diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -12508,6 +12508,7 @@ Value *CodeGenFunction::EmitX86BuiltinExpr(unsigned BuiltinID,
 
   SmallVector<Value*, 4> Ops;
   bool IsMaskFCmp = false;
+  bool IsConjFMA = false;
 
   // Find out if any arguments are required to be integer constant expressions.
   unsigned ICEArguments = 0;
@@ -15046,6 +15047,36 @@ Value *CodeGenFunction::EmitX86BuiltinExpr(unsigned BuiltinID,
     Builder.SetInsertPoint(End);
     return Builder.CreateExtractValue(Call, 0);
   }
+  case X86::BI__builtin_ia32_vfcmaddcph512_mask:
+    IsConjFMA = true;
+    LLVM_FALLTHROUGH;
+  case X86::BI__builtin_ia32_vfmaddcph512_mask: {
+    Intrinsic::ID IID = IsConjFMA
+                            ? Intrinsic::x86_avx512fp16_mask_vfcmadd_cph_512
+                            : Intrinsic::x86_avx512fp16_mask_vfmadd_cph_512;
+    Value *Call = Builder.CreateCall(CGM.getIntrinsic(IID), Ops);
+    return EmitX86Select(*this, Ops[3], Call, Ops[0]);
+  }
+  case X86::BI__builtin_ia32_vfcmaddcsh_round_mask:
+    IsConjFMA = true;
+    LLVM_FALLTHROUGH;
+  case X86::BI__builtin_ia32_vfmaddcsh_round_mask: {
+    Intrinsic::ID IID = IsConjFMA ? Intrinsic::x86_avx512fp16_mask_vfcmadd_csh
+                                  : Intrinsic::x86_avx512fp16_mask_vfmadd_csh;
+    Value *Call = Builder.CreateCall(CGM.getIntrinsic(IID), Ops);
+    Value *And = Builder.CreateAnd(Ops[3], llvm::ConstantInt::get(Int8Ty, 1));
+    return EmitX86Select(*this, And, Call, Ops[0]);
+  }
+  case X86::BI__builtin_ia32_vfcmaddcsh_round_mask3:
+    IsConjFMA = true;
+    LLVM_FALLTHROUGH;
+  case X86::BI__builtin_ia32_vfmaddcsh_round_mask3: {
+    Intrinsic::ID IID = IsConjFMA ? Intrinsic::x86_avx512fp16_mask_vfcmadd_csh
+                                  : Intrinsic::x86_avx512fp16_mask_vfmadd_csh;
+    Value *Call = Builder.CreateCall(CGM.getIntrinsic(IID), Ops);
+    static constexpr int Mask[] = {0, 5, 6, 7};
+    return Builder.CreateShuffleVector(Call, Ops[2], Mask);
+  }
   }
 }
 

diff --git a/clang/lib/CodeGen/CGOpenMPRuntime.cpp b/clang/lib/CodeGen/CGOpenMPRuntime.cpp
@@ -1560,13 +1560,22 @@ llvm::Type *CGOpenMPRuntime::getKmpc_MicroPointerTy() {
 }
 
 llvm::FunctionCallee
-CGOpenMPRuntime::createForStaticInitFunction(unsigned IVSize, bool IVSigned) {
+CGOpenMPRuntime::createForStaticInitFunction(unsigned IVSize, bool IVSigned,
+                                             bool IsGPUDistribute) {
   assert((IVSize == 32 || IVSize == 64) &&
          "IV size is not compatible with the omp runtime");
-  StringRef Name = IVSize == 32 ? (IVSigned ? "__kmpc_for_static_init_4"
-                                            : "__kmpc_for_static_init_4u")
-                                : (IVSigned ? "__kmpc_for_static_init_8"
-                                            : "__kmpc_for_static_init_8u");
+  StringRef Name;
+  if (IsGPUDistribute)
+    Name = IVSize == 32 ? (IVSigned ? "__kmpc_distribute_static_init_4"
+                                    : "__kmpc_distribute_static_init_4u")
+                        : (IVSigned ? "__kmpc_distribute_static_init_8"
+                                    : "__kmpc_distribute_static_init_8u");
+  else
+    Name = IVSize == 32 ? (IVSigned ? "__kmpc_for_static_init_4"
+                                    : "__kmpc_for_static_init_4u")
+                        : (IVSigned ? "__kmpc_for_static_init_8"
+                                    : "__kmpc_for_static_init_8u");
+
   llvm::Type *ITy = IVSize == 32 ? CGM.Int32Ty : CGM.Int64Ty;
   auto *PtrTy = llvm::PointerType::getUnqual(ITy);
   llvm::Type *TypeParams[] = {
@@ -2826,7 +2835,7 @@ void CGOpenMPRuntime::emitForStaticInit(CodeGenFunction &CGF,
                                                  : OMP_IDENT_WORK_SECTIONS);
   llvm::Value *ThreadId = getThreadID(CGF, Loc);
   llvm::FunctionCallee StaticInitFunction =
-      createForStaticInitFunction(Values.IVSize, Values.IVSigned);
+      createForStaticInitFunction(Values.IVSize, Values.IVSigned, false);
   auto DL = ApplyDebugLocation::CreateDefaultArtificial(CGF, Loc);
   emitForStaticInitCall(CGF, UpdatedLocation, ThreadId, StaticInitFunction,
                         ScheduleNum, ScheduleKind.M1, ScheduleKind.M2, Values);
@@ -2841,8 +2850,13 @@ void CGOpenMPRuntime::emitDistributeStaticInit(
   llvm::Value *UpdatedLocation =
       emitUpdateLocation(CGF, Loc, OMP_IDENT_WORK_DISTRIBUTE);
   llvm::Value *ThreadId = getThreadID(CGF, Loc);
-  llvm::FunctionCallee StaticInitFunction =
-      createForStaticInitFunction(Values.IVSize, Values.IVSigned);
+  llvm::FunctionCallee StaticInitFunction;
+  bool isGPUDistribute =
+      CGM.getLangOpts().OpenMPIsDevice &&
+      (CGM.getTriple().isAMDGCN() || CGM.getTriple().isNVPTX());
+  StaticInitFunction = createForStaticInitFunction(
+      Values.IVSize, Values.IVSigned, isGPUDistribute);
+
   emitForStaticInitCall(CGF, UpdatedLocation, ThreadId, StaticInitFunction,
                         ScheduleNum, OMPC_SCHEDULE_MODIFIER_unknown,
                         OMPC_SCHEDULE_MODIFIER_unknown, Values);
@@ -2863,9 +2877,16 @@ void CGOpenMPRuntime::emitForStaticFinish(CodeGenFunction &CGF,
                                    : OMP_IDENT_WORK_SECTIONS),
       getThreadID(CGF, Loc)};
   auto DL = ApplyDebugLocation::CreateDefaultArtificial(CGF, Loc);
-  CGF.EmitRuntimeCall(OMPBuilder.getOrCreateRuntimeFunction(
-                          CGM.getModule(), OMPRTL___kmpc_for_static_fini),
-                      Args);
+  if (isOpenMPDistributeDirective(DKind) && CGM.getLangOpts().OpenMPIsDevice &&
+      (CGM.getTriple().isAMDGCN() || CGM.getTriple().isNVPTX()))
+    CGF.EmitRuntimeCall(
+        OMPBuilder.getOrCreateRuntimeFunction(
+            CGM.getModule(), OMPRTL___kmpc_distribute_static_fini),
+        Args);
+  else
+    CGF.EmitRuntimeCall(OMPBuilder.getOrCreateRuntimeFunction(
+                            CGM.getModule(), OMPRTL___kmpc_for_static_fini),
+                        Args);
 }
 
 void CGOpenMPRuntime::emitForOrderedIterationEnd(CodeGenFunction &CGF,

diff --git a/clang/lib/CodeGen/CGOpenMPRuntime.h b/clang/lib/CodeGen/CGOpenMPRuntime.h
@@ -795,9 +795,11 @@ class CGOpenMPRuntime {
   llvm::Type *getKmpc_MicroPointerTy();
 
   /// Returns __kmpc_for_static_init_* runtime function for the specified
-  /// size \a IVSize and sign \a IVSigned.
+  /// size \a IVSize and sign \a IVSigned. Will create a distribute call
+  /// __kmpc_distribute_static_init* if \a IsGPUDistribute is set.
   llvm::FunctionCallee createForStaticInitFunction(unsigned IVSize,
-                                                   bool IVSigned);
+                                                   bool IVSigned,
+                                                   bool IsGPUDistribute);
 
   /// Returns __kmpc_dispatch_init_* runtime function for the specified
   /// size \a IVSize and sign \a IVSigned.

diff --git a/clang/lib/Driver/Driver.cpp b/clang/lib/Driver/Driver.cpp
@@ -1216,8 +1216,14 @@ Compilation *Driver::BuildCompilation(ArrayRef<const char *> ArgList) {
 
 static void printArgList(raw_ostream &OS, const llvm::opt::ArgList &Args) {
   llvm::opt::ArgStringList ASL;
-  for (const auto *A : Args)
+  for (const auto *A : Args) {
+    // Use user's original spelling of flags. For example, use
+    // `/source-charset:utf-8` instead of `-finput-charset=utf-8` if the user
+    // wrote the former.
+    while (A->getAlias())
+      A = A->getAlias();
     A->render(Args, ASL);
+  }
 
   for (auto I = ASL.begin(), E = ASL.end(); I != E; ++I) {
     if (I != ASL.begin())

diff --git a/clang/lib/Driver/ToolChain.cpp b/clang/lib/Driver/ToolChain.cpp
@@ -1026,7 +1026,7 @@ void ToolChain::AddCudaIncludeArgs(const ArgList &DriverArgs,
 void ToolChain::AddHIPIncludeArgs(const ArgList &DriverArgs,
                                   ArgStringList &CC1Args) const {}
 
-llvm::SmallVector<std::string, 12>
+llvm::SmallVector<ToolChain::BitCodeLibraryInfo, 12>
 ToolChain::getHIPDeviceLibs(const ArgList &DriverArgs) const {
   return {};
 }

diff --git a/clang/lib/Driver/ToolChains/Gnu.cpp b/clang/lib/Driver/ToolChains/Gnu.cpp
@@ -2084,21 +2084,19 @@ void Generic_GCC::GCCInstallationDetector::AddDefaultGCCPrefixes(
   static const char *const AArch64LibDirs[] = {"/lib64", "/lib"};
   static const char *const AArch64Triples[] = {
       "aarch64-none-linux-gnu", "aarch64-linux-gnu", "aarch64-redhat-linux",
-      "aarch64-suse-linux", "aarch64-linux-android"};
+      "aarch64-suse-linux"};
   static const char *const AArch64beLibDirs[] = {"/lib"};
   static const char *const AArch64beTriples[] = {"aarch64_be-none-linux-gnu",
                                                  "aarch64_be-linux-gnu"};
 
   static const char *const ARMLibDirs[] = {"/lib"};
-  static const char *const ARMTriples[] = {"arm-linux-gnueabi",
-                                           "arm-linux-androideabi"};
+  static const char *const ARMTriples[] = {"arm-linux-gnueabi"};
   static const char *const ARMHFTriples[] = {"arm-linux-gnueabihf",
                                              "armv7hl-redhat-linux-gnueabi",
                                              "armv6hl-suse-linux-gnueabi",
                                              "armv7hl-suse-linux-gnueabi"};
   static const char *const ARMebLibDirs[] = {"/lib"};
-  static const char *const ARMebTriples[] = {"armeb-linux-gnueabi",
-                                             "armeb-linux-androideabi"};
+  static const char *const ARMebTriples[] = {"armeb-linux-gnueabi"};
   static const char *const ARMebHFTriples[] = {
       "armeb-linux-gnueabihf", "armebv7hl-redhat-linux-gnueabi"};
 
@@ -2112,17 +2110,15 @@ void Generic_GCC::GCCInstallationDetector::AddDefaultGCCPrefixes(
       "x86_64-redhat-linux",    "x86_64-suse-linux",
       "x86_64-manbo-linux-gnu", "x86_64-linux-gnu",
       "x86_64-slackware-linux", "x86_64-unknown-linux",
-      "x86_64-amazon-linux",    "x86_64-linux-android"};
+      "x86_64-amazon-linux"};
   static const char *const X32Triples[] = {"x86_64-linux-gnux32",
                                            "x86_64-pc-linux-gnux32"};
   static const char *const X32LibDirs[] = {"/libx32", "/lib"};
   static const char *const X86LibDirs[] = {"/lib32", "/lib"};
   static const char *const X86Triples[] = {
-      "i586-linux-gnu",     "i686-linux-gnu",
-      "i686-pc-linux-gnu",  "i386-redhat-linux6E",
-      "i686-redhat-linux",  "i386-redhat-linux",
-      "i586-suse-linux",    "i686-montavista-linux",
-      "i686-linux-android", "i686-gnu",
+      "i586-linux-gnu",      "i686-linux-gnu",        "i686-pc-linux-gnu",
+      "i386-redhat-linux6E", "i686-redhat-linux",     "i386-redhat-linux",
+      "i586-suse-linux",     "i686-montavista-linux", "i686-gnu",
   };
 
   static const char *const M68kLibDirs[] = {"/lib"};
@@ -2135,8 +2131,7 @@ void Generic_GCC::GCCInstallationDetector::AddDefaultGCCPrefixes(
       "mips-img-linux-gnu", "mipsisa32r6-linux-gnu"};
   static const char *const MIPSELLibDirs[] = {"/lib"};
   static const char *const MIPSELTriples[] = {
-      "mipsel-linux-gnu", "mips-img-linux-gnu", "mipsisa32r6el-linux-gnu",
-      "mipsel-linux-android"};
+      "mipsel-linux-gnu", "mips-img-linux-gnu", "mipsisa32r6el-linux-gnu"};
 
   static const char *const MIPS64LibDirs[] = {"/lib64", "/lib"};
   static const char *const MIPS64Triples[] = {
@@ -2147,8 +2142,7 @@ void Generic_GCC::GCCInstallationDetector::AddDefaultGCCPrefixes(
   static const char *const MIPS64ELTriples[] = {
       "mips64el-linux-gnu",      "mips-mti-linux-gnu",
       "mips-img-linux-gnu",      "mips64el-linux-gnuabi64",
-      "mipsisa64r6el-linux-gnu", "mipsisa64r6el-linux-gnuabi64",
-      "mips64el-linux-android"};
+      "mipsisa64r6el-linux-gnu", "mipsisa64r6el-linux-gnuabi64"};
 
   static const char *const MIPSN32LibDirs[] = {"/lib32"};
   static const char *const MIPSN32Triples[] = {"mips64-linux-gnuabin32",

diff --git a/clang/lib/Driver/ToolChains/HIP.cpp b/clang/lib/Driver/ToolChains/HIP.cpp
@@ -88,8 +88,8 @@ void AMDGCN::Linker::constructLldCommand(Compilation &C, const JobAction &JA,
 
   if (Args.hasFlag(options::OPT_fgpu_sanitize, options::OPT_fno_gpu_sanitize,
                    false))
-    llvm::for_each(TC.getHIPDeviceLibs(Args), [&](StringRef BCFile) {
-      LldArgs.push_back(Args.MakeArgString(BCFile));
+    llvm::for_each(TC.getHIPDeviceLibs(Args), [&](auto BCFile) {
+      LldArgs.push_back(Args.MakeArgString(BCFile.Path));
     });
 
   const char *Lld = Args.MakeArgString(getToolChain().GetProgramPath("lld"));
@@ -276,9 +276,10 @@ void HIPToolChain::addClangTargetOptions(
     CC1Args.push_back("-fapply-global-visibility-to-externs");
   }
 
-  llvm::for_each(getHIPDeviceLibs(DriverArgs), [&](StringRef BCFile) {
-    CC1Args.push_back("-mlink-builtin-bitcode");
-    CC1Args.push_back(DriverArgs.MakeArgString(BCFile));
+  llvm::for_each(getHIPDeviceLibs(DriverArgs), [&](auto BCFile) {
+    CC1Args.push_back(BCFile.ShouldInternalize ? "-mlink-builtin-bitcode"
+                                               : "-mlink-bitcode-file");
+    CC1Args.push_back(DriverArgs.MakeArgString(BCFile.Path));
   });
 }
 
@@ -359,9 +360,9 @@ VersionTuple HIPToolChain::computeMSVCVersion(const Driver *D,
   return HostTC.computeMSVCVersion(D, Args);
 }
 
