE-FP-ATOMICS-NEXT:    store double [[ADD]], ptr addrspacecast (ptr addrspace(1) @dv to ptr), align 8
+// UNSAFE-FP-ATOMICS-NEXT:    ret void
+//
+void atomic_fadd_f64() {
+#pragma omp atomic capture
+  dv = dx = dx + dv;
+}
+
+#pragma omp end declare target
+//.
+// UNSAFE-FP-ATOMICS: [[META5]] = !{}
+//.
diff --git a/clang/test/Parser/namelookup-anonymous-struct.c b/clang/test/Parser/namelookup-anonymous-struct.c
@@ -0,0 +1,6 @@
+// RUN: %clang_cc1 -std=c11 -verify %s
+
+struct GH31295 {
+  struct { int x; };
+  int arr[sizeof(x)]; // expected-error{{use of undeclared identifier 'x'}}
+};
diff --git a/clang/test/Preprocessor/predefined-macros.c b/clang/test/Preprocessor/predefined-macros.c
@@ -70,7 +70,7 @@
 // RUN:   | FileCheck -match-full-lines %s --check-prefix=CHECK-NO-MATH-ERRNO
 // CHECK-NO-MATH-ERRNO: #define __NO_MATH_ERRNO__ 1
 //
-// RUN: %clang_cc1 %s -E -dM -ffinite-math-only -o - \
+// RUN: %clang_cc1 %s -E -dM -menable-no-nans -menable-no-infs -o - \
 // RUN:   | FileCheck -match-full-lines %s --check-prefix=CHECK-FINITE-MATH-ONLY
 // CHECK-FINITE-MATH-ONLY: #define __FINITE_MATH_ONLY__ 1
 //
@@ -316,4 +316,4 @@
 // RUN:  -triple amdgcn-amd-amdhsa -fcuda-is-device | FileCheck -match-full-lines \
 // RUN:  %s --check-prefix=CHECK-HIPSTDPAR-INTERPOSE-DEV-NEG
 // CHECK-HIPSTDPAR-INTERPOSE-DEV-NEG: #define __HIPSTDPAR__ 1
-// CHECK-HIPSTDPAR-INTERPOSE-DEV-NEG-NOT: #define __HIPSTDPAR_INTERPOSE_ALLOC__ 1
+// CHECK-HIPSTDPAR-INTERPOSE-DEV-NEG-NOT: #define __HIPSTDPAR_INTERPOSE_ALLOC__ 1
diff --git a/clang/test/Sema/aarch64-fmv-streaming.c b/clang/test/Sema/aarch64-fmv-streaming.c
@@ -0,0 +1,46 @@
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sme -Waarch64-sme-attributes -fsyntax-only -verify %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sme -Waarch64-sme-attributes -fsyntax-only -verify=expected-cpp -x c++ %s
+
+__attribute__((target_clones("sve", "simd"))) void ok_arm_streaming(void) __arm_streaming {}
+__arm_locally_streaming __attribute__((target_version("sme2"))) void ok_arm_streaming(void) __arm_streaming {}
+__attribute__((target_version("default"))) void ok_arm_streaming(void) __arm_streaming {}
+
+__attribute__((target_clones("sve", "simd"))) void ok_arm_streaming_compatible(void) __arm_streaming_compatible {}
+__arm_locally_streaming __attribute__((target_version("sme2"))) void ok_arm_streaming_compatible(void) __arm_streaming_compatible {}
+__attribute__((target_version("default"))) void ok_arm_streaming_compatible(void) __arm_streaming_compatible {}
+
+__arm_locally_streaming __attribute__((target_clones("sve", "simd"))) void ok_no_streaming(void) {}
+__attribute__((target_version("sme2"))) void ok_no_streaming(void) {}
+__attribute__((target_version("default"))) void ok_no_streaming(void) {}
+
+__attribute__((target_clones("sve", "simd"))) void bad_mixed_streaming(void) {}
+// expected-cpp-error@+2 {{multiversioned function declaration has a different calling convention}}
+// expected-error@+1 {{multiversioned function declaration has a different calling convention}}
+__attribute__((target_version("sme2"))) void bad_mixed_streaming(void) __arm_streaming {}
+// expected-cpp-error@+2 {{multiversioned function declaration has a different calling convention}}
+// expected-error@+1 {{multiversioned function declaration has a different calling convention}}
+__attribute__((target_version("default"))) void bad_mixed_streaming(void) __arm_streaming_compatible {}
+// expected-cpp-error@+2 {{multiversioned function declaration has a different calling convention}}
+// expected-error@+1 {{multiversioned function declaration has a different calling convention}}
+__arm_locally_streaming __attribute__((target_version("dotprod"))) void bad_mixed_streaming(void) __arm_streaming {}
+
+void n_caller(void) {
+  ok_arm_streaming();
+  ok_arm_streaming_compatible();
+  ok_no_streaming();
+  bad_mixed_streaming();
+}
+
+void s_caller(void) __arm_streaming {
+  ok_arm_streaming();
+  ok_arm_streaming_compatible();
+  ok_no_streaming();
+  bad_mixed_streaming();
+}
+
+void sc_caller(void) __arm_streaming_compatible {
+  ok_arm_streaming();
+  ok_arm_streaming_compatible();
+  ok_no_streaming();
+  bad_mixed_streaming();
+}
diff --git a/clang/test/Sema/aarch64-sme-func-attrs.c b/clang/test/Sema/aarch64-sme-func-attrs.c
@@ -455,48 +455,6 @@ void unimplemented_spill_fill_za(void (*share_zt0_only)(void) __arm_inout("zt0")
   share_zt0_only();
 }
 
-// expected-cpp-error@+2 {{streaming function cannot be multi-versioned}}
-// expected-error@+1 {{streaming function cannot be multi-versioned}}
-__attribute__((target_version("sme2")))
-void cannot_work_version(void) __arm_streaming {}
-// expected-cpp-error@+5 {{function declared 'void ()' was previously declared 'void () __arm_streaming', which has different SME function attributes}}
-// expected-cpp-note@-2 {{previous declaration is here}}
-// expected-error@+3 {{function declared 'void (void)' was previously declared 'void (void) __arm_streaming', which has different SME function attributes}}
-// expected-note@-4 {{previous declaration is here}}
-__attribute__((target_version("default")))
-void cannot_work_version(void) {}
-
-
-// expected-cpp-error@+2 {{streaming function cannot be multi-versioned}}
-// expected-error@+1 {{streaming function cannot be multi-versioned}}
-__attribute__((target_clones("sme2")))
-void cannot_work_clones(void) __arm_streaming {}
-
-
-__attribute__((target("sme2")))
-void just_fine_streaming(void) __arm_streaming {}
-__attribute__((target_version("sme2")))
-void just_fine(void) { just_fine_streaming(); }
-__attribute__((target_version("default")))
-void just_fine(void) {}
-
-
-__arm_locally_streaming
-__attribute__((target_version("sme2")))
-void incompatible_locally_streaming(void) {}
-// expected-error@-1 {{attribute 'target_version' multiversioning cannot be combined with attribute '__arm_locally_streaming'}}
-// expected-cpp-error@-2 {{attribute 'target_version' multiversioning cannot be combined with attribute '__arm_locally_streaming'}}
-__attribute__((target_version("default")))
-void incompatible_locally_streaming(void) {}
-
-
-void fmv_caller() {
-    cannot_work_version();
-    cannot_work_clones();
-    just_fine();
-    incompatible_locally_streaming();
-}
-
 void sme_streaming_with_vl_arg(__SVInt8_t a) __arm_streaming { }
 