-llvm::SmallVector<std::string, 12>
+llvm::SmallVector<ToolChain::BitCodeLibraryInfo, 12>
 HIPToolChain::getHIPDeviceLibs(const llvm::opt::ArgList &DriverArgs) const {
-  llvm::SmallVector<std::string, 12> BCLibs;
+  llvm::SmallVector<BitCodeLibraryInfo, 12> BCLibs;
   if (DriverArgs.hasArg(options::OPT_nogpulib))
     return {};
   ArgStringList LibraryPaths;
@@ -382,7 +383,7 @@ HIPToolChain::getHIPDeviceLibs(const llvm::opt::ArgList &DriverArgs) const {
         llvm::sys::path::append(Path, BCName);
         FullName = Path;
         if (llvm::sys::fs::exists(FullName)) {
-          BCLibs.push_back(FullName.str());
+          BCLibs.push_back(FullName);
           return;
         }
       }
@@ -409,22 +410,23 @@ HIPToolChain::getHIPDeviceLibs(const llvm::opt::ArgList &DriverArgs) const {
         getDriver().Diag(DiagID);
         return {};
       } else
-        BCLibs.push_back(AsanRTL.str());
+        BCLibs.push_back({AsanRTL.str(), /*ShouldInternalize=*/false});
     }
 
     // Add the HIP specific bitcode library.
-    BCLibs.push_back(RocmInstallation.getHIPPath().str());
+    BCLibs.push_back(RocmInstallation.getHIPPath());
 
     // Add common device libraries like ocml etc.
-    BCLibs.append(getCommonDeviceLibNames(DriverArgs, GpuArch.str()));
+    for (auto N : getCommonDeviceLibNames(DriverArgs, GpuArch.str()))
+      BCLibs.push_back(StringRef(N));
 
     // Add instrument lib.
     auto InstLib =
         DriverArgs.getLastArgValue(options::OPT_gpu_instrument_lib_EQ);
     if (InstLib.empty())
       return BCLibs;
     if (llvm::sys::fs::exists(InstLib))
-      BCLibs.push_back(InstLib.str());
+      BCLibs.push_back(InstLib);
     else
       getDriver().Diag(diag::err_drv_no_such_file) << InstLib;
   }

diff --git a/clang/lib/Driver/ToolChains/HIP.h b/clang/lib/Driver/ToolChains/HIP.h
@@ -83,7 +83,7 @@ class LLVM_LIBRARY_VISIBILITY HIPToolChain final : public ROCMToolChain {
                            llvm::opt::ArgStringList &CC1Args) const override;
   void AddHIPIncludeArgs(const llvm::opt::ArgList &DriverArgs,
                          llvm::opt::ArgStringList &CC1Args) const override;
-  llvm::SmallVector<std::string, 12>
+  llvm::SmallVector<BitCodeLibraryInfo, 12>
   getHIPDeviceLibs(const llvm::opt::ArgList &Args) const override;
 
   SanitizerMask getSupportedSanitizers() const override;

diff --git a/clang/lib/Headers/avx512fp16intrin.h b/clang/lib/Headers/avx512fp16intrin.h
@@ -2941,11 +2941,8 @@ static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_fcmadd_sch(__m128h __A,
 
 static __inline__ __m128h __DEFAULT_FN_ATTRS128
 _mm_mask_fcmadd_sch(__m128h __A, __mmask8 __U, __m128h __B, __m128h __C) {
-  return (__m128h)__builtin_ia32_selectps_128(
-      __U,
-      __builtin_ia32_vfcmaddcsh_mask((__v4sf)__A, (__v4sf)__B, (__v4sf)__C,
-                                     (__mmask8)__U, _MM_FROUND_CUR_DIRECTION),
-      (__v4sf)__A);
+  return (__m128h)__builtin_ia32_vfcmaddcsh_round_mask(
+      (__v4sf)__A, (__v4sf)(__B), (__v4sf)(__C), __U, _MM_FROUND_CUR_DIRECTION);
 }
 
 static __inline__ __m128h __DEFAULT_FN_ATTRS128
@@ -2957,10 +2954,8 @@ _mm_maskz_fcmadd_sch(__mmask8 __U, __m128h __A, __m128h __B, __m128h __C) {
 
 static __inline__ __m128h __DEFAULT_FN_ATTRS128
 _mm_mask3_fcmadd_sch(__m128h __A, __m128h __B, __m128h __C, __mmask8 __U) {
-  return (__m128h)_mm_move_ss((__m128)__C,
-                              (__m128)__builtin_ia32_vfcmaddcsh_mask(
-                                  (__v4sf)__A, (__v4sf)__B, (__v4sf)__C, __U,
-                                  _MM_FROUND_CUR_DIRECTION));
+  return (__m128h)__builtin_ia32_vfcmaddcsh_round_mask3(
+      (__v4sf)__A, (__v4sf)__B, (__v4sf)__C, __U, _MM_FROUND_CUR_DIRECTION);
 }
 
 #define _mm_fcmadd_round_sch(A, B, C, R)                                       \
@@ -2969,22 +2964,19 @@ _mm_mask3_fcmadd_sch(__m128h __A, __m128h __B, __m128h __C, __mmask8 __U) {
       (__mmask8)-1, (int)(R)))
 
 #define _mm_mask_fcmadd_round_sch(A, U, B, C, R)                               \
-  ((__m128h)__builtin_ia32_selectps_128(                                       \
-      (__mmask8)(U & 1),                                                       \
-      __builtin_ia32_vfcmaddcsh_mask(                                          \
-          (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B), (__v4sf)(__m128h)(C),    \
-          (__mmask8)(U), (int)(R)),                                            \
-      (__v4sf)(__m128h)(A)))
+  ((__m128h)__builtin_ia32_vfcmaddcsh_round_mask(                              \
+      (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B), (__v4sf)(__m128h)(C),        \
+      (__mmask8)(U), (int)(R)))
 
 #define _mm_maskz_fcmadd_round_sch(U, A, B, C, R)                              \
   ((__m128h)__builtin_ia32_vfcmaddcsh_maskz(                                   \
       (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B), (__v4sf)(__m128h)(C),        \
       (__mmask8)(U), (int)(R)))
 
 #define _mm_mask3_fcmadd_round_sch(A, B, C, U, R)                              \
-  ((__m128h)_mm_move_ss((__m128)(C),                                           \
-                        (__m128)__builtin_ia32_vfcmaddcsh_mask(                \
-                            (__v4sf)(A), (__v4sf)(B), (__v4sf)(C), (U), (R))))
+  ((__m128h)__builtin_ia32_vfcmaddcsh_round_mask3(                             \
+      (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B), (__v4sf)(__m128h)(C),        \
+      (__mmask8)(U), (int)(R)))
 
 static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_fmadd_sch(__m128h __A,
                                                               __m128h __B,
@@ -2996,11 +2988,8 @@ static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_fmadd_sch(__m128h __A,
 
 static __inline__ __m128h __DEFAULT_FN_ATTRS128
 _mm_mask_fmadd_sch(__m128h __A, __mmask8 __U, __m128h __B, __m128h __C) {
-  return (__m128h)__builtin_ia32_selectps_128(
-      __U,
-      __builtin_ia32_vfmaddcsh_mask((__v4sf)__A, (__v4sf)__B, (__v4sf)__C,
-                                    (__mmask8)__U, _MM_FROUND_CUR_DIRECTION),
-      (__v4sf)__A);
+  return (__m128h)__builtin_ia32_vfmaddcsh_round_mask(
+      (__v4sf)__A, (__v4sf)(__B), (__v4sf)(__C), __U, _MM_FROUND_CUR_DIRECTION);
 }
 
 static __inline__ __m128h __DEFAULT_FN_ATTRS128
@@ -3010,24 +2999,32 @@ _mm_maskz_fmadd_sch(__mmask8 __U, __m128h __A, __m128h __B, __m128h __C) {
                                                  _MM_FROUND_CUR_DIRECTION);
 }
 
+static __inline__ __m128h __DEFAULT_FN_ATTRS128
+_mm_mask3_fmadd_sch(__m128h __A, __m128h __B, __m128h __C, __mmask8 __U) {
+  return (__m128h)__builtin_ia32_vfmaddcsh_round_mask3(
+      (__v4sf)__A, (__v4sf)__B, (__v4sf)__C, __U, _MM_FROUND_CUR_DIRECTION);
+}
+
 #define _mm_fmadd_round_sch(A, B, C, R)                                        \
   ((__m128h)__builtin_ia32_vfmaddcsh_mask(                                     \
       (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B), (__v4sf)(__m128h)(C),        \
       (__mmask8)-1, (int)(R)))
 
 #define _mm_mask_fmadd_round_sch(A, U, B, C, R)                                \
-  ((__m128h)__builtin_ia32_selectps_128(                                       \
-      (__mmask8)(U & 1),                                                       \
-      __builtin_ia32_vfmaddcsh_mask(                                           \
-          (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B), (__v4sf)(__m128h)(C),    \
-          (__mmask8)(U), (int)(R)),                                            \
-      (__v4sf)(__m128h)(A)))
+  ((__m128h)__builtin_ia32_vfmaddcsh_round_mask(                               \
+      (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B), (__v4sf)(__m128h)(C),        \
+      (__mmask8)(U), (int)(R)))
 
 #define _mm_maskz_fmadd_round_sch(U, A, B, C, R)                               \
   ((__m128h)__builtin_ia32_vfmaddcsh_maskz(                                    \
       (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B), (__v4sf)(__m128h)(C),        \
       (__mmask8)(U), (int)(R)))
 
+#define _mm_mask3_fmadd_round_sch(A, B, C, U, R)                               \
+  ((__m128h)__builtin_ia32_vfmaddcsh_round_mask3(                              \
+      (__v4sf)(__m128h)(A), (__v4sf)(__m128h)(B), (__v4sf)(__m128h)(C),        \
+      (__mmask8)(U), (int)(R)))
+
 static __inline__ __m128h __DEFAULT_FN_ATTRS128 _mm_fcmul_sch(__m128h __A,
                                                               __m128h __B) {
   return (__m128h)__builtin_ia32_vfcmulcsh_mask(
@@ -3177,24 +3174,21 @@ _mm512_maskz_fmul_pch(__mmask16 __U, __m512h __A, __m512h __B) {
 static __inline__ __m512h __DEFAULT_FN_ATTRS512 _mm512_fcmadd_pch(__m512h __A,
                                                                   __m512h __B,
                                                                   __m512h __C) {
-  return (__m512h)__builtin_ia32_vfcmaddcph512_mask((__v16sf)__A, (__v16sf)__B,
-                                                    (__v16sf)__C, (__mmask16)-1,
-                                                    _MM_FROUND_CUR_DIRECTION);
+  return (__m512h)__builtin_ia32_vfcmaddcph512_mask3(
+      (__v16sf)__A, (__v16sf)__B, (__v16sf)__C, (__mmask16)-1,
+      _MM_FROUND_CUR_DIRECTION);
 }
 
 static __inline__ __m512h __DEFAULT_FN_ATTRS512
 _mm512_mask_fcmadd_pch(__m512h __A, __mmask16 __U, __m512h __B, __m512h __C) {
-  return (__m512h)__builtin_ia32_selectps_512(
-      __U,
-      __builtin_ia32_vfcmaddcph512_mask((__v16sf)__A, (__v16sf)__B,
-                                        (__v16sf)__C, (__mmask16)__U,
-                                        _MM_FROUND_CUR_DIRECTION),
-      (__v16sf)__A);
+  return (__m512h)__builtin_ia32_vfcmaddcph512_mask(
+      (__v16sf)__A, (__v16sf)__B, (__v16sf)__C, (__mmask16)__U,
+      _MM_FROUND_CUR_DIRECTION);
 }
 
 static __inline__ __m512h __DEFAULT_FN_ATTRS512
 _mm512_mask3_fcmadd_pch(__m512h __A, __m512h __B, __m512h __C, __mmask16 __U) {
-  return (__m512h)__builtin_ia32_vfcmaddcph512_mask(
+  return (__m512h)__builtin_ia32_vfcmaddcph512_mask3(
       (__v16sf)__A, (__v16sf)__B, (__v16sf)__C, (__mmask16)__U,
       _MM_FROUND_CUR_DIRECTION);
 }
@@ -3207,20 +3201,17 @@ _mm512_maskz_fcmadd_pch(__mmask16 __U, __m512h __A, __m512h __B, __m512h __C) {
 }
 
 #define _mm512_fcmadd_round_pch(A, B, C, R)                                    \
-  ((__m512h)__builtin_ia32_vfcmaddcph512_mask(                                 \
+  ((__m512h)__builtin_ia32_vfcmaddcph512_mask3(                                \
       (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B), (__v16sf)(__m512h)(C),     \
       (__mmask16)-1, (int)(R)))
 
 #define _mm512_mask_fcmadd_round_pch(A, U, B, C, R)                            \
-  ((__m512h)__builtin_ia32_selectps_512(                                       \
-      (__mmask16)(U),                                                          \
-      __builtin_ia32_vfcmaddcph512_mask(                                       \
-          (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B), (__v16sf)(__m512h)(C), \
-          (__mmask16)(U), (int)(R)),                                           \
-      (__v16sf)(__m512h)(A)))
+  ((__m512h)__builtin_ia32_vfcmaddcph512_mask(                                 \
+      (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B), (__v16sf)(__m512h)(C),     \
+      (__mmask16)(U), (int)(R)))
 
 #define _mm512_mask3_fcmadd_round_pch(A, B, C, U, R)                           \
-  ((__m512h)__builtin_ia32_vfcmaddcph512_mask(                                 \
+  ((__m512h)__builtin_ia32_vfcmaddcph512_mask3(                                \
       (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B), (__v16sf)(__m512h)(C),     \
       (__mmask16)(U), (int)(R)))
 
@@ -3232,26 +3223,23 @@ _mm512_maskz_fcmadd_pch(__mmask16 __U, __m512h __A, __m512h __B, __m512h __C) {
 static __inline__ __m512h __DEFAULT_FN_ATTRS512 _mm512_fmadd_pch(__m512h __A,
                                                                  __m512h __B,
                                                                  __m512h __C) {
-  return (__m512h)__builtin_ia32_vfmaddcph512_mask((__v16sf)__A, (__v16sf)__B,
-                                                   (__v16sf)__C, (__mmask16)-1,
-                                                   _MM_FROUND_CUR_DIRECTION);
+  return (__m512h)__builtin_ia32_vfmaddcph512_mask3((__v16sf)__A, (__v16sf)__B,
+                                                    (__v16sf)__C, (__mmask16)-1,
+                                                    _MM_FROUND_CUR_DIRECTION);
 }
 
 static __inline__ __m512h __DEFAULT_FN_ATTRS512
 _mm512_mask_fmadd_pch(__m512h __A, __mmask16 __U, __m512h __B, __m512h __C) {
-  return (__m512h)__builtin_ia32_selectps_512(
-      __U,
-      __builtin_ia32_vfmaddcph512_mask((__v16sf)__A, (__v16sf)__B, (__v16sf)__C,
-                                       (__mmask16)__U,
-                                       _MM_FROUND_CUR_DIRECTION),
-      (__v16sf)__A);
+  return (__m512h)__builtin_ia32_vfmaddcph512_mask((__v16sf)__A, (__v16sf)__B,
+                                                   (__v16sf)__C, (__mmask16)__U,
+                                                   _MM_FROUND_CUR_DIRECTION);
 }
 
 static __inline__ __m512h __DEFAULT_FN_ATTRS512
 _mm512_mask3_fmadd_pch(__m512h __A, __m512h __B, __m512h __C, __mmask16 __U) {
-  return (__m512h)__builtin_ia32_vfmaddcph512_mask((__v16sf)__A, (__v16sf)__B,
-                                                   (__v16sf)__C, (__mmask16)__U,
-                                                   _MM_FROUND_CUR_DIRECTION);
+  return (__m512h)__builtin_ia32_vfmaddcph512_mask3(
+      (__v16sf)__A, (__v16sf)__B, (__v16sf)__C, (__mmask16)__U,
+      _MM_FROUND_CUR_DIRECTION);
 }
 
 static __inline__ __m512h __DEFAULT_FN_ATTRS512
@@ -3262,20 +3250,17 @@ _mm512_maskz_fmadd_pch(__mmask16 __U, __m512h __A, __m512h __B, __m512h __C) {
 }
 
 #define _mm512_fmadd_round_pch(A, B, C, R)                                     \
-  ((__m512h)__builtin_ia32_vfmaddcph512_mask(                                  \
+  ((__m512h)__builtin_ia32_vfmaddcph512_mask3(                                 \
       (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B), (__v16sf)(__m512h)(C),     \
       (__mmask16)-1, (int)(R)))
 
 #define _mm512_mask_fmadd_round_pch(A, U, B, C, R)                             \
-  ((__m512h)__builtin_ia32_selectps_512(                                       \
-      (__mmask16)(U),                                                          \
-      __builtin_ia32_vfmaddcph512_mask(                                        \
-          (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B), (__v16sf)(__m512h)(C), \
-          (__mmask16)(U), (int)(R)),                                           \
-      (__v16sf)(__m512h)(A)))
+  ((__m512h)__builtin_ia32_vfmaddcph512_mask(                                  \
+      (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B), (__v16sf)(__m512h)(C),     \
+      (__mmask16)(U), (int)(R)))
 
 #define _mm512_mask3_fmadd_round_pch(A, B, C, U, R)                            \
-  ((__m512h)__builtin_ia32_vfmaddcph512_mask(                                  \
+  ((__m512h)__builtin_ia32_vfmaddcph512_mask3(                                 \
       (__v16sf)(__m512h)(A), (__v16sf)(__m512h)(B), (__v16sf)(__m512h)(C),     \
       (__mmask16)(U), (int)(R)))
 

diff --git a/clang/lib/Lex/LiteralSupport.cpp b/clang/lib/Lex/LiteralSupport.cpp
@@ -1654,9 +1654,9 @@ CharLiteralParser::CharLiteralParser(const char *begin, const char *end,
 ///
 StringLiteralParser::
 StringLiteralParser(ArrayRef<Token> StringToks,
-                    Preprocessor &PP, bool Complain)
+                    Preprocessor &PP)
   : SM(PP.getSourceManager()), Features(PP.getLangOpts()),
-    Target(PP.getTargetInfo()), Diags(Complain ? &PP.getDiagnostics() :nullptr),
+    Target(PP.getTargetInfo()), Diags(&PP.getDiagnostics()),
     MaxTokenLength(0), SizeBound(0), CharByteWidth(0), Kind(tok::unknown),
     ResultPtr(ResultBuf.data()), hadError(false), Pascal(false) {
   init(StringToks);

diff --git a/clang/lib/Lex/PPDirectives.cpp b/clang/lib/Lex/PPDirectives.cpp
@@ -2865,6 +2865,12 @@ void Preprocessor::HandleDefineDirective(
   if (MacroNameTok.is(tok::eod))
     return;
 
+  IdentifierInfo *II = MacroNameTok.getIdentifierInfo();
+  // Issue a final pragma warning if we're defining a macro that was has been
+  // undefined and is being redefined.
+  if (!II->hasMacroDefinition() && II->hadMacroDefinition() && II->isFinal())
+    emitFinalMacroWarning(MacroNameTok, /*IsUndef=*/false);
+
   // If we are supposed to keep comments in #defines, reenable comment saving
   // mode.
   if (CurLexer) CurLexer->SetCommentRetentionState(KeepMacroComments);
@@ -2907,6 +2913,12 @@ void Preprocessor::HandleDefineDirective(
   // Finally, if this identifier already had a macro defined for it, verify that
   // the macro bodies are identical, and issue diagnostics if they are not.
   if (const MacroInfo *OtherMI=getMacroInfo(MacroNameTok.getIdentifierInfo())) {
+    // Final macros are hard-mode: they always warn. Even if the bodies are
+    // identical. Even if they are in system headers. Even if they are things we
+    // would silently allow in the past.
+    if (MacroNameTok.getIdentifierInfo()->isFinal())
+      emitFinalMacroWarning(MacroNameTok, /*IsUndef=*/false);
+
     // In Objective-C, ignore attempts to directly redefine the builtin
     // definitions of the ownership qualifiers.  It's still possible to
     // #undef them.
@@ -2936,6 +2948,7 @@ void Preprocessor::HandleDefineDirective(
     // then don't bother calling MacroInfo::isIdenticalTo.
     if (!getDiagnostics().getSuppressSystemWarnings() ||
         !SourceMgr.isInSystemHeader(DefineTok.getLocation())) {
+
       if (!OtherMI->isUsed() && OtherMI->isWarnIfUnused())
         Diag(OtherMI->getDefinitionLoc(), diag::pp_macro_not_used);
 
@@ -3013,6 +3026,9 @@ void Preprocessor::HandleUndefDirective() {
   auto MD = getMacroDefinition(II);
   UndefMacroDirective *Undef = nullptr;
 
+  if (II->isFinal())
+    emitFinalMacroWarning(MacroNameTok, /*IsUndef=*/true);
+
   // If the macro is not defined, this is a noop undef.
   if (const MacroInfo *MI = MD.getMacroInfo()) {
     if (!MI->isUsed() && MI->isWarnIfUnused())

diff --git a/clang/lib/Lex/Pragma.cpp b/clang/lib/Lex/Pragma.cpp
@@ -1990,7 +1990,7 @@ static IdentifierInfo *HandleMacroAnnotationPragma(Preprocessor &PP, Token &Tok,
   IdentifierInfo *II = Tok.getIdentifierInfo();
 
   if (!II->hasMacroDefinition()) {
-    PP.Diag(Tok, diag::err_pp_visibility_non_macro) << II->getName();
+    PP.Diag(Tok, diag::err_pp_visibility_non_macro) << II;
     return nullptr;
   }
 
@@ -2025,8 +2025,8 @@ struct PragmaDeprecatedHandler : public PragmaHandler {
     if (IdentifierInfo *II = HandleMacroAnnotationPragma(
             PP, Tok, "#pragma clang deprecated", MessageString)) {
       II->setIsDeprecatedMacro(true);
-      if (!MessageString.empty())
-        PP.addMacroDeprecationMsg(II, std::move(MessageString));
+      PP.addMacroDeprecationMsg(II, std::move(MessageString),
+                                Tok.getLocation());
     }
   }
 };
@@ -2053,6 +2053,47 @@ struct PragmaRestrictExpansionHandler : public PragmaHandler {
   }
 };
 
+/// "\#pragma clang final(...)"
+///
+/// The syntax is
+/// \code
+///   #pragma clang final(MACRO_NAME)
+/// \endcode
+struct PragmaFinalHandler : public PragmaHandler {
+  PragmaFinalHandler() : PragmaHandler("final") {}
+
+  void HandlePragma(Preprocessor &PP, PragmaIntroducer Introducer,
+                    Token &Tok) override {
+    std::string Macro;
+
+    PP.Lex(Tok);
+    if (Tok.isNot(tok::l_paren)) {
+      PP.Diag(Tok, diag::err_expected) << "(";
+      return;
+    }
+
+    PP.LexUnexpandedToken(Tok);
+    if (!Tok.is(tok::identifier)) {
+      PP.Diag(Tok, diag::err_expected) << tok::identifier;
+      return;
+    }
+    IdentifierInfo *II = Tok.getIdentifierInfo();
+
+    if (!II->hasMacroDefinition()) {
+      PP.Diag(Tok, diag::err_pp_visibility_non_macro) << II;
+      return;
+    }
+
+    PP.Lex(Tok);
+    if (Tok.isNot(tok::r_paren)) {
+      PP.Diag(Tok, diag::err_expected) << ")";
+      return;
+    }
+    II->setIsFinal(true);
+    PP.addFinalLoc(II, Tok.getLocation());
+  }
+};
+
 } // namespace
 
 /// RegisterBuiltinPragmas - Install the standard preprocessor pragmas:
@@ -2084,6 +2125,7 @@ void Preprocessor::RegisterBuiltinPragmas() {
   AddPragmaHandler("clang", new PragmaAssumeNonNullHandler());
   AddPragmaHandler("clang", new PragmaDeprecatedHandler());
   AddPragmaHandler("clang", new PragmaRestrictExpansionHandler());
+  AddPragmaHandler("clang", new PragmaFinalHandler());
 
   // #pragma clang module ...
   auto *ModuleHandler = new PragmaNamespace("module");

diff --git a/clang/lib/Lex/Preprocessor.cpp b/clang/lib/Lex/Preprocessor.cpp
@@ -1409,25 +1409,46 @@ bool Preprocessor::HandleComment(Token &result, SourceRange Comment) {
   return true;
 }
 
-void Preprocessor::emitMacroDeprecationWarning(const Token &Identifier) {
-  auto DepMsg = getMacroDeprecationMsg(Identifier.getIdentifierInfo());
-  if (!DepMsg)
+void Preprocessor::emitMacroDeprecationWarning(const Token &Identifier) const {
+  const MacroAnnotations &A =
+      getMacroAnnotations(Identifier.getIdentifierInfo());
+  assert(A.DeprecationInfo &&
+         "Macro deprecation warning without recorded annotation!");
+  const MacroAnnotationInfo &Info = *A.DeprecationInfo;
+  if (Info.Message.empty())
     Diag(Identifier, diag::warn_pragma_deprecated_macro_use)
         << Identifier.getIdentifierInfo() << 0;
   else
     Diag(Identifier, diag::warn_pragma_deprecated_macro_use)
-        << Identifier.getIdentifierInfo() << 1 << *DepMsg;
+        << Identifier.getIdentifierInfo() << 1 << Info.Message;
+  Diag(Info.Location, diag::note_pp_macro_annotation) << 0;
 }
 
-void Preprocessor::emitMacroUnsafeHeaderWarning(const Token &Identifier) {
-  auto DepMsg = getRestrictExpansionMsg(Identifier.getIdentifierInfo());
-  if (DepMsg.first.empty())
+void Preprocessor::emitRestrictExpansionWarning(const Token &Identifier) const {
+  const MacroAnnotations &A =
+      getMacroAnnotations(Identifier.getIdentifierInfo());
+  assert(A.RestrictExpansionInfo &&
+         "Macro restricted expansion warning without recorded annotation!");
+  const MacroAnnotationInfo &Info = *A.RestrictExpansionInfo;
+  if (Info.Message.empty())
     Diag(Identifier, diag::warn_pragma_restrict_expansion_macro_use)
         << Identifier.getIdentifierInfo() << 0;
   else
     Diag(Identifier, diag::warn_pragma_restrict_expansion_macro_use)
-        << Identifier.getIdentifierInfo() << 1 << DepMsg.first;
-  Diag(DepMsg.second, diag::note_pp_macro_annotation) << 1;
+        << Identifier.getIdentifierInfo() << 1 << Info.Message;
+  Diag(Info.Location, diag::note_pp_macro_annotation) << 1;
+}
+
+void Preprocessor::emitFinalMacroWarning(const Token &Identifier,
+                                         bool IsUndef) const {
+  const MacroAnnotations &A =
+      getMacroAnnotations(Identifier.getIdentifierInfo());
+  assert(A.FinalAnnotationLoc &&
+         "Final macro warning without recorded annotation!");
+
+  Diag(Identifier, diag::warn_pragma_final_macro)
+      << Identifier.getIdentifierInfo() << (IsUndef ? 0 : 1);
+  Diag(*A.FinalAnnotationLoc, diag::note_pp_macro_annotation) << 2;
 }
 
 ModuleLoader::~ModuleLoader() = default;

diff --git a/clang/lib/Sema/SemaChecking.cpp b/clang/lib/Sema/SemaChecking.cpp
@@ -4129,11 +4129,17 @@ bool Sema::CheckX86BuiltinRoundingOrSAE(unsigned BuiltinID, CallExpr *TheCall) {
   case X86::BI__builtin_ia32_vfmaddsubph512_mask3:
   case X86::BI__builtin_ia32_vfmsubaddph512_mask3:
   case X86::BI__builtin_ia32_vfmaddcsh_mask:
+  case X86::BI__builtin_ia32_vfmaddcsh_round_mask:
+  case X86::BI__builtin_ia32_vfmaddcsh_round_mask3:
   case X86::BI__builtin_ia32_vfmaddcph512_mask:
   case X86::BI__builtin_ia32_vfmaddcph512_maskz:
+  case X86::BI__builtin_ia32_vfmaddcph512_mask3:
   case X86::BI__builtin_ia32_vfcmaddcsh_mask:
+  case X86::BI__builtin_ia32_vfcmaddcsh_round_mask:
+  case X86::BI__builtin_ia32_vfcmaddcsh_round_mask3:
   case X86::BI__builtin_ia32_vfcmaddcph512_mask:
   case X86::BI__builtin_ia32_vfcmaddcph512_maskz:
+  case X86::BI__builtin_ia32_vfcmaddcph512_mask3:
   case X86::BI__builtin_ia32_vfmulcsh_mask:
   case X86::BI__builtin_ia32_vfmulcph512_mask:
   case X86::BI__builtin_ia32_vfcmulcsh_mask:

diff --git a/clang/test/CodeGen/X86/avx512fp16-builtins.c b/clang/test/CodeGen/X86/avx512fp16-builtins.c
@@ -4086,10 +4086,8 @@ __m128h test_mm_mask3_fcmadd_sch(__m128h __A, __m128h __B, __m128h __C, __mmask8
   // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
   // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
   // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
-  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
   // CHECK:  %{{.*}} = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.csh(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}, i8 %{{.*}}, i32 4)
-  // CHECK:  %{{.*}} = extractelement <4 x float> %{{.*}}, i32 0
-  // CHECK:  %{{.*}} = insertelement <4 x float> %{{.*}}, float %{{.*}}, i32 0
+  // CHECK:  %{{.*}} = shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x i32> <i32 0, i32 5, i32 6, i32 7>
   // CHECK:  %{{.*}} = bitcast <4 x float> %{{.*}} to <8 x half>
   return _mm_mask3_fcmadd_sch(__A, __B, __C, __U);
 }
@@ -4119,10 +4117,8 @@ __m128h test_mm_mask3_fcmadd_round_sch(__m128h __A, __m128h __B, __m128h __C, __
   // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
   // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
   // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
-  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
   // CHECK:  %{{.*}} = call <4 x float> @llvm.x86.avx512fp16.mask.vfcmadd.csh(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}, i8 %{{.*}}, i32 11)
-  // CHECK:  %{{.*}} = extractelement <4 x float> %{{.*}}, i32 0
-  // CHECK:  %{{.*}} = insertelement <4 x float> %{{.*}}, float %{{.*}}, i32 0
+  // CHECK:  %{{.*}} = shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x i32> <i32 0, i32 5, i32 6, i32 7>
   // CHECK:  %{{.*}} = bitcast <4 x float> %{{.*}} to <8 x half>
   return _mm_mask3_fcmadd_round_sch(__A, __B, __C, __U, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
 }
@@ -4147,6 +4143,17 @@ __m128h test_mm_maskz_fmadd_sch(__mmask8 __U, __m128h __A, __m128h __B, __m128h
   return _mm_maskz_fmadd_sch(__U, __A, __B, __C);
 }
 
+__m128h test_mm_mask3_fmadd_sch(__m128h __A, __m128h __B, __m128h __C, __mmask8 __U) {
+  // CHECK-LABEL: @test_mm_mask3_fmadd_sch
+  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = call <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.csh(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}, i8 %{{.*}}, i32 4)
+  // CHECK:  %{{.*}} = shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x i32> <i32 0, i32 5, i32 6, i32 7>
+  // CHECK:  %{{.*}} = bitcast <4 x float> %{{.*}} to <8 x half>
+  return _mm_mask3_fmadd_sch(__A, __B, __C, __U);
+}
+
 __m128h test_mm_fmadd_round_sch(__m128h __A, __m128h __B, __m128h __C) {
   // CHECK-LABEL: @test_mm_fmadd_round_sch
   // CHECK: @llvm.x86.avx512fp16.mask.vfmadd.csh
@@ -4167,6 +4174,17 @@ __m128h test_mm_maskz_fmadd_round_sch(__mmask8 __U, __m128h __A, __m128h __B, __
   return _mm_maskz_fmadd_round_sch(__U, __A, __B, __C, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
 }
 