 __SVInt8_t sme_streaming_returns_vl(void) __arm_streaming { __SVInt8_t r; return r; }

diff --git a/clang/test/Sema/warn-infinity-nan-disabled-lnx.cpp b/clang/test/Sema/warn-infinity-nan-disabled-lnx.cpp
@@ -1,10 +1,11 @@
 // RUN: %clang_cc1 -x c++ -verify=no-inf-no-nan \
-// RUN: -triple powerpc64le-unknown-unknown %s -menable-no-infs \
-// RUN: -menable-no-nans -std=c++23
+// RUN: -triple powerpc64le-unknown-unknown %s \
+// RUN: -menable-no-infs -menable-no-nans -std=c++23
 
 // RUN: %clang_cc1 -x c++ -verify=no-inf-no-nan \
-// RUN: -triple powerpc64le-unknown-unknown %s -menable-no-infs \
-// RUN: -menable-no-nans -funsafe-math-optimizations -std=c++23
+// RUN: -triple powerpc64le-unknown-unknown %s \
+// RUN: -menable-no-infs -menable-no-nans -funsafe-math-optimizations \
+// RUN: -std=c++23
 
 // RUN: %clang_cc1 -x c++ -verify=no-fast -triple powerpc64le-unknown-unknown \
 // RUN: %s -std=c++23

diff --git a/clang/test/Sema/warn-infinity-nan-disabled-win.cpp b/clang/test/Sema/warn-infinity-nan-disabled-win.cpp
@@ -6,8 +6,9 @@
 // RUN: -menable-no-nans -std=c++23
 
 // RUN: %clang_cc1 -x c++ -verify=no-inf-no-nan \
-// RUN: -triple powerpc64le-unknown-unknown %s -menable-no-infs \
-// RUN: -menable-no-nans -funsafe-math-optimizations -std=c++23
+// RUN: -triple powerpc64le-unknown-unknown %s \
+// RUN: -menable-no-infs -menable-no-nans -funsafe-math-optimizations \
+// RUN: -std=c++23
 
 // RUN: %clang_cc1 -x c++ -verify=no-fast -triple powerpc64le-unknown-unknown \
 // RUN: %s -std=c++23

diff --git a/clang/test/SemaCXX/pr100095.cpp b/clang/test/SemaCXX/pr100095.cpp
@@ -0,0 +1,17 @@
+// RUN: %clang_cc1 -fsyntax-only -std=c++11 %s
+// XFAIL: asserts
+
+template <class> struct Pair;
+template <class...> struct Tuple {
+  template <class _Up> Tuple(_Up);
+};
+template <typename> struct StatusOr;
+template <int> using ElementType = int;
+template <int... fields>
+using Key = Tuple<ElementType<fields>...>;
+template <int... fields>
+StatusOr<Pair<Key<fields...>>> Parser();
+struct Helper { Helper(Tuple<>, Tuple<>, int, int); };
+struct D : Helper {
+  D(Key<> f, int n, int e) : Helper(f, Parser<>, n, e) {}
+};
diff --git a/clang/test/SemaCXX/type-traits.cpp b/clang/test/SemaCXX/type-traits.cpp
@@ -2402,11 +2402,11 @@ template<typename T> struct DerivedB : BaseA<T> { };
 template<typename T> struct CrazyDerived : T { };
 
 
-class class_forward; // expected-note 2 {{forward declaration of 'class_forward'}}
+class class_forward; // expected-note 4 {{forward declaration of 'class_forward'}}
 
 template <class T> class DerivedTemp : Base {};
 template <class T> class NonderivedTemp {};
-template <class T> class UndefinedTemp; // expected-note {{declared here}}
+template <class T> class UndefinedTemp; // expected-note 2 {{declared here}}
 
 void is_base_of() {
   static_assert(__is_base_of(Base, Derived));
@@ -2457,6 +2457,76 @@ void is_base_of() {
   static_assert(!__is_base_of(DerivedB<int>, BaseA<int>));
 }
 
+struct DerivedTransitiveViaNonVirtual : Derived3 {};
+struct DerivedTransitiveViaVirtual : virtual Derived3 {};
+
+template <typename T>
+struct CrazyDerivedVirtual : virtual T {};
+
+struct DerivedPrivate : private virtual Base {};
+struct DerivedProtected : protected virtual Base {};
+struct DerivedPrivatePrivate : private DerivedPrivate {};
+struct DerivedPrivateProtected : private DerivedProtected {};
+struct DerivedProtectedPrivate : protected DerivedProtected {};
+struct DerivedProtectedProtected : protected DerivedProtected {};
+
+void is_virtual_base_of(int n) {
+  static_assert(!__builtin_is_virtual_base_of(Base, Derived));
+  static_assert(!__builtin_is_virtual_base_of(const Base, Derived));
+  static_assert(!__builtin_is_virtual_base_of(Derived, Base));
+  static_assert(!__builtin_is_virtual_base_of(Derived, int));
+  static_assert(!__builtin_is_virtual_base_of(Base, Base));
+  static_assert(!__builtin_is_virtual_base_of(Base, Derived3));
+  static_assert(!__builtin_is_virtual_base_of(Derived, Derived3));
+  static_assert(__builtin_is_virtual_base_of(Derived2b, Derived3));
+  static_assert(__builtin_is_virtual_base_of(Derived2a, Derived3));
+  static_assert(!__builtin_is_virtual_base_of(BaseA<int>, DerivedB<int>));
+  static_assert(!__builtin_is_virtual_base_of(DerivedB<int>, BaseA<int>));
+  static_assert(!__builtin_is_virtual_base_of(Union, Union));
+  static_assert(!__builtin_is_virtual_base_of(Empty, Empty));
+  static_assert(!__builtin_is_virtual_base_of(class_forward, class_forward)); // expected-error {{incomplete type 'class_forward' where a complete type is required}}
+  static_assert(!__builtin_is_virtual_base_of(Empty, class_forward)); // expected-error {{incomplete type 'class_forward' where a complete type is required}}
+  static_assert(!__builtin_is_virtual_base_of(class_forward, Empty));
+  static_assert(!__builtin_is_virtual_base_of(Base&, Derived&));
+  static_assert(!__builtin_is_virtual_base_of(Base[10], Derived[10]));
+  static_assert(!__builtin_is_virtual_base_of(Base[n], Derived[n])); // expected-error 2 {{variable length arrays are not supported in '__builtin_is_virtual_base_of'}}
+  static_assert(!__builtin_is_virtual_base_of(int, int));
+  static_assert(!__builtin_is_virtual_base_of(int[], int[]));
+  static_assert(!__builtin_is_virtual_base_of(long, int));
+  static_assert(!__builtin_is_virtual_base_of(Base, DerivedTemp<int>));
+  static_assert(!__builtin_is_virtual_base_of(Base, NonderivedTemp<int>));
+  static_assert(!__builtin_is_virtual_base_of(Base, UndefinedTemp<int>)); // expected-error {{implicit instantiation of undefined template 'UndefinedTemp<int>'}}
+  static_assert(__builtin_is_virtual_base_of(Base, DerivedPrivate));
+  static_assert(__builtin_is_virtual_base_of(Base, DerivedProtected));
+  static_assert(__builtin_is_virtual_base_of(Base, DerivedPrivatePrivate));
+  static_assert(__builtin_is_virtual_base_of(Base, DerivedPrivateProtected));
+  static_assert(__builtin_is_virtual_base_of(Base, DerivedProtectedPrivate));
+  static_assert(__builtin_is_virtual_base_of(Base, DerivedProtectedProtected));
+  static_assert(__builtin_is_virtual_base_of(Derived2a, DerivedTransitiveViaNonVirtual));
+  static_assert(__builtin_is_virtual_base_of(Derived2b, DerivedTransitiveViaNonVirtual));
+  static_assert(__builtin_is_virtual_base_of(Derived2a, DerivedTransitiveViaVirtual));
+  static_assert(__builtin_is_virtual_base_of(Derived2b, DerivedTransitiveViaVirtual));
+  static_assert(!__builtin_is_virtual_base_of(Base, CrazyDerived<Base>));
+  static_assert(!__builtin_is_virtual_base_of(CrazyDerived<Base>, Base));
+  static_assert(__builtin_is_virtual_base_of(Base, CrazyDerivedVirtual<Base>));
+  static_assert(!__builtin_is_virtual_base_of(CrazyDerivedVirtual<Base>, Base));
+
+  static_assert(!__builtin_is_virtual_base_of(IncompleteUnion, IncompleteUnion));
+  static_assert(!__builtin_is_virtual_base_of(Union, IncompleteUnion));
+  static_assert(!__builtin_is_virtual_base_of(IncompleteUnion, Union));
+  static_assert(!__builtin_is_virtual_base_of(IncompleteStruct, IncompleteUnion));
+  static_assert(!__builtin_is_virtual_base_of(IncompleteUnion, IncompleteStruct));
+  static_assert(!__builtin_is_virtual_base_of(Empty, IncompleteUnion));
+  static_assert(!__builtin_is_virtual_base_of(IncompleteUnion, Empty));
+  static_assert(!__builtin_is_virtual_base_of(int, IncompleteUnion));
+  static_assert(!__builtin_is_virtual_base_of(IncompleteUnion, int));
+  static_assert(!__builtin_is_virtual_base_of(Empty, Union));
+  static_assert(!__builtin_is_virtual_base_of(Union, Empty));
+  static_assert(!__builtin_is_virtual_base_of(int, Empty));
+  static_assert(!__builtin_is_virtual_base_of(Union, int));
+  static_assert(!__builtin_is_virtual_base_of(IncompleteStruct, IncompleteStruct[n])); // expected-error {{variable length arrays are not supported in '__builtin_is_virtual_base_of'}}
+}
+
 template<class T, class U>
 class TemplateClass {};
 