+__m128h test_mm_mask3_fmadd_round_sch(__m128h __A, __m128h __B, __m128h __C, __mmask8 __U) {
+  // CHECK-LABEL: @test_mm_mask3_fmadd_round_sch
+  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = bitcast <8 x half> %{{.*}} to <4 x float>
+  // CHECK:  %{{.*}} = call <4 x float> @llvm.x86.avx512fp16.mask.vfmadd.csh(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}, i8 %{{.*}}, i32 11)
+  // CHECK:  %{{.*}} = shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x i32> <i32 0, i32 5, i32 6, i32 7>
+  // CHECK:  %{{.*}} = bitcast <4 x float> %{{.*}} to <8 x half>
+  return _mm_mask3_fmadd_round_sch(__A, __B, __C, __U, _MM_FROUND_TO_ZERO | _MM_FROUND_NO_EXC);
+}
+
 __m128h test_mm_fcmul_sch(__m128h __A, __m128h __B) {
   // CHECK-LABEL: @test_mm_fcmul_sch
   // CHECK: @llvm.x86.avx512fp16.mask.vfcmul.csh

diff --git a/clang/test/CodeGen/thinlto-distributed-cfi-devirt.ll b/clang/test/CodeGen/thinlto-distributed-cfi-devirt.ll
@@ -7,7 +7,7 @@
 ; RUN: opt -thinlto-bc -thinlto-split-lto-unit -o %t.o %s
 
 ; FIXME: Fix machine verifier issues and remove -verify-machineinstrs=0. PR39436.
-; RUN: llvm-lto2 run -thinlto-distributed-indexes %t.o \
+; RUN: llvm-lto2 run -thinlto-distributed-indexes -disable-thinlto-funcattrs=0 %t.o \
 ; RUN:   -whole-program-visibility \
 ; RUN:   -verify-machineinstrs=0 \
 ; RUN:   -o %t2.index \
@@ -36,7 +36,7 @@
 ; Round trip it through llvm-as
 ; RUN: llvm-dis %t.o.thinlto.bc -o - | llvm-as -o - | llvm-dis -o - | FileCheck %s --check-prefix=CHECK-DIS
 ; CHECK-DIS: ^0 = module: (path: "{{.*}}thinlto-distributed-cfi-devirt.ll.tmp.o", hash: ({{.*}}, {{.*}}, {{.*}}, {{.*}}, {{.*}}))
-; CHECK-DIS: ^1 = gv: (guid: 8346051122425466633, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 1, dsoLocal: 0, canAutoHide: 0), insts: 18, typeIdInfo: (typeTests: (^2), typeCheckedLoadVCalls: (vFuncId: (^2, offset: 8), vFuncId: (^2, offset: 0))))))
+; CHECK-DIS: ^1 = gv: (guid: 8346051122425466633, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 1, dsoLocal: 0, canAutoHide: 0), insts: 18, funcFlags: (readNone: 0, readOnly: 0, noRecurse: 0, returnDoesNotAlias: 0, noInline: 0, alwaysInline: 0, noUnwind: 0, mayThrow: 0, hasUnknownCall: 1), typeIdInfo: (typeTests: (^2), typeCheckedLoadVCalls: (vFuncId: (^2, offset: 8), vFuncId: (^2, offset: 0))))))
 ; CHECK-DIS: ^2 = typeid: (name: "_ZTS1A", summary: (typeTestRes: (kind: allOnes, sizeM1BitWidth: 7), wpdResolutions: ((offset: 0, wpdRes: (kind: branchFunnel)), (offset: 8, wpdRes: (kind: singleImpl, singleImplName: "_ZN1A1nEi"))))) ; guid = 7004155349499253778
 
 ; RUN: %clang_cc1 -triple x86_64-grtev4-linux-gnu \

diff --git a/clang/test/CodeGen/thinlto-distributed-cfi.ll b/clang/test/CodeGen/thinlto-distributed-cfi.ll
@@ -4,7 +4,7 @@
 
 ; RUN: opt -thinlto-bc -thinlto-split-lto-unit -o %t.o %s
 
-; RUN: llvm-lto2 run -thinlto-distributed-indexes %t.o \
+; RUN: llvm-lto2 run -thinlto-distributed-indexes -disable-thinlto-funcattrs=0 %t.o \
 ; RUN:   -o %t2.index \
 ; RUN:   -r=%t.o,test,px \
 ; RUN:   -r=%t.o,_ZTV1B, \
@@ -24,7 +24,7 @@
 ; Round trip it through llvm-as
 ; RUN: llvm-dis %t.o.thinlto.bc -o - | llvm-as -o - | llvm-dis -o - | FileCheck %s --check-prefix=CHECK-DIS
 ; CHECK-DIS: ^0 = module: (path: "{{.*}}thinlto-distributed-cfi.ll.tmp.o", hash: ({{.*}}, {{.*}}, {{.*}}, {{.*}}, {{.*}}))
-; CHECK-DIS: ^1 = gv: (guid: 8346051122425466633, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 1, dsoLocal: 0, canAutoHide: 0), insts: 7, typeIdInfo: (typeTests: (^2)))))
+; CHECK-DIS: ^1 = gv: (guid: 8346051122425466633, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 1, dsoLocal: 0, canAutoHide: 0), insts: 7, funcFlags: (readNone: 0, readOnly: 0, noRecurse: 1, returnDoesNotAlias: 0, noInline: 0, alwaysInline: 0, noUnwind: 1, mayThrow: 0, hasUnknownCall: 0), typeIdInfo: (typeTests: (^2)))))
 ; CHECK-DIS: ^2 = typeid: (name: "_ZTS1A", summary: (typeTestRes: (kind: single, sizeM1BitWidth: 0))) ; guid = 7004155349499253778
 
 ; RUN: %clang_cc1 -triple x86_64-grtev4-linux-gnu \

diff --git a/clang/test/CodeGen/thinlto-funcattr-prop.ll b/clang/test/CodeGen/thinlto-funcattr-prop.ll
@@ -0,0 +1,43 @@
+; REQUIRES: x86-registered-target
+
+; Test that FunctionAttr Propagation is generating correct summaries
+
+; RUN: split-file %s %t
+; RUN: opt -module-summary %t/a.ll -o %t/a.bc
+; RUN: opt -module-summary %t/b.ll -o %t/b.bc
+
+; RUN: llvm-lto2 run -disable-thinlto-funcattrs=0 %t/a.bc %t/b.bc -o %t1.o -save-temps \
+; RUN:   -r=%t/a.bc,call_extern,plx \
+; RUN:   -r=%t/a.bc,extern, \
+; RUN:   -r=%t/b.bc,extern,p
+
+; RUN: llvm-dis %t1.o.index.bc -o - | FileCheck %s --check-prefix=CHECK-INDEX
+; RUN: llvm-dis %t1.o.1.1.promote.bc -o - | FileCheck %s --check-prefix=CHECK-IR
+
+;; Summary for call_extern. Note that llvm-lto2 writes out the index before propagation occurs so call_extern doesn't have its flags updated.
+; CHECK-INDEX: ^2 = gv: (guid: 13959900437860518209, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 1, dsoLocal: 1, canAutoHide: 0), insts: 2, calls: ((callee: ^3)))))
+;; Summary for extern
+; CHECK-INDEX: ^3 = gv: (guid: 14959766916849974397, summaries: (function: (module: ^1, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 1, dsoLocal: 0, canAutoHide: 0), insts: 1, funcFlags: (readNone: 0, readOnly: 0, noRecurse: 1, returnDoesNotAlias: 0, noInline: 0, alwaysInline: 0, noUnwind: 1, mayThrow: 0, hasUnknownCall: 0))))
+
+;--- a.ll
+target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-unknown-linux-gnu"
+
+declare void @extern()
+
+; CHECK-IR: Function Attrs: norecurse nounwind
+; CHECK-IR-NEXT: define dso_local void @call_extern()
+define void @call_extern() {
+    call void @extern()
+    ret void
+}
+
+;--- b.ll
+target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-unknown-linux-gnu"
+
+attributes #0 = { nounwind norecurse }
+
+define void @extern() #0 {
+  ret void
+}
diff --git a/clang/test/CodeGenCUDA/Inputs/amdgpu-asanrtl.ll b/clang/test/CodeGenCUDA/Inputs/amdgpu-asanrtl.ll
@@ -0,0 +1,13 @@
+; Sample code for amdgpu address sanitizer runtime.
+
+; Note the runtime functions need to have weak linkage and default
+; visibility, otherwise they may be internalized and removed by GlobalOptPass.
+
+define weak void @__amdgpu_device_library_preserve_asan_functions() {
+  tail call void @__asan_report_load1(i64 0)
+  ret void
+}
+
+define weak void @__asan_report_load1(i64 %0) {
+  ret void
+}
diff --git a/clang/test/CodeGenCUDA/amdgpu-asan.cu b/clang/test/CodeGenCUDA/amdgpu-asan.cu
@@ -1,15 +1,31 @@
+// Create a sample address sanitizer bitcode library.
+
+// RUN: %clang_cc1 -x ir -fcuda-is-device -triple amdgcn-amd-amdhsa -emit-llvm-bc \
+// RUN:   -disable-llvm-passes -o %t.asanrtl.bc %S/Inputs/amdgpu-asanrtl.ll
+
+// Check sanitizer runtime library functions survive
+// optimizations without being removed or parameters altered.
+
+// RUN: %clang_cc1 %s -emit-llvm -o - -triple=amdgcn-amd-amdhsa \
+// RUN:   -fcuda-is-device -target-cpu gfx906 -fsanitize=address \
+// RUN:   -mlink-bitcode-file %t.asanrtl.bc -x hip \
+// RUN:   | FileCheck -check-prefix=ASAN %s
+
 // RUN: %clang_cc1 %s -emit-llvm -o - -triple=amdgcn-amd-amdhsa \
 // RUN:   -fcuda-is-device -target-cpu gfx906 -fsanitize=address \
-// RUN:   -x hip | FileCheck -check-prefix=ASAN %s
+// RUN:   -O3 -mlink-bitcode-file %t.asanrtl.bc -x hip \
+// RUN:   | FileCheck -check-prefix=ASAN %s
 
 // RUN: %clang_cc1 %s -emit-llvm -o - -triple=amdgcn-amd-amdhsa \
 // RUN:   -fcuda-is-device -target-cpu gfx906 -x hip \
 // RUN:   | FileCheck %s
 
 // REQUIRES: amdgpu-registered-target
 
-// ASAN-DAG: declare void @__amdgpu_device_library_preserve_asan_functions()
+// ASAN-DAG: define weak void @__amdgpu_device_library_preserve_asan_functions()
 // ASAN-DAG: @__amdgpu_device_library_preserve_asan_functions_ptr = weak addrspace(1) constant void ()* @__amdgpu_device_library_preserve_asan_functions
 // ASAN-DAG: @llvm.compiler.used = {{.*}}@__amdgpu_device_library_preserve_asan_functions_ptr
+// ASAN-DAG: define weak void @__asan_report_load1(i64 %{{.*}})
 
-// CHECK-NOT: @__amdgpu_device_library_preserve_asan_functions_ptr
+// CHECK-NOT: @__amdgpu_device_library_preserve_asan_functions
+// CHECK-NOT: @__asan_report_load1
diff --git a/clang/test/Driver/android-gcc-toolchain.c b/clang/test/Driver/android-gcc-toolchain.c
diff --git a/clang/test/Driver/compilation_database.c b/clang/test/Driver/compilation_database.c
@@ -1,9 +1,9 @@
 // RUN: mkdir -p %t.workdir && cd %t.workdir
-// RUN: %clang -MD -MP --sysroot=somewhere -c -x c %s -xc++ %s -Wall -MJ - -no-canonical-prefixes 2>&1 | FileCheck %s
+// RUN: %clang -fintegrated-as -MD -MP --sysroot=somewhere -c -x c %s -xc++ %s -Wall -MJ - -no-canonical-prefixes 2>&1 | FileCheck %s
 // RUN: not %clang -c -x c %s -MJ %s/non-existant -no-canonical-prefixes 2>&1 | FileCheck --check-prefix=ERROR %s
 
-// CHECK: { "directory": "{{[^"]*}}workdir",  "file": "[[SRC:[^"]+[/|\\]compilation_database.c]]", "output": "compilation_database.o", "arguments": ["{{[^"]*}}clang{{[^"]*}}", "-xc", "[[SRC]]", "--sysroot=somewhere", "-c", "-Wall",{{.*}} "--target={{[^"]+}}"]},
-// CHECK: { "directory": "{{.*}}",  "file": "[[SRC:[^"]+[/|\\]compilation_database.c]]", "output": "compilation_database.o", "arguments": ["{{[^"]*}}clang{{[^"]*}}", "-xc++", "[[SRC]]", "--sysroot=somewhere", "-c", "-Wall",{{.*}} "--target={{[^"]+}}"]},
+// CHECK: { "directory": "{{[^"]*}}workdir",  "file": "[[SRC:[^"]+[/|\\]compilation_database.c]]", "output": "compilation_database.o", "arguments": ["{{[^"]*}}clang{{[^"]*}}", "-xc", "[[SRC]]", "-fintegrated-as", "--sysroot=somewhere", "-c", "-Wall",{{.*}} "--target={{[^"]+}}"]},
+// CHECK: { "directory": "{{.*}}",  "file": "[[SRC:[^"]+[/|\\]compilation_database.c]]", "output": "compilation_database.o", "arguments": ["{{[^"]*}}clang{{[^"]*}}", "-xc++", "[[SRC]]", "-fintegrated-as", "--sysroot=somewhere", "-c", "-Wall",{{.*}} "--target={{[^"]+}}"]},
 // ERROR: error: compilation database '{{.*}}/non-existant' could not be opened:
 
 int main(void) {

diff --git a/clang/test/Driver/crash-report-clang-cl.c b/clang/test/Driver/crash-report-clang-cl.c
@@ -0,0 +1,24 @@
+// RUN: rm -rf %t
+// RUN: mkdir %t
+
+// RUN: env TMPDIR=%t TEMP=%t TMP=%t RC_DEBUG_OPTIONS=1           \
+// RUN: not %clang_cl -fsyntax-only /Brepro /source-charset:utf-8 \
+// RUN:     -- %s 2>&1 | FileCheck %s
+// RUN: cat %t/crash-report-*.sh | FileCheck --check-prefix=CHECKSH %s
+
+// REQUIRES: crash-recovery
+
+#pragma clang __debug crash
+
+// CHECK: Preprocessed source(s) and associated run script(s) are located at:
+// CHECK-NEXT: note: diagnostic msg: {{.*}}crash-report-clang-cl-{{.*}}.c
+// CHECKSH: # Crash reproducer
+// CHECKSH-NEXT: # Driver args: {{.*}}"-fsyntax-only"
+// CHECKSH-SAME: /Brepro
+// CHECKSH-SAME: /source-charset:utf-8
+// CHECKSH-NOT: -mno-incremental-linker-compatible
+// CHECKSH-NOT: -finput-charset=utf-8
+// CHECKSH-NEXT: # Original command: {{.*$}}
+// CHECKSH-NEXT: "-cc1"
+// CHECKSH: "-main-file-name" "crash-report-clang-cl.c"
+// CHECKSH: "crash-report-{{[^ ]*}}.c"
diff --git a/clang/test/Driver/hip-sanitize-options.hip b/clang/test/Driver/hip-sanitize-options.hip
@@ -34,7 +34,7 @@
 // CHECK-NOT: {{"[^"]*lld(\.exe){0,1}".* ".*hip.bc"}}
 // CHECK: {{"[^"]*clang[^"]*".* "-triple" "x86_64-unknown-linux-gnu".* "-fsanitize=address"}}
 
-// NORDC: {{"[^"]*clang[^"]*".* "-emit-obj".* "-fcuda-is-device".* "-fsanitize=address".*}} "-o" "[[OUT:[^"]*.o]]"
+// NORDC: {{"[^"]*clang[^"]*".* "-emit-obj".* "-fcuda-is-device".* "-mlink-bitcode-file" ".*asanrtl.bc".* "-mlink-builtin-bitcode" ".*hip.bc".* "-fsanitize=address".*}} "-o" "[[OUT:[^"]*.o]]"
 // NORDC: {{"[^"]*lld(\.exe){0,1}".*}} "[[OUT]]" {{".*asanrtl.bc" ".*hip.bc"}}
 // NORDC: {{"[^"]*clang[^"]*".* "-triple" "x86_64-unknown-linux-gnu".* "-fsanitize=address"}}
 

diff --git a/clang/test/Lexer/Inputs/final-macro.h b/clang/test/Lexer/Inputs/final-macro.h
@@ -0,0 +1,4 @@
+// expected-warning@+2{{macro 'Foo' has been marked as deprecated}}
+// expected-warning@+1{{macro 'Foo' has been marked as unsafe for use in headers}}
+#if Foo
+#endif
diff --git a/clang/test/Lexer/Inputs/unsafe-macro.h b/clang/test/Lexer/Inputs/unsafe-macro.h
@@ -4,7 +4,7 @@
 // expected-error@+1{{expected identifier}}
 #pragma clang restrict_expansion(4
 
-// expected-error@+1{{no macro named foo}}
+// expected-error@+1{{no macro named 'foo'}}
 #pragma clang restrict_expansion(foo)
 
 

diff --git a/clang/test/Lexer/deprecate-macro.c b/clang/test/Lexer/deprecate-macro.c
@@ -6,17 +6,26 @@
 // expected-error@+1{{expected identifier}}
 #pragma clang deprecated(4
 
-// expected-error@+1{{no macro named foo}}
+// expected-error@+1{{no macro named 'foo'}}
 #pragma clang deprecated(foo)
 
 #define bar 1
+// expected-note@+1{{macro marked 'deprecated' here}} 
 #pragma clang deprecated(bar, "bar is deprecated use 1")
 
 // expected-warning@+1{{macro 'bar' has been marked as deprecated: bar is deprecated use 1}}
 #if bar
 #endif
 
 #define foo 1
+// expected-note@+8{{macro marked 'deprecated' here}} 
+// expected-note@+7{{macro marked 'deprecated' here}} 
+// expected-note@+6{{macro marked 'deprecated' here}} 
+// expected-note@+5{{macro marked 'deprecated' here}} 
+// expected-note@+4{{macro marked 'deprecated' here}} 
+// expected-note@+3{{macro marked 'deprecated' here}} 
+// expected-note@+2{{macro marked 'deprecated' here}} 
+// expected-note@+1{{macro marked 'deprecated' here}} 
 #pragma clang deprecated(foo)
 
 // expected-error@+1{{expected )}}
@@ -39,7 +48,7 @@
 #endif
 
 int main(int argc, char** argv) {
-  // expected-error@+1{{no macro named main}}
+// expected-error@+1{{no macro named 'main'}}
 #pragma clang deprecated(main)
 
   // expected-warning@+1{{macro 'foo' has been marked as deprecated}}

diff --git a/clang/test/Lexer/final-macro.c b/clang/test/Lexer/final-macro.c
@@ -0,0 +1,45 @@
+// RUN: %clang_cc1 -Wfinal-macro %s -fsyntax-only -verify
+
+// Test warning production
+#define Foo 1
+// expected-note@+1 4{{macro marked 'final' here}}
+#pragma clang final(Foo)
+
+// expected-warning@+2{{macro 'Foo' has been marked as final and should not be redefined}}
+// expected-note@+1{{previous definition is here}}
+#define Foo 1
+
+// expected-warning@+2{{macro 'Foo' has been marked as final and should not be redefined}}
+// expected-warning@+1{{'Foo' macro redefined}}
+#define Foo 2
+
+// expected-warning@+1{{redefining builtin macro}}
+#define __TIME__ 1
+
+// expected-warning@+1{{undefining builtin macro}}
+#undef __TIMESTAMP__
+
+// expected-warning@+1{{macro 'Foo' has been marked as final and should not be undefined}}
+#undef Foo
+// expected-warning@+1{{macro 'Foo' has been marked as final and should not be redefined}}
+#define Foo 3
+
+// Test parse errors
+// expected-error@+1{{expected (}}
+#pragma clang final
+
+// expected-error@+1{{expected )}}
+#pragma clang final(Foo
+
+// expected-error@+1{{no macro named 'Baz'}}
+#pragma clang final(Baz)
+
+// expected-error@+1{{expected identifier}}
+#pragma clang final(4)
+
+// expected-error@+1{{expected (}}
+#pragma clang final Baz
+
+// no diagnostics triggered by these pragmas.
+#pragma clang deprecated(Foo)
+#pragma clang restrict_expansion(Foo)
diff --git a/clang/test/Lexer/pedantic-macro-interplay.c b/clang/test/Lexer/pedantic-macro-interplay.c
@@ -11,4 +11,17 @@
 // not-expected-warning@+1{{macro 'UNSAFE_MACRO_2' has been marked as deprecated: Don't use this!}}
 #pragma clang restrict_expansion(UNSAFE_MACRO_2, "Don't use this!")
 
-// expected-no-diagnostics
+
+#define Foo 1
+#pragma clang final(Foo)
+// expected-note@+2{{macro marked 'deprecated' here}}
+// expected-note@+1{{macro marked 'deprecated' here}}
+#pragma clang deprecated(Foo)
+// expected-note@+1{{macro marked 'restrict_expansion' here}}
+#pragma clang restrict_expansion(Foo)
+
+// Test that unsafe_header and deprecated markings stick around after the undef
+#include "Inputs/final-macro.h"
+
+// expected-warning@+1{{macro 'Foo' has been marked as deprecated}}
+const int X = Foo;
diff --git a/clang/test/OpenMP/nvptx_distribute_parallel_generic_mode_codegen.cpp b/clang/test/OpenMP/nvptx_distribute_parallel_generic_mode_codegen.cpp
diff --git a/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen_tbaa_PR46146.cpp b/clang/test/OpenMP/nvptx_target_parallel_reduction_codegen_tbaa_PR46146.cpp
diff --git a/clang/test/OpenMP/nvptx_target_teams_distribute_codegen.cpp b/clang/test/OpenMP/nvptx_target_teams_distribute_codegen.cpp
@@ -615,7 +615,7 @@ int bar(int n){
 // CHECK1-NEXT:    store i32 0, i32* [[DOTOMP_IS_LAST]], align 4
 // CHECK1-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
 // CHECK1-NEXT:    [[TMP1:%.*]] = load i32, i32* [[TMP0]], align 4
-// CHECK1-NEXT:    call void @__kmpc_for_static_init_4(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP1]], i32 92, i32* [[DOTOMP_IS_LAST]], i32* [[DOTOMP_LB]], i32* [[DOTOMP_UB]], i32* [[DOTOMP_STRIDE]], i32 1, i32 1)
+// CHECK1-NEXT:    call void @__kmpc_distribute_static_init_4(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP1]], i32 92, i32* [[DOTOMP_IS_LAST]], i32* [[DOTOMP_LB]], i32* [[DOTOMP_UB]], i32* [[DOTOMP_STRIDE]], i32 1, i32 1)
 // CHECK1-NEXT:    [[TMP2:%.*]] = load i32, i32* [[DOTOMP_UB]], align 4
 // CHECK1-NEXT:    [[CMP:%.*]] = icmp sgt i32 [[TMP2]], 9
 // CHECK1-NEXT:    br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
@@ -656,7 +656,7 @@ int bar(int n){
 // CHECK1:       omp.inner.for.end:
 // CHECK1-NEXT:    br label [[OMP_LOOP_EXIT:%.*]]
 // CHECK1:       omp.loop.exit:
-// CHECK1-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP1]])
+// CHECK1-NEXT:    call void @__kmpc_distribute_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP1]])
 // CHECK1-NEXT:    call void @__kmpc_free_shared(i8* [[I]], i64 4)
 // CHECK1-NEXT:    ret void
 //
@@ -737,7 +737,7 @@ int bar(int n){
 // CHECK2-NEXT:    store i32 0, i32* [[DOTOMP_IS_LAST]], align 4
 // CHECK2-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 4
 // CHECK2-NEXT:    [[TMP1:%.*]] = load i32, i32* [[TMP0]], align 4
-// CHECK2-NEXT:    call void @__kmpc_for_static_init_4(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP1]], i32 92, i32* [[DOTOMP_IS_LAST]], i32* [[DOTOMP_LB]], i32* [[DOTOMP_UB]], i32* [[DOTOMP_STRIDE]], i32 1, i32 1)
+// CHECK2-NEXT:    call void @__kmpc_distribute_static_init_4(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP1]], i32 92, i32* [[DOTOMP_IS_LAST]], i32* [[DOTOMP_LB]], i32* [[DOTOMP_UB]], i32* [[DOTOMP_STRIDE]], i32 1, i32 1)
 // CHECK2-NEXT:    [[TMP2:%.*]] = load i32, i32* [[DOTOMP_UB]], align 4
 // CHECK2-NEXT:    [[CMP:%.*]] = icmp sgt i32 [[TMP2]], 9
 // CHECK2-NEXT:    br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
@@ -778,7 +778,7 @@ int bar(int n){
 // CHECK2:       omp.inner.for.end:
 // CHECK2-NEXT:    br label [[OMP_LOOP_EXIT:%.*]]
 // CHECK2:       omp.loop.exit:
-// CHECK2-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP1]])
+// CHECK2-NEXT:    call void @__kmpc_distribute_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP1]])
 // CHECK2-NEXT:    call void @__kmpc_free_shared(i8* [[I]], i32 4)
 // CHECK2-NEXT:    ret void
 //
@@ -859,7 +859,7 @@ int bar(int n){
 // CHECK3-NEXT:    store i32 0, i32* [[DOTOMP_IS_LAST]], align 4
 // CHECK3-NEXT:    [[TMP0:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 4
 // CHECK3-NEXT:    [[TMP1:%.*]] = load i32, i32* [[TMP0]], align 4
-// CHECK3-NEXT:    call void @__kmpc_for_static_init_4(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP1]], i32 92, i32* [[DOTOMP_IS_LAST]], i32* [[DOTOMP_LB]], i32* [[DOTOMP_UB]], i32* [[DOTOMP_STRIDE]], i32 1, i32 1)
+// CHECK3-NEXT:    call void @__kmpc_distribute_static_init_4(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP1]], i32 92, i32* [[DOTOMP_IS_LAST]], i32* [[DOTOMP_LB]], i32* [[DOTOMP_UB]], i32* [[DOTOMP_STRIDE]], i32 1, i32 1)
 // CHECK3-NEXT:    [[TMP2:%.*]] = load i32, i32* [[DOTOMP_UB]], align 4
 // CHECK3-NEXT:    [[CMP:%.*]] = icmp sgt i32 [[TMP2]], 9
 // CHECK3-NEXT:    br i1 [[CMP]], label [[COND_TRUE:%.*]], label [[COND_FALSE:%.*]]
@@ -900,7 +900,7 @@ int bar(int n){
 // CHECK3:       omp.inner.for.end:
 // CHECK3-NEXT:    br label [[OMP_LOOP_EXIT:%.*]]
 // CHECK3:       omp.loop.exit:
-// CHECK3-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP1]])
+// CHECK3-NEXT:    call void @__kmpc_distribute_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP1]])
 // CHECK3-NEXT:    call void @__kmpc_free_shared(i8* [[I]], i32 4)
 // CHECK3-NEXT:    ret void
 //