diff --git a/clang/tools/clang-nvlink-wrapper/ClangNVLinkWrapper.cpp b/clang/tools/clang-nvlink-wrapper/ClangNVLinkWrapper.cpp
@@ -598,10 +598,11 @@ Expected<SmallVector<StringRef>> getInput(const ArgList &Args) {
         Res.Prevailing = !Sym.isUndefined() && ObjSym.File == *BitcodeFile;
 
         // We need LTO to preseve the following global symbols:
-        // 1) Symbols used in regular objects.
-        // 2) Prevailing symbols that are needed visible to the gpu runtime.
+        // 1) All symbols during a relocatable link.
+        // 2) Symbols used in regular objects.
+        // 3) Prevailing symbols that are needed visible to the gpu runtime.
         Res.VisibleToRegularObj =
-            ObjSym.UsedInRegularObj ||
+            Args.hasArg(OPT_relocatable) || ObjSym.UsedInRegularObj ||
             (Res.Prevailing &&
              (Sym.getVisibility() != GlobalValue::HiddenVisibility &&
               !Sym.canBeOmittedFromSymbolTable()));

diff --git a/clang/tools/clang-nvlink-wrapper/NVLinkOpts.td b/clang/tools/clang-nvlink-wrapper/NVLinkOpts.td
@@ -73,6 +73,10 @@ def plugin_opt : Joined<["--", "-"], "plugin-opt=">, Flags<[WrapperOnlyOption]>,
 def save_temps : Flag<["--", "-"], "save-temps">,
   Flags<[WrapperOnlyOption]>, HelpText<"Save intermediate results">;
 
+def relocatable : Flag<["--", "-"], "relocatable">,
+  Flags<[WrapperOnlyOption]>, HelpText<"Perform a relocatable link (LTO only)">;
+def r : Flag<["-"], "r">, Flags<[WrapperOnlyOption]>, Alias<relocatable>;
+
 def whole_archive : Flag<["--", "-"], "whole-archive">,
   Flags<[WrapperOnlyOption, HelpHidden]>;
 def no_whole_archive : Flag<["--", "-"], "no-whole-archive">,
@@ -83,8 +87,7 @@ def mllvm : Separate<["-"], "mllvm">, Flags<[WrapperOnlyOption]>,
   HelpText<"Arguments passed to LLVM, including Clang invocations, for which "
            "the '-mllvm' prefix is preserved. Use '-mllvm --help' for a list "
            "of options.">;
-def mllvm_EQ : Joined<["-"], "mllvm=">, Flags<[HelpHidden]>,
-  Alias<mllvm>;
+def mllvm_EQ : Joined<["-"], "mllvm=">, Flags<[HelpHidden]>, Alias<mllvm>;
 
 def dry_run : Flag<["--", "-"], "dry-run">, Flags<[WrapperOnlyOption]>,
   HelpText<"Print generated commands without running.">;
diff --git a/clang/unittests/AST/ASTImporterTest.cpp b/clang/unittests/AST/ASTImporterTest.cpp
@@ -9783,6 +9783,43 @@ TEST_P(ASTImporterOptionSpecificTestBase, ImportExistingEmptyAnonymousEnums) {
   EXPECT_EQ(ImportedE2, ToE1);
 }
 
+TEST_P(ASTImporterOptionSpecificTestBase, ImportMultipleAnonymousEnumDecls) {
+  Decl *ToTU = getToTuDecl("", Lang_CXX03);
+  Decl *FromTU = getTuDecl(
+      R"(
+        struct foo {
+          enum { A };
+          enum { B };
+        };
+      )",
+      Lang_CXX03);
+
+  auto EnumConstA = enumConstantDecl(hasName("A"));
+  auto EnumConstB = enumConstantDecl(hasName("B"));
+
+  auto *FromA = FirstDeclMatcher<EnumConstantDecl>().match(FromTU, EnumConstA);
+  auto *FromB = FirstDeclMatcher<EnumConstantDecl>().match(FromTU, EnumConstB);
+
+  auto *ToA = Import(FromA, Lang_CXX03);
+  auto *ToB = Import(FromB, Lang_CXX03);
+
+  ASSERT_TRUE(ToA);
+  ASSERT_TRUE(ToB);
+
+  auto *ToFooA = FirstDeclMatcher<CXXRecordDecl>().match(
+      ToTU, tagDecl(has(enumDecl(has(EnumConstA)))));
+  auto *ToFooB = FirstDeclMatcher<CXXRecordDecl>().match(
+      ToTU, tagDecl(has(enumDecl(has(EnumConstB)))));
+  ASSERT_EQ(ToFooA, ToFooB);
+
+  // different EnumDecl
+  auto *ToEnumDeclA =
+      FirstDeclMatcher<EnumDecl>().match(ToTU, enumDecl(has(EnumConstA)));
+  auto *ToEnumDeclB =
+      FirstDeclMatcher<EnumDecl>().match(ToTU, enumDecl(has(EnumConstB)));
+  ASSERT_NE(ToEnumDeclA, ToEnumDeclB);
+}
+
 INSTANTIATE_TEST_SUITE_P(ParameterizedTests, ASTImporterLookupTableTest,
                          DefaultTestValuesForRunOptions);
 

diff --git a/clang/unittests/AST/StructuralEquivalenceTest.cpp b/clang/unittests/AST/StructuralEquivalenceTest.cpp
@@ -1109,6 +1109,20 @@ TEST_F(StructuralEquivalenceEnumTest, EnumsWithDifferentBody) {
   EXPECT_FALSE(testStructuralMatch(t));
 }
 
+TEST_F(StructuralEquivalenceEnumTest, AnonymousEnumsWithSameConsts) {
+  // field x is required to trigger comparison of the anonymous enum
+  auto t = makeNamedDecls("struct foo { enum { A } x; };",
+                          "struct foo { enum { A } x;};", Lang_CXX11);
+  EXPECT_TRUE(testStructuralMatch(t));
+}
+
+TEST_F(StructuralEquivalenceEnumTest, AnonymousEnumsWithDiffConsts) {
+  // field x is required to trigger comparison of the anonymous enum
+  auto t = makeNamedDecls("struct foo { enum { A } x; };",
+                          "struct foo { enum { B } x;};", Lang_CXX11);
+  EXPECT_FALSE(testStructuralMatch(t));
+}
+
 struct StructuralEquivalenceEnumConstantTest : StructuralEquivalenceTest {};
 