diff --git a/clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_codegen.cpp b/clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_codegen.cpp
diff --git a/clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_generic_mode_codegen.cpp b/clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_generic_mode_codegen.cpp
@@ -116,7 +116,7 @@ int main(int argc, char **argv) {
 // CHECK1-NEXT:    [[NVPTX_NUM_THREADS:%.*]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x()
 // CHECK1-NEXT:    [[TMP5:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
 // CHECK1-NEXT:    [[TMP6:%.*]] = load i32, i32* [[TMP5]], align 4
-// CHECK1-NEXT:    call void @__kmpc_for_static_init_4(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP6]], i32 91, i32* [[DOTOMP_IS_LAST]], i32* [[DOTOMP_COMB_LB]], i32* [[DOTOMP_COMB_UB]], i32* [[DOTOMP_STRIDE]], i32 1, i32 [[NVPTX_NUM_THREADS]])
+// CHECK1-NEXT:    call void @__kmpc_distribute_static_init_4(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP6]], i32 91, i32* [[DOTOMP_IS_LAST]], i32* [[DOTOMP_COMB_LB]], i32* [[DOTOMP_COMB_UB]], i32* [[DOTOMP_STRIDE]], i32 1, i32 [[NVPTX_NUM_THREADS]])
 // CHECK1-NEXT:    [[TMP7:%.*]] = load i32, i32* [[DOTOMP_COMB_UB]], align 4
 // CHECK1-NEXT:    [[TMP8:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_3]], align 4
 // CHECK1-NEXT:    [[CMP6:%.*]] = icmp sgt i32 [[TMP7]], [[TMP8]]
@@ -206,7 +206,7 @@ int main(int argc, char **argv) {
 // CHECK1:       omp.loop.exit:
 // CHECK1-NEXT:    [[TMP46:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
 // CHECK1-NEXT:    [[TMP47:%.*]] = load i32, i32* [[TMP46]], align 4
-// CHECK1-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP47]])
+// CHECK1-NEXT:    call void @__kmpc_distribute_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP47]])
 // CHECK1-NEXT:    br label [[OMP_PRECOND_END]]
 // CHECK1:       omp.precond.end:
 // CHECK1-NEXT:    ret void
@@ -333,7 +333,7 @@ int main(int argc, char **argv) {
 // CHECK1:       omp.dispatch.end:
 // CHECK1-NEXT:    [[TMP25:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
 // CHECK1-NEXT:    [[TMP26:%.*]] = load i32, i32* [[TMP25]], align 4
-// CHECK1-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP26]])
+// CHECK1-NEXT:    call void @__kmpc_distribute_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP26]])
 // CHECK1-NEXT:    br label [[OMP_PRECOND_END]]
 // CHECK1:       omp.precond.end:
 // CHECK1-NEXT:    ret void
@@ -420,7 +420,7 @@ int main(int argc, char **argv) {
 // CHECK2-NEXT:    [[NVPTX_NUM_THREADS:%.*]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x()
 // CHECK2-NEXT:    [[TMP5:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 4
 // CHECK2-NEXT:    [[TMP6:%.*]] = load i32, i32* [[TMP5]], align 4
-// CHECK2-NEXT:    call void @__kmpc_for_static_init_4(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP6]], i32 91, i32* [[DOTOMP_IS_LAST]], i32* [[DOTOMP_COMB_LB]], i32* [[DOTOMP_COMB_UB]], i32* [[DOTOMP_STRIDE]], i32 1, i32 [[NVPTX_NUM_THREADS]])
+// CHECK2-NEXT:    call void @__kmpc_distribute_static_init_4(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP6]], i32 91, i32* [[DOTOMP_IS_LAST]], i32* [[DOTOMP_COMB_LB]], i32* [[DOTOMP_COMB_UB]], i32* [[DOTOMP_STRIDE]], i32 1, i32 [[NVPTX_NUM_THREADS]])
 // CHECK2-NEXT:    [[TMP7:%.*]] = load i32, i32* [[DOTOMP_COMB_UB]], align 4
 // CHECK2-NEXT:    [[TMP8:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2]], align 4
 // CHECK2-NEXT:    [[CMP5:%.*]] = icmp sgt i32 [[TMP7]], [[TMP8]]
@@ -506,7 +506,7 @@ int main(int argc, char **argv) {
 // CHECK2:       omp.loop.exit:
 // CHECK2-NEXT:    [[TMP44:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 4
 // CHECK2-NEXT:    [[TMP45:%.*]] = load i32, i32* [[TMP44]], align 4
-// CHECK2-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP45]])
+// CHECK2-NEXT:    call void @__kmpc_distribute_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP45]])
 // CHECK2-NEXT:    br label [[OMP_PRECOND_END]]
 // CHECK2:       omp.precond.end:
 // CHECK2-NEXT:    ret void
@@ -627,7 +627,7 @@ int main(int argc, char **argv) {
 // CHECK2:       omp.dispatch.end:
 // CHECK2-NEXT:    [[TMP25:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 4
 // CHECK2-NEXT:    [[TMP26:%.*]] = load i32, i32* [[TMP25]], align 4
-// CHECK2-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP26]])
+// CHECK2-NEXT:    call void @__kmpc_distribute_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP26]])
 // CHECK2-NEXT:    br label [[OMP_PRECOND_END]]
 // CHECK2:       omp.precond.end:
 // CHECK2-NEXT:    ret void
@@ -714,7 +714,7 @@ int main(int argc, char **argv) {
 // CHECK3-NEXT:    [[NVPTX_NUM_THREADS:%.*]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x()
 // CHECK3-NEXT:    [[TMP5:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 4
 // CHECK3-NEXT:    [[TMP6:%.*]] = load i32, i32* [[TMP5]], align 4
-// CHECK3-NEXT:    call void @__kmpc_for_static_init_4(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP6]], i32 91, i32* [[DOTOMP_IS_LAST]], i32* [[DOTOMP_COMB_LB]], i32* [[DOTOMP_COMB_UB]], i32* [[DOTOMP_STRIDE]], i32 1, i32 [[NVPTX_NUM_THREADS]])
+// CHECK3-NEXT:    call void @__kmpc_distribute_static_init_4(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP6]], i32 91, i32* [[DOTOMP_IS_LAST]], i32* [[DOTOMP_COMB_LB]], i32* [[DOTOMP_COMB_UB]], i32* [[DOTOMP_STRIDE]], i32 1, i32 [[NVPTX_NUM_THREADS]])
 // CHECK3-NEXT:    [[TMP7:%.*]] = load i32, i32* [[DOTOMP_COMB_UB]], align 4
 // CHECK3-NEXT:    [[TMP8:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2]], align 4
 // CHECK3-NEXT:    [[CMP5:%.*]] = icmp sgt i32 [[TMP7]], [[TMP8]]
@@ -800,7 +800,7 @@ int main(int argc, char **argv) {
 // CHECK3:       omp.loop.exit:
 // CHECK3-NEXT:    [[TMP44:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 4
 // CHECK3-NEXT:    [[TMP45:%.*]] = load i32, i32* [[TMP44]], align 4
-// CHECK3-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP45]])
+// CHECK3-NEXT:    call void @__kmpc_distribute_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP45]])
 // CHECK3-NEXT:    br label [[OMP_PRECOND_END]]
 // CHECK3:       omp.precond.end:
 // CHECK3-NEXT:    ret void
@@ -921,7 +921,7 @@ int main(int argc, char **argv) {
 // CHECK3:       omp.dispatch.end:
 // CHECK3-NEXT:    [[TMP25:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 4
 // CHECK3-NEXT:    [[TMP26:%.*]] = load i32, i32* [[TMP25]], align 4
-// CHECK3-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP26]])
+// CHECK3-NEXT:    call void @__kmpc_distribute_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP26]])
 // CHECK3-NEXT:    br label [[OMP_PRECOND_END]]
 // CHECK3:       omp.precond.end:
 // CHECK3-NEXT:    ret void
@@ -1014,7 +1014,7 @@ int main(int argc, char **argv) {
 // CHECK4-NEXT:    [[NVPTX_NUM_THREADS:%.*]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x()
 // CHECK4-NEXT:    [[TMP5:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
 // CHECK4-NEXT:    [[TMP6:%.*]] = load i32, i32* [[TMP5]], align 4
-// CHECK4-NEXT:    call void @__kmpc_for_static_init_4(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP6]], i32 91, i32* [[DOTOMP_IS_LAST]], i32* [[DOTOMP_COMB_LB]], i32* [[DOTOMP_COMB_UB]], i32* [[DOTOMP_STRIDE]], i32 1, i32 [[NVPTX_NUM_THREADS]])
+// CHECK4-NEXT:    call void @__kmpc_distribute_static_init_4(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP6]], i32 91, i32* [[DOTOMP_IS_LAST]], i32* [[DOTOMP_COMB_LB]], i32* [[DOTOMP_COMB_UB]], i32* [[DOTOMP_STRIDE]], i32 1, i32 [[NVPTX_NUM_THREADS]])
 // CHECK4-NEXT:    [[TMP7:%.*]] = load i32, i32* [[DOTOMP_COMB_UB]], align 4
 // CHECK4-NEXT:    [[TMP8:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_3]], align 4
 // CHECK4-NEXT:    [[CMP6:%.*]] = icmp sgt i32 [[TMP7]], [[TMP8]]
@@ -1104,7 +1104,7 @@ int main(int argc, char **argv) {
 // CHECK4:       omp.loop.exit:
 // CHECK4-NEXT:    [[TMP46:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
 // CHECK4-NEXT:    [[TMP47:%.*]] = load i32, i32* [[TMP46]], align 4
-// CHECK4-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP47]])
+// CHECK4-NEXT:    call void @__kmpc_distribute_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP47]])
 // CHECK4-NEXT:    br label [[OMP_PRECOND_END]]
 // CHECK4:       omp.precond.end:
 // CHECK4-NEXT:    ret void
@@ -1231,7 +1231,7 @@ int main(int argc, char **argv) {
 // CHECK4:       omp.dispatch.end:
 // CHECK4-NEXT:    [[TMP25:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 8
 // CHECK4-NEXT:    [[TMP26:%.*]] = load i32, i32* [[TMP25]], align 4
-// CHECK4-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP26]])
+// CHECK4-NEXT:    call void @__kmpc_distribute_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP26]])
 // CHECK4-NEXT:    br label [[OMP_PRECOND_END]]
 // CHECK4:       omp.precond.end:
 // CHECK4-NEXT:    ret void
@@ -1318,7 +1318,7 @@ int main(int argc, char **argv) {
 // CHECK5-NEXT:    [[NVPTX_NUM_THREADS:%.*]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x()
 // CHECK5-NEXT:    [[TMP5:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 4
 // CHECK5-NEXT:    [[TMP6:%.*]] = load i32, i32* [[TMP5]], align 4
-// CHECK5-NEXT:    call void @__kmpc_for_static_init_4(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP6]], i32 91, i32* [[DOTOMP_IS_LAST]], i32* [[DOTOMP_COMB_LB]], i32* [[DOTOMP_COMB_UB]], i32* [[DOTOMP_STRIDE]], i32 1, i32 [[NVPTX_NUM_THREADS]])
+// CHECK5-NEXT:    call void @__kmpc_distribute_static_init_4(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP6]], i32 91, i32* [[DOTOMP_IS_LAST]], i32* [[DOTOMP_COMB_LB]], i32* [[DOTOMP_COMB_UB]], i32* [[DOTOMP_STRIDE]], i32 1, i32 [[NVPTX_NUM_THREADS]])
 // CHECK5-NEXT:    [[TMP7:%.*]] = load i32, i32* [[DOTOMP_COMB_UB]], align 4
 // CHECK5-NEXT:    [[TMP8:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2]], align 4
 // CHECK5-NEXT:    [[CMP5:%.*]] = icmp sgt i32 [[TMP7]], [[TMP8]]
@@ -1404,7 +1404,7 @@ int main(int argc, char **argv) {
 // CHECK5:       omp.loop.exit:
 // CHECK5-NEXT:    [[TMP44:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 4
 // CHECK5-NEXT:    [[TMP45:%.*]] = load i32, i32* [[TMP44]], align 4
-// CHECK5-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP45]])
+// CHECK5-NEXT:    call void @__kmpc_distribute_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP45]])
 // CHECK5-NEXT:    br label [[OMP_PRECOND_END]]
 // CHECK5:       omp.precond.end:
 // CHECK5-NEXT:    ret void
@@ -1525,7 +1525,7 @@ int main(int argc, char **argv) {
 // CHECK5:       omp.dispatch.end:
 // CHECK5-NEXT:    [[TMP25:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 4
 // CHECK5-NEXT:    [[TMP26:%.*]] = load i32, i32* [[TMP25]], align 4
-// CHECK5-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP26]])
+// CHECK5-NEXT:    call void @__kmpc_distribute_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP26]])
 // CHECK5-NEXT:    br label [[OMP_PRECOND_END]]
 // CHECK5:       omp.precond.end:
 // CHECK5-NEXT:    ret void
@@ -1612,7 +1612,7 @@ int main(int argc, char **argv) {
 // CHECK6-NEXT:    [[NVPTX_NUM_THREADS:%.*]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x()
 // CHECK6-NEXT:    [[TMP5:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 4
 // CHECK6-NEXT:    [[TMP6:%.*]] = load i32, i32* [[TMP5]], align 4
-// CHECK6-NEXT:    call void @__kmpc_for_static_init_4(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP6]], i32 91, i32* [[DOTOMP_IS_LAST]], i32* [[DOTOMP_COMB_LB]], i32* [[DOTOMP_COMB_UB]], i32* [[DOTOMP_STRIDE]], i32 1, i32 [[NVPTX_NUM_THREADS]])
+// CHECK6-NEXT:    call void @__kmpc_distribute_static_init_4(%struct.ident_t* @[[GLOB2:[0-9]+]], i32 [[TMP6]], i32 91, i32* [[DOTOMP_IS_LAST]], i32* [[DOTOMP_COMB_LB]], i32* [[DOTOMP_COMB_UB]], i32* [[DOTOMP_STRIDE]], i32 1, i32 [[NVPTX_NUM_THREADS]])
 // CHECK6-NEXT:    [[TMP7:%.*]] = load i32, i32* [[DOTOMP_COMB_UB]], align 4
 // CHECK6-NEXT:    [[TMP8:%.*]] = load i32, i32* [[DOTCAPTURE_EXPR_2]], align 4
 // CHECK6-NEXT:    [[CMP5:%.*]] = icmp sgt i32 [[TMP7]], [[TMP8]]
@@ -1698,7 +1698,7 @@ int main(int argc, char **argv) {
 // CHECK6:       omp.loop.exit:
 // CHECK6-NEXT:    [[TMP44:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 4
 // CHECK6-NEXT:    [[TMP45:%.*]] = load i32, i32* [[TMP44]], align 4
-// CHECK6-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP45]])
+// CHECK6-NEXT:    call void @__kmpc_distribute_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP45]])
 // CHECK6-NEXT:    br label [[OMP_PRECOND_END]]
 // CHECK6:       omp.precond.end:
 // CHECK6-NEXT:    ret void
@@ -1819,7 +1819,7 @@ int main(int argc, char **argv) {
 // CHECK6:       omp.dispatch.end:
 // CHECK6-NEXT:    [[TMP25:%.*]] = load i32*, i32** [[DOTGLOBAL_TID__ADDR]], align 4
 // CHECK6-NEXT:    [[TMP26:%.*]] = load i32, i32* [[TMP25]], align 4
-// CHECK6-NEXT:    call void @__kmpc_for_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP26]])
+// CHECK6-NEXT:    call void @__kmpc_distribute_static_fini(%struct.ident_t* @[[GLOB2]], i32 [[TMP26]])
 // CHECK6-NEXT:    br label [[OMP_PRECOND_END]]
 // CHECK6:       omp.precond.end:
 // CHECK6-NEXT:    ret void

diff --git a/clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_simd_codegen.cpp b/clang/test/OpenMP/nvptx_target_teams_distribute_parallel_for_simd_codegen.cpp
diff --git a/clang/test/OpenMP/nvptx_target_teams_distribute_simd_codegen.cpp b/clang/test/OpenMP/nvptx_target_teams_distribute_simd_codegen.cpp
@@ -73,24 +73,24 @@ int bar(int n){
 // CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
 // CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 false)
 
-// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 91,
-// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: call void @__kmpc_distribute_static_init_4({{.+}}, {{.+}}, {{.+}} 91,
+// CHECK: call void @__kmpc_distribute_static_fini(
 // CHECK: ret void
 
 // CHECK-LABEL: define {{.*}}void {{@__omp_offloading_.+}}_l43(
 // CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
 // CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 false)
 
-// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 91,
-// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: call void @__kmpc_distribute_static_init_4({{.+}}, {{.+}}, {{.+}} 91,
+// CHECK: call void @__kmpc_distribute_static_fini(
 // CHECK: ret void
 
 // CHECK-LABEL: define {{.*}}void {{@__omp_offloading_.+}}_l48(
 // CHECK: call i32 @__kmpc_target_init({{.*}}, i8 2, i1 false, i1 false)
 // CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 false)
 
-// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 91,
-// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: call void @__kmpc_distribute_static_init_4({{.+}}, {{.+}}, {{.+}} 91,
+// CHECK: call void @__kmpc_distribute_static_fini(
 // CHECK: ret void
 
 // CHECK: define {{.*}}void {{@__omp_offloading_.+}}_l53({{.+}}, i{{32|64}} [[F_IN:%.+]])
@@ -99,8 +99,8 @@ int bar(int n){
 // CHECK: call void @__kmpc_target_deinit({{.*}}, i8 2, i1 false)
 
 // CHECK: store {{.+}} 99, {{.+}}* [[COMB_UB:%.+]], align
-// CHECK: call void @__kmpc_for_static_init_4({{.+}}, {{.+}}, {{.+}} 91, {{.+}}, {{.+}}, {{.+}}* [[COMB_UB]],
-// CHECK: call void @__kmpc_for_static_fini(
+// CHECK: call void @__kmpc_distribute_static_init_4({{.+}}, {{.+}}, {{.+}} 91, {{.+}}, {{.+}}, {{.+}}* [[COMB_UB]],
+// CHECK: call void @__kmpc_distribute_static_fini(
 // CHECK: ret void
 
 #endif
diff --git a/clang/tools/scan-build-py/lib/libscanbuild/analyze.py b/clang/tools/scan-build-py/lib/libscanbuild/analyze.py
@@ -39,7 +39,7 @@
 
 __all__ = ['scan_build', 'analyze_build', 'analyze_compiler_wrapper']
 
-scanbuild_dir = os.path.dirname(__import__('sys').argv[0])
+scanbuild_dir = os.path.dirname(os.path.realpath(__import__('sys').argv[0]))
 
 COMPILER_WRAPPER_CC = os.path.join(scanbuild_dir, '..', 'libexec', 'analyze-cc')
 COMPILER_WRAPPER_CXX = os.path.join(scanbuild_dir, '..', 'libexec', 'analyze-c++')

diff --git a/clang/unittests/AST/ASTImporterTest.cpp b/clang/unittests/AST/ASTImporterTest.cpp
@@ -290,6 +290,25 @@ TEST_P(ImportExpr, ImportChooseExpr) {
              functionDecl(hasDescendant(chooseExpr())));
 }
 
+const internal::VariadicDynCastAllOfMatcher<Stmt, ShuffleVectorExpr>
+    shuffleVectorExpr;
+
+TEST_P(ImportExpr, ImportShuffleVectorExpr) {
+  MatchVerifier<Decl> Verifier;
+  constexpr auto Code = R"code(
+    typedef double vector4double __attribute__((__vector_size__(32)));
+    vector4double declToImport(vector4double a, vector4double b) {
+      return __builtin_shufflevector(a, b, 0, 1, 2, 3);
+    }
+  )code";
+  const auto Pattern = functionDecl(hasDescendant(shuffleVectorExpr(
+      allOf(has(declRefExpr(to(parmVarDecl(hasName("a"))))),
+            has(declRefExpr(to(parmVarDecl(hasName("b"))))),
+            has(integerLiteral(equals(0))), has(integerLiteral(equals(1))),
+            has(integerLiteral(equals(2))), has(integerLiteral(equals(3)))))));
+  testImport(Code, Lang_C99, "", Lang_C99, Verifier, Pattern);
+}
+
 TEST_P(ImportExpr, ImportGNUNullExpr) {
   MatchVerifier<Decl> Verifier;
   testImport("void declToImport() { (void)__null; }", Lang_CXX03, "",

diff --git a/compiler-rt/lib/tsan/rtl/tsan_platform_linux.cpp b/compiler-rt/lib/tsan/rtl/tsan_platform_linux.cpp
@@ -122,7 +122,7 @@ void FillProfileCallback(uptr p, uptr rss, bool file,
 void WriteMemoryProfile(char *buf, uptr buf_size, u64 uptime_ns) {
   uptr mem[MemCount];
   internal_memset(mem, 0, sizeof(mem));
-  GetMemoryProfile(FillProfileCallback, mem, 7);
+  GetMemoryProfile(FillProfileCallback, mem, MemCount);
   auto meta = ctx->metamap.GetMemoryStats();
   StackDepotStats *stacks = StackDepotGetStats();
   uptr nthread, nlive;
@@ -453,18 +453,19 @@ static void InitializeLongjmpXorKey() {
 }
 #endif
 
+extern "C" void __tsan_tls_initialization() {}
+
 void ImitateTlsWrite(ThreadState *thr, uptr tls_addr, uptr tls_size) {
-  // Check that the thr object is in tls;
   const uptr thr_beg = (uptr)thr;
   const uptr thr_end = (uptr)thr + sizeof(*thr);
-  CHECK_GE(thr_beg, tls_addr);
-  CHECK_LE(thr_beg, tls_addr + tls_size);
-  CHECK_GE(thr_end, tls_addr);
-  CHECK_LE(thr_end, tls_addr + tls_size);
-  // Since the thr object is huge, skip it.
-  MemoryRangeImitateWrite(thr, /*pc=*/2, tls_addr, thr_beg - tls_addr);
-  MemoryRangeImitateWrite(thr, /*pc=*/2, thr_end,
-                          tls_addr + tls_size - thr_end);
+  // ThreadState is normally allocated in TLS and is large,
+  // so we skip it. But unit tests allocate ThreadState outside of TLS.
+  if (thr_beg < tls_addr || thr_end >= tls_addr + tls_size)
+    return;
+  const uptr pc = StackTrace::GetNextInstructionPc(
+      reinterpret_cast<uptr>(__tsan_tls_initialization));
+  MemoryRangeImitateWrite(thr, pc, tls_addr, thr_beg - tls_addr);
+  MemoryRangeImitateWrite(thr, pc, thr_end, tls_addr + tls_size - thr_end);
 }
 
 // Note: this function runs with async signals enabled,

diff --git a/compiler-rt/lib/tsan/rtl/tsan_rtl.cpp b/compiler-rt/lib/tsan/rtl/tsan_rtl.cpp
@@ -145,6 +145,17 @@ ThreadState::ThreadState(Context *ctx, Tid tid, int unique_id, u64 epoch,
       last_sleep_clock(tid)
 #endif
 {
+  CHECK_EQ(reinterpret_cast<uptr>(this) % SANITIZER_CACHE_LINE_SIZE, 0);
+#if !SANITIZER_GO
+  shadow_stack_pos = shadow_stack;
+  shadow_stack_end = shadow_stack + kShadowStackSize;
+#else
+  // Setup dynamic shadow stack.
+  const int kInitStackSize = 8;
+  shadow_stack = (uptr *)Alloc(kInitStackSize * sizeof(uptr));
+  shadow_stack_pos = shadow_stack;
+  shadow_stack_end = shadow_stack + kInitStackSize;
+#endif
 }
 
 #if !SANITIZER_GO

diff --git a/compiler-rt/lib/tsan/rtl/tsan_rtl.h b/compiler-rt/lib/tsan/rtl/tsan_rtl.h
@@ -159,10 +159,12 @@ struct ThreadState {
 #if !SANITIZER_GO
   IgnoreSet mop_ignore_set;
   IgnoreSet sync_ignore_set;
-#endif
-  // C/C++ uses fixed size shadow stack embed into Trace.
+  // C/C++ uses fixed size shadow stack.
+  uptr shadow_stack[kShadowStackSize];
+#else
   // Go uses malloc-allocated shadow stack with dynamic size.
   uptr *shadow_stack;
+#endif
   uptr *shadow_stack_end;
   uptr *shadow_stack_pos;
   RawShadow *racy_shadow_addr;
@@ -221,7 +223,7 @@ struct ThreadState {
   explicit ThreadState(Context *ctx, Tid tid, int unique_id, u64 epoch,
                        unsigned reuse_count, uptr stk_addr, uptr stk_size,
                        uptr tls_addr, uptr tls_size);
-};
+} ALIGNED(SANITIZER_CACHE_LINE_SIZE);
 
 #if !SANITIZER_GO
 #if SANITIZER_MAC || SANITIZER_ANDROID

diff --git a/compiler-rt/lib/tsan/rtl/tsan_rtl_thread.cpp b/compiler-rt/lib/tsan/rtl/tsan_rtl_thread.cpp
@@ -138,6 +138,8 @@ void ThreadContext::OnCreated(void *arg) {
   creation_stack_id = CurrentStackId(args->thr, args->pc);
 }
 
+extern "C" void __tsan_stack_initialization() {}
+
 struct OnStartedArgs {
   ThreadState *thr;
   uptr stk_addr;
@@ -156,13 +158,6 @@ void ThreadStart(ThreadState *thr, Tid tid, tid_t os_id,
   if (thread_type != ThreadType::Fiber)
     GetThreadStackAndTls(tid == kMainTid, &stk_addr, &stk_size, &tls_addr,
                          &tls_size);
-
-  if (tid != kMainTid) {
-    if (stk_addr && stk_size)
-      MemoryRangeImitateWrite(thr, /*pc=*/ 1, stk_addr, stk_size);
-
-    if (tls_addr && tls_size) ImitateTlsWrite(thr, tls_addr, tls_size);
-  }
 #endif
 
   ThreadRegistry *tr = &ctx->thread_registry;
@@ -178,6 +173,22 @@ void ThreadStart(ThreadState *thr, Tid tid, tid_t os_id,
     ThreadIgnoreSyncBegin(thr, 0);
   }
 #endif
+
+#if !SANITIZER_GO
+  // Don't imitate stack/TLS writes for the main thread,
+  // because its initialization is synchronized with all
+  // subsequent threads anyway.
+  if (tid != kMainTid) {
+    if (stk_addr && stk_size) {
+      const uptr pc = StackTrace::GetNextInstructionPc(
+          reinterpret_cast<uptr>(__tsan_stack_initialization));
+      MemoryRangeImitateWrite(thr, pc, stk_addr, stk_size);
+    }
+
+    if (tls_addr && tls_size)
+      ImitateTlsWrite(thr, tls_addr, tls_size);
+  }
+#endif
 }
 
 void ThreadContext::OnStarted(void *arg) {
@@ -190,17 +201,6 @@ void ThreadContext::OnStarted(void *arg) {
   new (thr)
       ThreadState(ctx, tid, unique_id, epoch0, reuse_count, args->stk_addr,
                   args->stk_size, args->tls_addr, args->tls_size);
-#if !SANITIZER_GO
-  thr->shadow_stack = &ThreadTrace(thr->tid)->shadow_stack[0];
-  thr->shadow_stack_pos = thr->shadow_stack;
-  thr->shadow_stack_end = thr->shadow_stack + kShadowStackSize;
-#else
-  // Setup dynamic shadow stack.
-  const int kInitStackSize = 8;
-  thr->shadow_stack = (uptr *)Alloc(kInitStackSize * sizeof(uptr));
-  thr->shadow_stack_pos = thr->shadow_stack;
-  thr->shadow_stack_end = thr->shadow_stack + kInitStackSize;
-#endif
   if (common_flags()->detect_deadlocks)
     thr->dd_lt = ctx->dd->CreateLogicalThread(unique_id);
   thr->fast_state.SetHistorySize(flags()->history_size);

diff --git a/compiler-rt/lib/tsan/tests/unit/tsan_stack_test.cpp b/compiler-rt/lib/tsan/tests/unit/tsan_stack_test.cpp
@@ -19,10 +19,6 @@ namespace __tsan {
 template <typename StackTraceTy>
 static void TestStackTrace(StackTraceTy *trace) {
   ThreadState thr(0, 0, 0, 0, 0, 0, 0, 0, 0);
-  uptr stack[128];
-  thr.shadow_stack = &stack[0];
-  thr.shadow_stack_pos = &stack[0];
-  thr.shadow_stack_end = &stack[128];
 
   ObtainCurrentStack(&thr, 0, trace);
   EXPECT_EQ(0U, trace->size);
@@ -48,13 +44,8 @@ static void TestStackTrace(StackTraceTy *trace) {
 template<typename StackTraceTy>
 static void TestTrim(StackTraceTy *trace) {
   ThreadState thr(0, 0, 0, 0, 0, 0, 0, 0, 0);
-  const uptr kShadowStackSize = 2 * kStackTraceMax;
-  uptr stack[kShadowStackSize];
-  thr.shadow_stack = &stack[0];
-  thr.shadow_stack_pos = &stack[0];
-  thr.shadow_stack_end = &stack[kShadowStackSize];
 
-  for (uptr i = 0; i < kShadowStackSize; ++i)
+  for (uptr i = 0; i < 2 * kStackTraceMax; ++i)
     *thr.shadow_stack_pos++ = 100 + i;
 
   ObtainCurrentStack(&thr, 0, trace);

diff --git a/compiler-rt/lib/tsan/tests/unit/tsan_trace_test.cpp b/compiler-rt/lib/tsan/tests/unit/tsan_trace_test.cpp
@@ -22,102 +22,115 @@ using namespace v3;
 
 // We need to run all trace tests in a new thread,
 // so that the thread trace is empty initially.
-static void run_in_thread(void *(*f)(void *), void *arg = nullptr) {
-  pthread_t th;
-  pthread_create(&th, nullptr, f, arg);
-  pthread_join(th, nullptr);
-}
-
-#if SANITIZER_MAC
-// These tests are currently failing on Mac.
-// See https://reviews.llvm.org/D107911 for more details.
-#  define MAYBE_RestoreAccess DISABLED_RestoreAccess
-#  define MAYBE_MemoryAccessSize DISABLED_MemoryAccessSize
-#  define MAYBE_RestoreMutexLock DISABLED_RestoreMutexLock
-#  define MAYBE_MultiPart DISABLED_MultiPart
-#else
-#  define MAYBE_RestoreAccess RestoreAccess
-#  define MAYBE_MemoryAccessSize MemoryAccessSize
-#  define MAYBE_RestoreMutexLock RestoreMutexLock
-#  define MAYBE_MultiPart MultiPart
-#endif
+template <uptr N>
+struct ThreadArray {
+  ThreadArray() {
+    for (auto *&thr : threads) {
+      thr = static_cast<ThreadState *>(
+          MmapOrDie(sizeof(ThreadState), "ThreadState"));
+      Tid tid = ThreadCreate(cur_thread(), 0, 0, true);
+      Processor *proc = ProcCreate();
+      ProcWire(proc, thr);
+      ThreadStart(thr, tid, 0, ThreadType::Regular);
+    }
+  }
 
-TEST(Trace, MAYBE_RestoreAccess) {
-  struct Thread {
-    static void *Func(void *arg) {
-      // A basic test with some function entry/exit events,
-      // some mutex lock/unlock events and some other distracting
-      // memory events.
-      ThreadState *thr = cur_thread();
-      TraceFunc(thr, 0x1000);
-      TraceFunc(thr, 0x1001);
-      TraceMutexLock(thr, v3::EventType::kLock, 0x4000, 0x5000, 0x6000);
-      TraceMutexLock(thr, v3::EventType::kLock, 0x4001, 0x5001, 0x6001);
-      TraceMutexUnlock(thr, 0x5000);
-      TraceFunc(thr);
-      CHECK(TryTraceMemoryAccess(thr, 0x2001, 0x3001, 8, kAccessRead));
-      TraceMutexLock(thr, v3::EventType::kRLock, 0x4002, 0x5002, 0x6002);
-      TraceFunc(thr, 0x1002);
-      CHECK(TryTraceMemoryAccess(thr, 0x2000, 0x3000, 8, kAccessRead));
-      // This is the access we want to find.
-      // The previous one is equivalent, but RestoreStack must prefer
-      // the last of the matchig accesses.
-      CHECK(TryTraceMemoryAccess(thr, 0x2002, 0x3000, 8, kAccessRead));
-      Lock lock1(&ctx->slot_mtx);
-      ThreadRegistryLock lock2(&ctx->thread_registry);
-      VarSizeStackTrace stk;
-      MutexSet mset;
-      uptr tag = kExternalTagNone;
-      bool res =
-          RestoreStack(thr->tid, v3::EventType::kAccessExt, thr->sid,
-                       thr->epoch, 0x3000, 8, kAccessRead, &stk, &mset, &tag);
-      CHECK(res);
-      CHECK_EQ(stk.size, 3);
-      CHECK_EQ(stk.trace[0], 0x1000);
-      CHECK_EQ(stk.trace[1], 0x1002);
-      CHECK_EQ(stk.trace[2], 0x2002);
-      CHECK_EQ(mset.Size(), 2);
-      CHECK_EQ(mset.Get(0).addr, 0x5001);
-      CHECK_EQ(mset.Get(0).stack_id, 0x6001);
-      CHECK_EQ(mset.Get(0).write, true);
-      CHECK_EQ(mset.Get(1).addr, 0x5002);
-      CHECK_EQ(mset.Get(1).stack_id, 0x6002);
-      CHECK_EQ(mset.Get(1).write, false);
-      CHECK_EQ(tag, kExternalTagNone);
-      return nullptr;
+  ~ThreadArray() {
+    for (uptr i = 0; i < N; i++) {
+      if (threads[i])
+        Finish(i);
     }
-  };
-  run_in_thread(Thread::Func);
+  }
+
+  void Finish(uptr i) {
+    auto *thr = threads[i];
+    threads[i] = nullptr;
+    Processor *proc = thr->proc();
+    ThreadFinish(thr);
+    ProcUnwire(proc, thr);
+    ProcDestroy(proc);
+    UnmapOrDie(thr, sizeof(ThreadState));
+  }
+
+  ThreadState *threads[N];
+  ThreadState *operator[](uptr i) { return threads[i]; }
+  ThreadState *operator->() { return threads[0]; }
+  operator ThreadState *() { return threads[0]; }
+};
+
+TEST(Trace, RestoreAccess) {
+  // A basic test with some function entry/exit events,
+  // some mutex lock/unlock events and some other distracting
+  // memory events.
+  ThreadArray<1> thr;
+  TraceFunc(thr, 0x1000);
+  TraceFunc(thr, 0x1001);
+  TraceMutexLock(thr, v3::EventType::kLock, 0x4000, 0x5000, 0x6000);
+  TraceMutexLock(thr, v3::EventType::kLock, 0x4001, 0x5001, 0x6001);
+  TraceMutexUnlock(thr, 0x5000);
+  TraceFunc(thr);
+  CHECK(TryTraceMemoryAccess(thr, 0x2001, 0x3001, 8, kAccessRead));
+  TraceMutexLock(thr, v3::EventType::kRLock, 0x4002, 0x5002, 0x6002);
+  TraceFunc(thr, 0x1002);
+  CHECK(TryTraceMemoryAccess(thr, 0x2000, 0x3000, 8, kAccessRead));
+  // This is the access we want to find.
+  // The previous one is equivalent, but RestoreStack must prefer
+  // the last of the matchig accesses.
+  CHECK(TryTraceMemoryAccess(thr, 0x2002, 0x3000, 8, kAccessRead));
+  Lock lock1(&ctx->slot_mtx);
+  ThreadRegistryLock lock2(&ctx->thread_registry);
+  VarSizeStackTrace stk;
+  MutexSet mset;
+  uptr tag = kExternalTagNone;
+  bool res =
+      RestoreStack(thr->tid, v3::EventType::kAccessExt, thr->sid, thr->epoch,
+                   0x3000, 8, kAccessRead, &stk, &mset, &tag);
+  CHECK(res);
+  CHECK_EQ(stk.size, 3);
+  CHECK_EQ(stk.trace[0], 0x1000);
+  CHECK_EQ(stk.trace[1], 0x1002);
+  CHECK_EQ(stk.trace[2], 0x2002);
+  CHECK_EQ(mset.Size(), 2);
+  CHECK_EQ(mset.Get(0).addr, 0x5001);
+  CHECK_EQ(mset.Get(0).stack_id, 0x6001);
+  CHECK_EQ(mset.Get(0).write, true);
+  CHECK_EQ(mset.Get(1).addr, 0x5002);
+  CHECK_EQ(mset.Get(1).stack_id, 0x6002);
+  CHECK_EQ(mset.Get(1).write, false);
+  CHECK_EQ(tag, kExternalTagNone);
 }
 
-TEST(Trace, MAYBE_MemoryAccessSize) {
-  struct Thread {
-    struct Params {
-      uptr access_size, offset, size;
-      bool res;
-      int type;
-    };
-    static void *Func(void *arg) {
-      // Test tracing and matching of accesses of different sizes.
-      const Params *params = static_cast<Params *>(arg);
+TEST(Trace, MemoryAccessSize) {
+  // Test tracing and matching of accesses of different sizes.
+  struct Params {
+    uptr access_size, offset, size;
+    bool res;
+  };
+  Params tests[] = {
+      {1, 0, 1, true},  {4, 0, 2, true},
+      {4, 2, 2, true},  {8, 3, 1, true},
+      {2, 1, 1, true},  {1, 1, 1, false},
+      {8, 5, 4, false}, {4, static_cast<uptr>(-1l), 4, false},
+  };
+  for (auto params : tests) {
+    for (int type = 0; type < 3; type++) {
+      ThreadArray<1> thr;
       Printf("access_size=%zu, offset=%zu, size=%zu, res=%d, type=%d\n",
-             params->access_size, params->offset, params->size, params->res,
-             params->type);
-      ThreadState *thr = cur_thread();
+             params.access_size, params.offset, params.size, params.res, type);
       TraceFunc(thr, 0x1000);
-      switch (params->type) {
+      switch (type) {
         case 0:
           // This should emit compressed event.
-          CHECK(TryTraceMemoryAccess(thr, 0x2000, 0x3000, params->access_size,
+          CHECK(TryTraceMemoryAccess(thr, 0x2000, 0x3000, params.access_size,
                                      kAccessRead));
           break;
         case 1:
           // This should emit full event.
-          CHECK(TryTraceMemoryAccess(thr, 0x2000000, 0x3000,
-                                     params->access_size, kAccessRead));
+          CHECK(TryTraceMemoryAccess(thr, 0x2000000, 0x3000, params.access_size,
+                                     kAccessRead));
           break;
         case 2:
-          TraceMemoryAccessRange(thr, 0x2000000, 0x3000, params->access_size,
+          TraceMemoryAccessRange(thr, 0x2000000, 0x3000, params.access_size,
                                  kAccessRead);
           break;
       }
@@ -127,105 +140,82 @@ TEST(Trace, MAYBE_MemoryAccessSize) {
       MutexSet mset;
       uptr tag = kExternalTagNone;
       bool res = RestoreStack(thr->tid, v3::EventType::kAccessExt, thr->sid,
-                              thr->epoch, 0x3000 + params->offset, params->size,
+                              thr->epoch, 0x3000 + params.offset, params.size,
                               kAccessRead, &stk, &mset, &tag);
-      CHECK_EQ(res, params->res);
-      if (params->res) {
+      CHECK_EQ(res, params.res);
+      if (params.res) {
         CHECK_EQ(stk.size, 2);
         CHECK_EQ(stk.trace[0], 0x1000);
-        CHECK_EQ(stk.trace[1], params->type ? 0x2000000 : 0x2000);
+        CHECK_EQ(stk.trace[1], type ? 0x2000000 : 0x2000);
       }
-      return nullptr;
     }
-  };
-  Thread::Params tests[] = {
-      {1, 0, 1, true, 0},  {4, 0, 2, true, 0},
-      {4, 2, 2, true, 0},  {8, 3, 1, true, 0},
-      {2, 1, 1, true, 0},  {1, 1, 1, false, 0},
-      {8, 5, 4, false, 0}, {4, static_cast<uptr>(-1l), 4, false, 0},
-  };
-  for (auto params : tests) {
-    for (params.type = 0; params.type < 3; params.type++)
-      run_in_thread(Thread::Func, &params);
   }
 }
 