 TEST_F(StructuralEquivalenceEnumConstantTest, EnumConstantsWithSameValues) {

diff --git a/clang/unittests/Analysis/FlowSensitive/DataflowEnvironmentTest.cpp b/clang/unittests/Analysis/FlowSensitive/DataflowEnvironmentTest.cpp
@@ -442,6 +442,55 @@ TEST_F(EnvironmentTest, CXXDefaultInitExprResultObjIsWrappedExprResultObj) {
             &Env.getResultObjectLocation(*DefaultInit->getExpr()));
 }
 
+// This test verifies the behavior of `getResultObjectLocation()` in
+// scenarios involving inherited constructors.
+// Since the specific AST node of interest `CXXConstructorDecl` is implicitly
+// generated, we cannot annotate any statements inside of it as we do in tests
+// within TransferTest. Thus, the only way to get the right `Environment` is by
+// explicitly initializing it as we do in tests within EnvironmentTest.
+// This is why this test is not inside TransferTest, where most of the tests for
+// `getResultObjectLocation()` are located.
+TEST_F(EnvironmentTest, ResultObjectLocationForInheritedCtorInitExpr) {
+  using namespace ast_matchers;
+
+  std::string Code = R"(
+    struct Base {
+      Base(int b) {}
+    };
+    struct Derived : Base {
+      using Base::Base;
+    };
+
+    Derived d = Derived(0);
+  )";
+
+  auto Unit =
+      tooling::buildASTFromCodeWithArgs(Code, {"-fsyntax-only", "-std=c++20"});
+  auto &Context = Unit->getASTContext();
+
+  ASSERT_EQ(Context.getDiagnostics().getClient()->getNumErrors(), 0U);
+
+  auto Results =
+      match(cxxConstructorDecl(
+                hasAnyConstructorInitializer(cxxCtorInitializer(
+                    withInitializer(expr().bind("inherited_ctor_init_expr")))))
+                .bind("ctor"),
+            Context);
+  const auto *Constructor = selectFirst<CXXConstructorDecl>("ctor", Results);
+  const auto *InheritedCtorInit = selectFirst<CXXInheritedCtorInitExpr>(
+      "inherited_ctor_init_expr", Results);
+
+  EXPECT_EQ(InheritedCtorInit->child_begin(), InheritedCtorInit->child_end());
+
+  Environment Env(DAContext, *Constructor);
+  Env.initialize();
+
+  RecordStorageLocation &Loc = Env.getResultObjectLocation(*InheritedCtorInit);
+  EXPECT_NE(&Loc, nullptr);
+
+  EXPECT_EQ(&Loc, Env.getThisPointeeStorageLocation());
+}
+
 TEST_F(EnvironmentTest, Stmt) {
   using namespace ast_matchers;
 

diff --git a/flang/lib/Frontend/CompilerInvocation.cpp b/flang/lib/Frontend/CompilerInvocation.cpp
@@ -1072,7 +1072,7 @@ static bool parseFloatingPointArgs(CompilerInvocation &invoc,
     opts.setFPContractMode(fpContractMode);
   }
 
-  if (args.getLastArg(clang::driver::options::OPT_menable_no_infinities)) {
+  if (args.getLastArg(clang::driver::options::OPT_menable_no_infs)) {
     opts.NoHonorInfs = true;
   }
 

diff --git a/flang/runtime/transformational.cpp b/flang/runtime/transformational.cpp
@@ -508,7 +508,8 @@ void RTDEF(CshiftVector)(Descriptor &result, const Descriptor &source,
   SubscriptValue lb{sourceDim.LowerBound()};
   for (SubscriptValue j{0}; j < extent; ++j) {
     SubscriptValue resultAt{1 + j};
-    SubscriptValue sourceAt{lb + (j + shift) % extent};
+    SubscriptValue sourceAt{
+        lb + static_cast<SubscriptValue>(j + shift) % extent};
     if (sourceAt < lb) {
       sourceAt += extent;
     }
@@ -619,7 +620,7 @@ void RTDEF(EoshiftVector)(Descriptor &result, const Descriptor &source,
   }
   SubscriptValue lb{source.GetDimension(0).LowerBound()};
   for (SubscriptValue j{1}; j <= extent; ++j) {
-    SubscriptValue sourceAt{lb + j - 1 + shift};
+    SubscriptValue sourceAt{lb + j - 1 + static_cast<SubscriptValue>(shift)};
     if (sourceAt >= lb && sourceAt < lb + extent) {
       CopyElement(result, &j, source, &sourceAt, terminator);
     } else if (boundary) {

diff --git a/libc/config/darwin/arm/entrypoints.txt b/libc/config/darwin/arm/entrypoints.txt
@@ -120,6 +120,7 @@ set(TARGET_LIBM_ENTRYPOINTS
     libc.src.math.acoshf
     libc.src.math.asinf
     libc.src.math.asinhf
+    libc.src.math.atan2
     libc.src.math.atan2f
     libc.src.math.atanf
     libc.src.math.atanhf

diff --git a/libc/config/linux/aarch64/entrypoints.txt b/libc/config/linux/aarch64/entrypoints.txt
@@ -343,6 +343,7 @@ set(TARGET_LIBM_ENTRYPOINTS
     libc.src.math.acoshf
     libc.src.math.asinf
     libc.src.math.asinhf
+    libc.src.math.atan2
     libc.src.math.atan2f
     libc.src.math.atanf
     libc.src.math.atanhf

diff --git a/libc/config/linux/arm/entrypoints.txt b/libc/config/linux/arm/entrypoints.txt
@@ -213,6 +213,7 @@ set(TARGET_LIBM_ENTRYPOINTS
     libc.src.math.acoshf
     libc.src.math.asinf
     libc.src.math.asinhf
+    libc.src.math.atan2
     libc.src.math.atan2f
     libc.src.math.atanf
     libc.src.math.atanhf

diff --git a/libc/config/linux/riscv/entrypoints.txt b/libc/config/linux/riscv/entrypoints.txt
@@ -365,6 +365,7 @@ set(TARGET_LIBM_ENTRYPOINTS
     libc.src.math.acoshf
     libc.src.math.asinf
     libc.src.math.asinhf
+    libc.src.math.atan2
     libc.src.math.atan2f
     libc.src.math.atanf
     libc.src.math.atanhf

diff --git a/libc/config/linux/x86_64/entrypoints.txt b/libc/config/linux/x86_64/entrypoints.txt
@@ -365,6 +365,7 @@ set(TARGET_LIBM_ENTRYPOINTS
     libc.src.math.acoshf
     libc.src.math.asinf
     libc.src.math.asinhf
+    libc.src.math.atan2
     libc.src.math.atan2f
     libc.src.math.atanf
     libc.src.math.atanhf

diff --git a/libc/config/windows/entrypoints.txt b/libc/config/windows/entrypoints.txt
@@ -118,6 +118,7 @@ set(TARGET_LIBM_ENTRYPOINTS
     libc.src.math.acoshf
     libc.src.math.asinf
     libc.src.math.asinhf
+    libc.src.math.atan2
     libc.src.math.atan2f
     libc.src.math.atanf
     libc.src.math.atanhf

diff --git a/libc/docs/math/index.rst b/libc/docs/math/index.rst
@@ -260,7 +260,7 @@ Higher Math Functions
 +-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
 | atan      | |check|          |                 |                        |                      |                        | 7.12.4.3               | F.10.1.3                   |
 +-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
-| atan2     | |check|          |                 |                        |                      |                        | 7.12.4.4               | F.10.1.4                   |
+| atan2     | |check|          | 1 ULP           |                        |                      |                        | 7.12.4.4               | F.10.1.4                   |
 +-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+
 | atan2pi   |                  |                 |                        |                      |                        | 7.12.4.11              | F.10.1.11                  |
 +-----------+------------------+-----------------+------------------------+----------------------+------------------------+------------------------+----------------------------+

diff --git a/libc/spec/stdc.td b/libc/spec/stdc.td
@@ -692,6 +692,7 @@ def StdC : StandardSpec<"stdc"> {
 
           FunctionSpec<"atanf", RetValSpec<FloatType>, [ArgSpec<FloatType>]>,
 
+          FunctionSpec<"atan2", RetValSpec<DoubleType>, [ArgSpec<DoubleType>, ArgSpec<DoubleType>]>,
           FunctionSpec<"atan2f", RetValSpec<FloatType>, [ArgSpec<FloatType>, ArgSpec<FloatType>]>,
 
           FunctionSpec<"acoshf", RetValSpec<FloatType>, [ArgSpec<FloatType>]>,

diff --git a/libc/src/math/generic/CMakeLists.txt b/libc/src/math/generic/CMakeLists.txt
@@ -3848,6 +3848,26 @@ add_entrypoint_object(
     libc.src.__support.macros.optimization
 )
 
+add_entrypoint_object(
+  atan2
+  SRCS
+    atan2.cpp
+  HDRS
+    ../atan2.h
+  COMPILE_OPTIONS
+    -O3
+  DEPENDS
+    .inv_trigf_utils
+    libc.src.__support.FPUtil.double_double
+    libc.src.__support.FPUtil.dyadic_float
+    libc.src.__support.FPUtil.fp_bits
+    libc.src.__support.FPUtil.multiply_add
+    libc.src.__support.FPUtil.nearest_integer
+    libc.src.__support.FPUtil.polyeval
+    libc.src.__support.FPUtil.rounding_mode
+    libc.src.__support.macros.optimization
+)
+
 add_entrypoint_object(
   scalblnf16
   SRCS