-TEST(Trace, MAYBE_RestoreMutexLock) {
-  struct Thread {
-    static void *Func(void *arg) {
-      // Check of restoration of a mutex lock event.
-      ThreadState *thr = cur_thread();
-      TraceFunc(thr, 0x1000);
-      TraceMutexLock(thr, v3::EventType::kLock, 0x4000, 0x5000, 0x6000);
-      TraceMutexLock(thr, v3::EventType::kRLock, 0x4001, 0x5001, 0x6001);
-      TraceMutexLock(thr, v3::EventType::kRLock, 0x4002, 0x5001, 0x6002);
-      Lock lock1(&ctx->slot_mtx);
-      ThreadRegistryLock lock2(&ctx->thread_registry);
-      VarSizeStackTrace stk;
-      MutexSet mset;
-      uptr tag = kExternalTagNone;
-      bool res = RestoreStack(thr->tid, v3::EventType::kLock, thr->sid,
-                              thr->epoch, 0x5001, 0, 0, &stk, &mset, &tag);
-      CHECK(res);
-      CHECK_EQ(stk.size, 2);
-      CHECK_EQ(stk.trace[0], 0x1000);
-      CHECK_EQ(stk.trace[1], 0x4002);
-      CHECK_EQ(mset.Size(), 2);
-      CHECK_EQ(mset.Get(0).addr, 0x5000);
-      CHECK_EQ(mset.Get(0).stack_id, 0x6000);
-      CHECK_EQ(mset.Get(0).write, true);
-      CHECK_EQ(mset.Get(1).addr, 0x5001);
-      CHECK_EQ(mset.Get(1).stack_id, 0x6001);
-      CHECK_EQ(mset.Get(1).write, false);
-      return nullptr;
-    }
-  };
-  run_in_thread(Thread::Func);
+TEST(Trace, RestoreMutexLock) {
+  // Check of restoration of a mutex lock event.
+  ThreadArray<1> thr;
+  TraceFunc(thr, 0x1000);
+  TraceMutexLock(thr, v3::EventType::kLock, 0x4000, 0x5000, 0x6000);
+  TraceMutexLock(thr, v3::EventType::kRLock, 0x4001, 0x5001, 0x6001);
+  TraceMutexLock(thr, v3::EventType::kRLock, 0x4002, 0x5001, 0x6002);
+  Lock lock1(&ctx->slot_mtx);
+  ThreadRegistryLock lock2(&ctx->thread_registry);
+  VarSizeStackTrace stk;
+  MutexSet mset;
+  uptr tag = kExternalTagNone;
+  bool res = RestoreStack(thr->tid, v3::EventType::kLock, thr->sid, thr->epoch,
+                          0x5001, 0, 0, &stk, &mset, &tag);
+  CHECK(res);
+  CHECK_EQ(stk.size, 2);
+  CHECK_EQ(stk.trace[0], 0x1000);
+  CHECK_EQ(stk.trace[1], 0x4002);
+  CHECK_EQ(mset.Size(), 2);
+  CHECK_EQ(mset.Get(0).addr, 0x5000);
+  CHECK_EQ(mset.Get(0).stack_id, 0x6000);
+  CHECK_EQ(mset.Get(0).write, true);
+  CHECK_EQ(mset.Get(1).addr, 0x5001);
+  CHECK_EQ(mset.Get(1).stack_id, 0x6001);
+  CHECK_EQ(mset.Get(1).write, false);
 }
 
-TEST(Trace, MAYBE_MultiPart) {
-  struct Thread {
-    static void *Func(void *arg) {
-      // Check replay of a trace with multiple parts.
-      ThreadState *thr = cur_thread();
-      TraceFunc(thr, 0x1000);
-      TraceFunc(thr, 0x2000);
-      TraceMutexLock(thr, v3::EventType::kLock, 0x4000, 0x5000, 0x6000);
-      const uptr kEvents = 3 * sizeof(TracePart) / sizeof(v3::Event);
-      for (uptr i = 0; i < kEvents; i++) {
-        TraceFunc(thr, 0x3000);
-        TraceMutexLock(thr, v3::EventType::kLock, 0x4002, 0x5002, 0x6002);
-        TraceMutexUnlock(thr, 0x5002);
-        TraceFunc(thr);
-      }
-      TraceFunc(thr, 0x4000);
-      TraceMutexLock(thr, v3::EventType::kRLock, 0x4001, 0x5001, 0x6001);
-      CHECK(TryTraceMemoryAccess(thr, 0x2002, 0x3000, 8, kAccessRead));
-      Lock lock1(&ctx->slot_mtx);
-      ThreadRegistryLock lock2(&ctx->thread_registry);
-      VarSizeStackTrace stk;
-      MutexSet mset;
-      uptr tag = kExternalTagNone;
-      bool res =
-          RestoreStack(thr->tid, v3::EventType::kAccessExt, thr->sid,
-                       thr->epoch, 0x3000, 8, kAccessRead, &stk, &mset, &tag);
-      CHECK(res);
-      CHECK_EQ(stk.size, 4);
-      CHECK_EQ(stk.trace[0], 0x1000);
-      CHECK_EQ(stk.trace[1], 0x2000);
-      CHECK_EQ(stk.trace[2], 0x4000);
-      CHECK_EQ(stk.trace[3], 0x2002);
-      CHECK_EQ(mset.Size(), 2);
-      CHECK_EQ(mset.Get(0).addr, 0x5000);
-      CHECK_EQ(mset.Get(0).stack_id, 0x6000);
-      CHECK_EQ(mset.Get(0).write, true);
-      CHECK_EQ(mset.Get(1).addr, 0x5001);
-      CHECK_EQ(mset.Get(1).stack_id, 0x6001);
-      CHECK_EQ(mset.Get(1).write, false);
-      return nullptr;
-    }
-  };
-  run_in_thread(Thread::Func);
+TEST(Trace, MultiPart) {
+  // Check replay of a trace with multiple parts.
+  ThreadArray<1> thr;
+  TraceFunc(thr, 0x1000);
+  TraceFunc(thr, 0x2000);
+  TraceMutexLock(thr, v3::EventType::kLock, 0x4000, 0x5000, 0x6000);
+  const uptr kEvents = 3 * sizeof(TracePart) / sizeof(v3::Event);
+  for (uptr i = 0; i < kEvents; i++) {
+    TraceFunc(thr, 0x3000);
+    TraceMutexLock(thr, v3::EventType::kLock, 0x4002, 0x5002, 0x6002);
+    TraceMutexUnlock(thr, 0x5002);
+    TraceFunc(thr);
+  }
+  TraceFunc(thr, 0x4000);
+  TraceMutexLock(thr, v3::EventType::kRLock, 0x4001, 0x5001, 0x6001);
+  CHECK(TryTraceMemoryAccess(thr, 0x2002, 0x3000, 8, kAccessRead));
+  Lock lock1(&ctx->slot_mtx);
+  ThreadRegistryLock lock2(&ctx->thread_registry);
+  VarSizeStackTrace stk;
+  MutexSet mset;
+  uptr tag = kExternalTagNone;
+  bool res =
+      RestoreStack(thr->tid, v3::EventType::kAccessExt, thr->sid, thr->epoch,
+                   0x3000, 8, kAccessRead, &stk, &mset, &tag);
+  CHECK(res);
+  CHECK_EQ(stk.size, 4);
+  CHECK_EQ(stk.trace[0], 0x1000);
+  CHECK_EQ(stk.trace[1], 0x2000);
+  CHECK_EQ(stk.trace[2], 0x4000);
+  CHECK_EQ(stk.trace[3], 0x2002);
+  CHECK_EQ(mset.Size(), 2);
+  CHECK_EQ(mset.Get(0).addr, 0x5000);
+  CHECK_EQ(mset.Get(0).stack_id, 0x6000);
+  CHECK_EQ(mset.Get(0).write, true);
+  CHECK_EQ(mset.Get(1).addr, 0x5001);
+  CHECK_EQ(mset.Get(1).stack_id, 0x6001);
+  CHECK_EQ(mset.Get(1).write, false);
 }
 
 }  // namespace __tsan
diff --git a/compiler-rt/test/tsan/stack_race3.cpp b/compiler-rt/test/tsan/stack_race3.cpp
@@ -0,0 +1,34 @@
+// RUN: %clangxx_tsan -O1 %s -o %t && %deflake %run %t | FileCheck %s
+
+// Race with initial stack initialization:
+// there is no explicit second write,
+// but the stack variable is published unsafely.
+#include "test.h"
+
+long *P;
+
+void *Thread(void *a) {
+  long X;
+  __atomic_store_n(&P, &X, __ATOMIC_RELAXED);
+  barrier_wait(&barrier);
+  barrier_wait(&barrier);
+  return 0;
+}
+
+int main() {
+  barrier_init(&barrier, 2);
+  pthread_t t;
+  pthread_create(&t, NULL, Thread, NULL);
+  barrier_wait(&barrier);
+  long *p = __atomic_load_n(&P, __ATOMIC_RELAXED);
+  *p = 42;
+  barrier_wait(&barrier);
+  pthread_join(t, 0);
+}
+
+// CHECK: WARNING: ThreadSanitizer: data race
+// CHECK:   Write of size 8 at {{.*}} by main thread:
+// CHECK:     #0 main
+// CHECK:   Previous write of size 8 at {{.*}} by thread T1:
+// CHECK:     #0 __tsan_stack_initialization
+// CHECK:   Location is stack of thread T1
diff --git a/compiler-rt/test/tsan/tls_race3.cpp b/compiler-rt/test/tsan/tls_race3.cpp
@@ -0,0 +1,35 @@
+// RUN: %clangxx_tsan %darwin_min_target_with_tls_support -O1 %s -o %t && \
+// RUN:   %deflake %run %t | FileCheck %s
+
+// Race with initial TLS initialization:
+// there is no explicit second write,
+// but the TLS variable is published unsafely.
+#include "test.h"
+
+__thread long X;
+long *P;
+
+void *Thread(void *a) {
+  __atomic_store_n(&P, &X, __ATOMIC_RELAXED);
+  barrier_wait(&barrier);
+  barrier_wait(&barrier);
+  return 0;
+}
+
+int main() {
+  barrier_init(&barrier, 2);
+  pthread_t t;
+  pthread_create(&t, NULL, Thread, NULL);
+  barrier_wait(&barrier);
+  long *p = __atomic_load_n(&P, __ATOMIC_RELAXED);
+  *p = 42;
+  barrier_wait(&barrier);
+  pthread_join(t, 0);
+}
+
+// CHECK: WARNING: ThreadSanitizer: data race
+// CHECK:   Write of size 8 at {{.*}} by main thread:
+// CHECK:     #0 main
+// CHECK:   Previous write of size 8 at {{.*}} by thread T1:
+// CHECK:     #0 __tsan_tls_initialization
+// CHECK:   Location is TLS of thread T1.
diff --git a/flang/include/flang/Semantics/expression.h b/flang/include/flang/Semantics/expression.h
@@ -105,9 +105,11 @@ class ExpressionAnalyzer {
   explicit ExpressionAnalyzer(semantics::SemanticsContext &sc) : context_{sc} {}
   ExpressionAnalyzer(semantics::SemanticsContext &sc, FoldingContext &fc)
       : context_{sc}, foldingContext_{fc} {}
-  ExpressionAnalyzer(ExpressionAnalyzer &) = default;
+  ExpressionAnalyzer(const ExpressionAnalyzer &) = default;
 
   semantics::SemanticsContext &context() const { return context_; }
+  bool inWhereBody() const { return inWhereBody_; }
+  void set_inWhereBody(bool yes = true) { inWhereBody_ = yes; }
 
   FoldingContext &GetFoldingContext() const { return foldingContext_; }
 
@@ -366,6 +368,7 @@ class ExpressionAnalyzer {
   std::map<parser::CharBlock, int> impliedDos_; // values are INTEGER kinds
   bool isWholeAssumedSizeArrayOk_{false};
   bool useSavedTypedExprs_{true};
+  bool inWhereBody_{false};
   friend class ArgumentAnalyzer;
 };
 
@@ -402,12 +405,6 @@ evaluate::Expr<evaluate::SubscriptInteger> AnalyzeKindSelector(
     SemanticsContext &, common::TypeCategory,
     const std::optional<parser::KindSelector> &);
 
-void AnalyzeCallStmt(SemanticsContext &, const parser::CallStmt &);
-const evaluate::Assignment *AnalyzeAssignmentStmt(
-    SemanticsContext &, const parser::AssignmentStmt &);
-const evaluate::Assignment *AnalyzePointerAssignmentStmt(
-    SemanticsContext &, const parser::PointerAssignmentStmt &);
-
 // Semantic analysis of all expressions in a parse tree, which becomes
 // decorated with typed representations for top-level expressions.
 class ExprChecker {
@@ -445,18 +442,38 @@ class ExprChecker {
   bool Pre(const parser::DataImpliedDo &);
 
   bool Pre(const parser::CallStmt &x) {
-    AnalyzeCallStmt(context_, x);
+    exprAnalyzer_.Analyze(x);
     return false;
   }
   bool Pre(const parser::AssignmentStmt &x) {
-    AnalyzeAssignmentStmt(context_, x);
+    exprAnalyzer_.Analyze(x);
     return false;
   }
   bool Pre(const parser::PointerAssignmentStmt &x) {
-    AnalyzePointerAssignmentStmt(context_, x);
+    exprAnalyzer_.Analyze(x);
     return false;
   }
 
+  // Track whether we're in a WHERE statement or construct body
+  bool Pre(const parser::WhereStmt &) {
+    ++whereDepth_;
+    exprAnalyzer_.set_inWhereBody(InWhereBody());
+    return true;
+  }
+  void Post(const parser::WhereStmt &) {
+    --whereDepth_;
+    exprAnalyzer_.set_inWhereBody(InWhereBody());
+  }
+  bool Pre(const parser::WhereBodyConstruct &) {
+    ++whereDepth_;
+    exprAnalyzer_.set_inWhereBody(InWhereBody());
+    return true;
+  }
+  void Post(const parser::WhereBodyConstruct &) {
+    --whereDepth_;
+    exprAnalyzer_.set_inWhereBody(InWhereBody());
+  }
+
   template <typename A> bool Pre(const parser::Scalar<A> &x) {
     exprAnalyzer_.Analyze(x);
     return false;
@@ -479,8 +496,11 @@ class ExprChecker {
   }
 
 private:
+  bool InWhereBody() const { return whereDepth_ > 0; }
+
   SemanticsContext &context_;
   evaluate::ExpressionAnalyzer exprAnalyzer_{context_};
+  int whereDepth_{0}; // nesting of WHERE statements & constructs
 };
 } // namespace Fortran::semantics
 #endif // FORTRAN_SEMANTICS_EXPRESSION_H_