diff --git a/libc/src/math/generic/atan2.cpp b/libc/src/math/generic/atan2.cpp
@@ -0,0 +1,313 @@
+//===-- Double-precision atan2 function -----------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "src/math/atan2.h"
+#include "inv_trigf_utils.h"
+#include "src/__support/FPUtil/FPBits.h"
+#include "src/__support/FPUtil/PolyEval.h"
+#include "src/__support/FPUtil/double_double.h"
+#include "src/__support/FPUtil/multiply_add.h"
+#include "src/__support/FPUtil/nearest_integer.h"
+#include "src/__support/FPUtil/rounding_mode.h"
+#include "src/__support/macros/config.h"
+#include "src/__support/macros/optimization.h" // LIBC_UNLIKELY
+
+namespace LIBC_NAMESPACE_DECL {
+
+namespace {
+
+using DoubleDouble = fputil::DoubleDouble;
+
+// atan(i/64) with i = 0..64, generated by Sollya with:
+// > for i from 0 to 64 do {
+//     a = round(atan(i/64), D, RN);
+//     b = round(atan(i/64) - a, D, RN);
+//     print("{", b, ",", a, "},");
+//   };
+constexpr fputil::DoubleDouble ATAN_I[65] = {
+    {0.0, 0.0},
+    {-0x1.220c39d4dff5p-61, 0x1.fff555bbb729bp-7},
+    {-0x1.5ec431444912cp-60, 0x1.ffd55bba97625p-6},
+    {-0x1.86ef8f794f105p-63, 0x1.7fb818430da2ap-5},
+    {-0x1.c934d86d23f1dp-60, 0x1.ff55bb72cfdeap-5},
+    {0x1.ac4ce285df847p-58, 0x1.3f59f0e7c559dp-4},
+    {-0x1.cfb654c0c3d98p-58, 0x1.7ee182602f10fp-4},
+    {0x1.f7b8f29a05987p-58, 0x1.be39ebe6f07c3p-4},
+    {-0x1.cd37686760c17p-59, 0x1.fd5ba9aac2f6ep-4},
+    {-0x1.b485914dacf8cp-59, 0x1.1e1fafb043727p-3},
+    {0x1.61a3b0ce9281bp-57, 0x1.3d6eee8c6626cp-3},
+    {-0x1.054ab2c010f3dp-58, 0x1.5c9811e3ec26ap-3},
+    {0x1.347b0b4f881cap-58, 0x1.7b97b4bce5b02p-3},
+    {0x1.cf601e7b4348ep-59, 0x1.9a6a8e96c8626p-3},
+    {0x1.17b10d2e0e5abp-61, 0x1.b90d7529260a2p-3},
+    {0x1.c648d1534597ep-57, 0x1.d77d5df205736p-3},
+    {0x1.8ab6e3cf7afbdp-57, 0x1.f5b75f92c80ddp-3},
+    {0x1.62e47390cb865p-56, 0x1.09dc597d86362p-2},
+    {0x1.30ca4748b1bf9p-57, 0x1.18bf5a30bf178p-2},
+    {-0x1.077cdd36dfc81p-56, 0x1.278372057ef46p-2},
+    {-0x1.963a544b672d8p-57, 0x1.362773707ebccp-2},
+    {-0x1.5d5e43c55b3bap-56, 0x1.44aa436c2af0ap-2},
+    {-0x1.2566480884082p-57, 0x1.530ad9951cd4ap-2},
+    {-0x1.a725715711fp-56, 0x1.614840309cfe2p-2},
+    {-0x1.c63aae6f6e918p-56, 0x1.6f61941e4def1p-2},
+    {0x1.69c885c2b249ap-56, 0x1.7d5604b63b3f7p-2},
+    {0x1.b6d0ba3748fa8p-56, 0x1.8b24d394a1b25p-2},
+    {0x1.9e6c988fd0a77p-56, 0x1.98cd5454d6b18p-2},
+    {-0x1.24dec1b50b7ffp-56, 0x1.a64eec3cc23fdp-2},
+    {0x1.ae187b1ca504p-56, 0x1.b3a911da65c6cp-2},
+    {-0x1.cc1ce70934c34p-56, 0x1.c0db4c94ec9fp-2},
+    {-0x1.a2cfa4418f1adp-56, 0x1.cde53432c1351p-2},
+    {0x1.a2b7f222f65e2p-56, 0x1.dac670561bb4fp-2},
+    {0x1.0e53dc1bf3435p-56, 0x1.e77eb7f175a34p-2},
+    {-0x1.a3992dc382a23p-57, 0x1.f40dd0b541418p-2},
+    {-0x1.b32c949c9d593p-55, 0x1.0039c73c1a40cp-1},
+    {-0x1.d5b495f6349e6p-56, 0x1.0657e94db30dp-1},
+    {0x1.974fa13b5404fp-58, 0x1.0c6145b5b43dap-1},
+    {-0x1.2bdaee1c0ee35p-58, 0x1.1255d9bfbd2a9p-1},
+    {0x1.c621cec00c301p-55, 0x1.1835a88be7c13p-1},
+    {-0x1.928df287a668fp-58, 0x1.1e00babdefeb4p-1},
+    {0x1.c421c9f38224ep-57, 0x1.23b71e2cc9e6ap-1},
+    {-0x1.09e73b0c6c087p-56, 0x1.2958e59308e31p-1},
+    {0x1.c5d5e9ff0cf8dp-55, 0x1.2ee628406cbcap-1},
+    {0x1.1021137c71102p-55, 0x1.345f01cce37bbp-1},
+    {-0x1.2304331d8bf46p-55, 0x1.39c391cd4171ap-1},
+    {0x1.ecf8b492644fp-56, 0x1.3f13fb89e96f4p-1},
+    {-0x1.f76d0163f79c8p-56, 0x1.445065b795b56p-1},
+    {0x1.2419a87f2a458p-56, 0x1.4978fa3269ee1p-1},
+    {0x1.4a33dbeb3796cp-55, 0x1.4e8de5bb6ec04p-1},
+    {-0x1.1bb74abda520cp-55, 0x1.538f57b89061fp-1},
+    {-0x1.5e5c9d8c5a95p-56, 0x1.587d81f732fbbp-1},
+    {0x1.0028e4bc5e7cap-57, 0x1.5d58987169b18p-1},
+    {-0x1.2b785350ee8c1p-57, 0x1.6220d115d7b8ep-1},
+    {-0x1.6ea6febe8bbbap-56, 0x1.66d663923e087p-1},
+    {-0x1.a80386188c50ep-55, 0x1.6b798920b3d99p-1},
+    {-0x1.8c34d25aadef6p-56, 0x1.700a7c5784634p-1},
+    {0x1.7b2a6165884a1p-59, 0x1.748978fba8e0fp-1},
+    {0x1.406a08980374p-55, 0x1.78f6bbd5d315ep-1},
+    {0x1.560821e2f3aa9p-55, 0x1.7d528289fa093p-1},
+    {-0x1.bf76229d3b917p-56, 0x1.819d0b7158a4dp-1},
+    {0x1.6b66e7fc8b8c3p-57, 0x1.85d69576cc2c5p-1},
+    {-0x1.55b9a5e177a1bp-55, 0x1.89ff5ff57f1f8p-1},
+    {-0x1.ec182ab042f61p-56, 0x1.8e17aa99cc05ep-1},
+    {0x1.1a62633145c07p-55, 0x1.921fb54442d18p-1},
+};
+
+// Approximate atan(x) for |x| <= 2^-7.
+// Using degree-9 Taylor polynomial:
+//  P = x - x^3/3 + x^5/5 -x^7/7 + x^9/9;
+// Then the absolute error is bounded by:
+//   |atan(x) - P(x)| < |x|^11/11 < 2^(-7*11) / 11 < 2^-80.
+// And the relative error is bounded by:
+//   |(atan(x) - P(x))/atan(x)| < |x|^10 / 10 < 2^-73.
+// For x = x_hi + x_lo, fully expand the polynomial and drop any terms less than
+//   ulp(x_hi^3 / 3) gives us:
+// P(x) ~ x_hi - x_hi^3/3 + x_hi^5/5 - x_hi^7/7 + x_hi^9/9 +
+//        + x_lo * (1 - x_hi^2 + x_hi^4)
+DoubleDouble atan_eval(const DoubleDouble &x) {
+  DoubleDouble p;
+  p.hi = x.hi;
+  double x_hi_sq = x.hi * x.hi;
+  // c0 ~ x_hi^2 * 1/5 - 1/3
+  double c0 = fputil::multiply_add(x_hi_sq, 0x1.999999999999ap-3,
+                                   -0x1.5555555555555p-2);
+  // c1 ~ x_hi^2 * 1/9 - 1/7
+  double c1 = fputil::multiply_add(x_hi_sq, 0x1.c71c71c71c71cp-4,
+                                   -0x1.2492492492492p-3);
+  // x_hi^3
+  double x_hi_3 = x_hi_sq * x.hi;
+  // x_hi^4
+  double x_hi_4 = x_hi_sq * x_hi_sq;
+  // d0 ~ 1/3 - x_hi^2 / 5 + x_hi^4 / 7 - x_hi^6 / 9
+  double d0 = fputil::multiply_add(x_hi_4, c1, c0);
+  // x_lo - x_lo * x_hi^2 + x_lo * x_hi^4
+  double d1 = fputil::multiply_add(x_hi_4 - x_hi_sq, x.lo, x.lo);
+  // p.lo ~ -x_hi^3/3 + x_hi^5/5 - x_hi^7/7 + x_hi^9/9 +
+  //        + x_lo * (1 - x_hi^2 + x_hi^4)
+  p.lo = fputil::multiply_add(x_hi_3, d0, d1);
+  return p;
+}
+
+} // anonymous namespace
+
+// There are several range reduction steps we can take for atan2(y, x) as
+// follow:
+
+// * Range reduction 1: signness
+// atan2(y, x) will return a number between -PI and PI representing the angle
+// forming by the 0x axis and the vector (x, y) on the 0xy-plane.
+// In particular, we have that:
+//   atan2(y, x) = atan( y/x )         if x >= 0 and y >= 0 (I-quadrant)
+//               = pi + atan( y/x )    if x < 0 and y >= 0  (II-quadrant)
+//               = -pi + atan( y/x )   if x < 0 and y < 0   (III-quadrant)
+//               = atan( y/x )         if x >= 0 and y < 0  (IV-quadrant)
+// Since atan function is odd, we can use the formula:
+//   atan(-u) = -atan(u)
+// to adjust the above conditions a bit further:
+//   atan2(y, x) = atan( |y|/|x| )         if x >= 0 and y >= 0 (I-quadrant)
+//               = pi - atan( |y|/|x| )    if x < 0 and y >= 0  (II-quadrant)
+//               = -pi + atan( |y|/|x| )   if x < 0 and y < 0   (III-quadrant)
+//               = -atan( |y|/|x| )        if x >= 0 and y < 0  (IV-quadrant)
+// Which can be simplified to:
+//   atan2(y, x) = sign(y) * atan( |y|/|x| )             if x >= 0
+//               = sign(y) * (pi - atan( |y|/|x| ))      if x < 0
+
+// * Range reduction 2: reciprocal
+// Now that the argument inside atan is positive, we can use the formula:
+//   atan(1/x) = pi/2 - atan(x)
+// to make the argument inside atan <= 1 as follow:
+//   atan2(y, x) = sign(y) * atan( |y|/|x|)            if 0 <= |y| <= x
+//               = sign(y) * (pi/2 - atan( |x|/|y| )   if 0 <= x < |y|
+//               = sign(y) * (pi - atan( |y|/|x| ))    if 0 <= |y| <= -x
+//               = sign(y) * (pi/2 + atan( |x|/|y| ))  if 0 <= -x < |y|
+
+// * Range reduction 3: look up table.
+// After the previous two range reduction steps, we reduce the problem to
+// compute atan(u) with 0 <= u <= 1, or to be precise:
+//   atan( n / d ) where n = min(|x|, |y|) and d = max(|x|, |y|).
+// An accurate polynomial approximation for the whole [0, 1] input range will
+// require a very large degree.  To make it more efficient, we reduce the input
+// range further by finding an integer idx such that:
+//   | n/d - idx/64 | <= 1/128.
+// In particular,
+//   idx := round(2^6 * n/d)
+// Then for the fast pass, we find a polynomial approximation for:
+//   atan( n/d ) ~ atan( idx/64 ) + (n/d - idx/64) * Q(n/d - idx/64)
+// For the accurate pass, we use the addition formula:
+//   atan( n/d ) - atan( idx/64 ) = atan( (n/d - idx/64)/(1 + (n*idx)/(64*d)) )
+//                                = atan( (n - d*(idx/64))/(d + n*(idx/64)) )
+// And for the fast pass, we use degree-9 Taylor polynomial to compute the RHS:
+//   atan(u) ~ P(u) = u - u^3/3 + u^5/5 - u^7/7 + u^9/9
+// with absolute errors bounded by:
+//   |atan(u) - P(u)| < |u|^11 / 11 < 2^-80
+// and relative errors bounded by:
+//   |(atan(u) - P(u)) / P(u)| < u^10 / 11 < 2^-73.
+
+LLVM_LIBC_FUNCTION(double, atan2, (double y, double x)) {
+  using FPBits = fputil::FPBits<double>;
+
+  constexpr double IS_NEG[2] = {1.0, -1.0};
+  constexpr DoubleDouble ZERO = {0.0, 0.0};
+  constexpr DoubleDouble MZERO = {-0.0, -0.0};
+  constexpr DoubleDouble PI = {0x1.1a62633145c07p-53, 0x1.921fb54442d18p+1};
+  constexpr DoubleDouble MPI = {-0x1.1a62633145c07p-53, -0x1.921fb54442d18p+1};
+  constexpr DoubleDouble PI_OVER_2 = {0x1.1a62633145c07p-54,
+                                      0x1.921fb54442d18p0};
+  constexpr DoubleDouble MPI_OVER_2 = {-0x1.1a62633145c07p-54,
+                                       -0x1.921fb54442d18p0};
+  constexpr DoubleDouble PI_OVER_4 = {0x1.1a62633145c07p-55,
+                                      0x1.921fb54442d18p-1};
+  constexpr DoubleDouble THREE_PI_OVER_4 = {0x1.a79394c9e8a0ap-54,
+                                            0x1.2d97c7f3321d2p+1};
+  // Adjustment for constant term:
+  //   CONST_ADJ[x_sign][y_sign][recip]
+  constexpr DoubleDouble CONST_ADJ[2][2][2] = {
+      {{ZERO, MPI_OVER_2}, {MZERO, MPI_OVER_2}},
+      {{MPI, PI_OVER_2}, {MPI, PI_OVER_2}}};
+
+  FPBits x_bits(x), y_bits(y);
+  bool x_sign = x_bits.sign().is_neg();
+  bool y_sign = y_bits.sign().is_neg();
+  x_bits = x_bits.abs();
+  y_bits = y_bits.abs();
+  uint64_t x_abs = x_bits.uintval();
+  uint64_t y_abs = y_bits.uintval();
+  bool recip = x_abs < y_abs;
+  uint64_t min_abs = recip ? x_abs : y_abs;
+  uint64_t max_abs = !recip ? x_abs : y_abs;
+  unsigned min_exp = static_cast<unsigned>(min_abs >> FPBits::FRACTION_LEN);
+  unsigned max_exp = static_cast<unsigned>(max_abs >> FPBits::FRACTION_LEN);
+
+  double num = FPBits(min_abs).get_val();
+  double den = FPBits(max_abs).get_val();
+
+  // Check for exceptional cases, whether inputs are 0, inf, nan, or close to
+  // overflow, or close to underflow.
+  if (LIBC_UNLIKELY(max_exp > 0x7ffU - 128U || min_exp < 128U)) {
+    if (x_bits.is_nan() || y_bits.is_nan())
+      return FPBits::quiet_nan().get_val();
+    unsigned x_except = x_abs == 0 ? 0 : (FPBits(x_abs).is_inf() ? 2 : 1);
+    unsigned y_except = y_abs == 0 ? 0 : (FPBits(y_abs).is_inf() ? 2 : 1);
+
+    // Exceptional cases:
+    //   EXCEPT[y_except][x_except][x_is_neg]
+    // with x_except & y_except:
+    //   0: zero
+    //   1: finite, non-zero
+    //   2: infinity
+    constexpr DoubleDouble EXCEPTS[3][3][2] = {
+        {{ZERO, PI}, {ZERO, PI}, {ZERO, PI}},
+        {{PI_OVER_2, PI_OVER_2}, {ZERO, ZERO}, {ZERO, PI}},
+        {{PI_OVER_2, PI_OVER_2},
+         {PI_OVER_2, PI_OVER_2},
+         {PI_OVER_4, THREE_PI_OVER_4}},
+    };
+
+    if ((x_except != 1) || (y_except != 1)) {
+      DoubleDouble r = EXCEPTS[y_except][x_except][x_sign];
+      return fputil::multiply_add(IS_NEG[y_sign], r.hi, IS_NEG[y_sign] * r.lo);
+    }
+    bool scale_up = min_exp < 128U;
+    bool scale_down = max_exp > 0x7ffU - 128U;
+    // At least one input is denormal, multiply both numerator and denominator
+    // by some large enough power of 2 to normalize denormal inputs.
+    if (scale_up) {
+      num *= 0x1.0p64;
+      if (!scale_down)
+        den *= 0x1.0p64;
+    } else if (scale_down) {
+      den *= 0x1.0p-64;
+      if (!scale_up)
+        num *= 0x1.0p-64;
+    }
+
+    min_abs = FPBits(num).uintval();
+    max_abs = FPBits(den).uintval();
+    min_exp = static_cast<unsigned>(min_abs >> FPBits::FRACTION_LEN);
+    max_exp = static_cast<unsigned>(max_abs >> FPBits::FRACTION_LEN);
+  }
+
+  double final_sign = IS_NEG[(x_sign != y_sign) != recip];
+  DoubleDouble const_term = CONST_ADJ[x_sign][y_sign][recip];
+  unsigned exp_diff = max_exp - min_exp;
+  // We have the following bound for normalized n and d:
+  //   2^(-exp_diff - 1) < n/d < 2^(-exp_diff + 1).
+  if (LIBC_UNLIKELY(exp_diff > 54)) {
+    return fputil::multiply_add(final_sign, const_term.hi,
+                                final_sign * (const_term.lo + num / den));
+  }
+
+  double k = fputil::nearest_integer(64.0 * num / den);
+  unsigned idx = static_cast<unsigned>(k);
+  // k = idx / 64
+  k *= 0x1.0p-6;
+
+  // Range reduction:
+  // atan(n/d) - atan(k/64) = atan((n/d - k/64) / (1 + (n/d) * (k/64)))
+  //                        = atan((n - d * k/64)) / (d + n * k/64))
+  DoubleDouble num_k = fputil::exact_mult(num, k);
+  DoubleDouble den_k = fputil::exact_mult(den, k);
+
+  // num_dd = n - d * k
+  DoubleDouble num_dd = fputil::exact_add(num - den_k.hi, -den_k.lo);
+  // den_dd = d + n * k
+  DoubleDouble den_dd = fputil::exact_add(den, num_k.hi);
+  den_dd.lo += num_k.lo;
+
+  // q = (n - d * k) / (d + n * k)
+  DoubleDouble q = fputil::div(num_dd, den_dd);
+  // p ~ atan(q)
+  DoubleDouble p = atan_eval(q);
+
+  DoubleDouble r = fputil::add(const_term, fputil::add(ATAN_I[idx], p));
+  r.hi *= final_sign;
+  r.lo *= final_sign;
+
+  return r.hi + r.lo;
+}
+
+} // namespace LIBC_NAMESPACE_DECL
diff --git a/libc/startup/gpu/CMakeLists.txt b/libc/startup/gpu/CMakeLists.txt
@@ -26,6 +26,16 @@ function(add_startup_object name)
     PROPERTIES
       OUTPUT_NAME ${name}.o
   )
+
+  # Make an executable target of relocatable bitcode for clang if needed.
+  if(LLVM_ENABLE_PER_TARGET_RUNTIME_DIR)
+    add_executable(${fq_target_name}.exe $<TARGET_OBJECTS:${fq_target_name}>)
+    set_target_properties(${fq_target_name}.exe PROPERTIES
+      RUNTIME_OUTPUT_DIRECTORY ${LIBC_LIBRARY_DIR}
+      RUNTIME_OUTPUT_NAME ${name}.o)
+    target_link_options(${fq_target_name}.exe PRIVATE
+                        "-nostdlib" "-flto" "-Wl,--lto-emit-llvm" "-march= ")
+  endif()
 endfunction()
 
 if(EXISTS ${CMAKE_CURRENT_SOURCE_DIR}/${LIBC_TARGET_ARCHITECTURE})

diff --git a/libc/test/src/math/CMakeLists.txt b/libc/test/src/math/CMakeLists.txt
@@ -2044,6 +2044,18 @@ add_fp_unittest(
     libc.src.__support.FPUtil.fp_bits
 )
 
+add_fp_unittest(
+  atan2_test
+  NEED_MPFR
+  SUITE
+    libc-math-unittests
+  SRCS
+    atan2_test.cpp
+  DEPENDS
+    libc.src.math.atan2
+    libc.src.__support.FPUtil.fp_bits
+)
+
 add_fp_unittest(
   f16add_test
   NEED_MPFR

diff --git a/libc/test/src/math/atan2_test.cpp b/libc/test/src/math/atan2_test.cpp
@@ -0,0 +1,125 @@
+//===-- Unittests for atan2 -----------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "src/__support/FPUtil/FPBits.h"
+#include "src/math/atan2.h"
+#include "test/UnitTest/FPMatcher.h"
+#include "test/UnitTest/Test.h"
+#include "utils/MPFRWrapper/MPFRUtils.h"
+
+using LlvmLibcAtan2Test = LIBC_NAMESPACE::testing::FPTest<double>;
+using LIBC_NAMESPACE::testing::tlog;
+
+namespace mpfr = LIBC_NAMESPACE::testing::mpfr;
+
+TEST_F(LlvmLibcAtan2Test, TrickyInputs) {
+  mpfr::BinaryInput<double> inputs[] = {
+      {0x1.0853408534085p-2, 0x1.e7b54166c6126p-2},
+      {FPBits::inf().get_val(), 0x0.0000000000001p-1022},
+  };
+
+  for (mpfr::BinaryInput<double> &input : inputs) {
+    double x = input.x;
+    double y = input.y;
+    mpfr::RoundingMode rm = mpfr::RoundingMode::Downward;
+    mpfr::ForceRoundingMode rr(rm);
+    ASSERT_MPFR_MATCH(mpfr::Operation::Atan2, input,
+                      LIBC_NAMESPACE::atan2(x, y), 0.5, rm);
+    input.x = -input.x;
+    ASSERT_MPFR_MATCH_ALL_ROUNDING(mpfr::Operation::Atan2, input,
+                                   LIBC_NAMESPACE::atan2(-x, y), 0.5);
+    input.y = -input.y;
+    ASSERT_MPFR_MATCH_ALL_ROUNDING(mpfr::Operation::Atan2, input,
+                                   LIBC_NAMESPACE::atan2(-x, -y), 0.5);
+    input.x = -input.x;
+    ASSERT_MPFR_MATCH_ALL_ROUNDING(mpfr::Operation::Atan2, input,
+                                   LIBC_NAMESPACE::atan2(x, -y), 0.5);
+  }
+}
+
+TEST_F(LlvmLibcAtan2Test, InDoubleRange) {
+  constexpr uint64_t X_COUNT = 123;
+  constexpr uint64_t X_START = FPBits(0.25).uintval();
+  constexpr uint64_t X_STOP = FPBits(4.0).uintval();
+  constexpr uint64_t X_STEP = (X_STOP - X_START) / X_COUNT;
+
+  constexpr uint64_t Y_COUNT = 137;
+  constexpr uint64_t Y_START = FPBits(0.25).uintval();
+  constexpr uint64_t Y_STOP = FPBits(4.0).uintval();
+  constexpr uint64_t Y_STEP = (Y_STOP - Y_START) / Y_COUNT;
+
+  auto test = [&](mpfr::RoundingMode rounding_mode) {
+    mpfr::ForceRoundingMode __r(rounding_mode);
+    if (!__r.success)
+      return;
+
+    uint64_t fails = 0;
+    uint64_t finite_count = 0;
+    uint64_t total_count = 0;
+    double failed_x = 0.0, failed_y = 0.0, failed_r = 0.0;
+    double tol = 0.5;
+
+    for (uint64_t i = 0, v = X_START; i <= X_COUNT; ++i, v += X_STEP) {
+      double x = FPBits(v).get_val();
+      if (FPBits(x).is_inf_or_nan() || x < 0.0)
+        continue;
+
+      for (uint64_t j = 0, w = Y_START; j <= Y_COUNT; ++j, w += Y_STEP) {
+        double y = FPBits(w).get_val();
+        if (FPBits(y).is_inf_or_nan())
+          continue;
+
+        double result = LIBC_NAMESPACE::atan2(x, y);
+        ++total_count;
+        if (FPBits(result).is_inf_or_nan())
+          continue;
+
+        ++finite_count;
+        mpfr::BinaryInput<double> inputs{x, y};
+
+        if (!TEST_MPFR_MATCH_ROUNDING_SILENTLY(mpfr::Operation::Atan2, inputs,
+                                               result, 0.5, rounding_mode)) {
+          ++fails;
+          while (!TEST_MPFR_MATCH_ROUNDING_SILENTLY(
+              mpfr::Operation::Atan2, inputs, result, tol, rounding_mode)) {
+            failed_x = x;
+            failed_y = y;
+            failed_r = result;
+
+            if (tol > 1000.0)
+              break;
+
+            tol *= 2.0;
+          }
+        }
+      }
+    }
+    if (fails || (finite_count < total_count)) {
+      tlog << " Atan2 failed: " << fails << "/" << finite_count << "/"
+           << total_count << " tests.\n"
+           << "   Max ULPs is at most: " << static_cast<uint64_t>(tol) << ".\n";
+    }
+    if (fails) {
+      mpfr::BinaryInput<double> inputs{failed_x, failed_y};
+      EXPECT_MPFR_MATCH(mpfr::Operation::Atan2, inputs, failed_r, 0.5,
+                        rounding_mode);
+    }
+  };
+
+  tlog << " Test Rounding To Nearest...\n";
+  test(mpfr::RoundingMode::Nearest);
+
+  tlog << " Test Rounding Downward...\n";
+  test(mpfr::RoundingMode::Downward);
+
+  tlog << " Test Rounding Upward...\n";
+  test(mpfr::RoundingMode::Upward);
+
+  tlog << " Test Rounding Toward Zero...\n";
+  test(mpfr::RoundingMode::TowardZero);
+}
diff --git a/libc/test/src/math/smoke/CMakeLists.txt b/libc/test/src/math/smoke/CMakeLists.txt
@@ -3531,6 +3531,16 @@ add_fp_unittest(
     libc.src.__support.FPUtil.fp_bits
 )
 
+add_fp_unittest(
+  atan2_test
+  SUITE
+    libc-math-smoke-tests
+  SRCS
+    atan2_test.cpp
+  DEPENDS
+    libc.src.math.atan2
+)
+
 add_fp_unittest(
   scalblnf16_test
   SUITE

diff --git a/libc/test/src/math/smoke/atan2_test.cpp b/libc/test/src/math/smoke/atan2_test.cpp
@@ -0,0 +1,22 @@
+//===-- Unittests for atan2 -----------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "src/math/atan2.h"
+#include "test/UnitTest/FPMatcher.h"
+#include "test/UnitTest/Test.h"
+
+using LlvmLibcAtan2Test = LIBC_NAMESPACE::testing::FPTest<double>;
+
+TEST_F(LlvmLibcAtan2Test, SpecialNumbers) {
+  EXPECT_FP_EQ_ALL_ROUNDING(aNaN, LIBC_NAMESPACE::atan2(aNaN, zero));
+  EXPECT_FP_EQ_ALL_ROUNDING(aNaN, LIBC_NAMESPACE::atan2(1.0, aNaN));
+  EXPECT_FP_EQ_ALL_ROUNDING(0.0, LIBC_NAMESPACE::atan2(zero, zero));
+  EXPECT_FP_EQ_ALL_ROUNDING(-0.0, LIBC_NAMESPACE::atan2(-0.0, zero));
+  EXPECT_FP_EQ_ALL_ROUNDING(0.0, LIBC_NAMESPACE::atan2(1.0, inf));
+  EXPECT_FP_EQ_ALL_ROUNDING(-0.0, LIBC_NAMESPACE::atan2(-1.0, inf));
+}