diff --git a/clang/include/clang/Basic/Builtins.def b/clang/include/clang/Basic/Builtins.def
index d2cb14d2fd8cc..a17d2e59f5d8b 100644
--- a/clang/include/clang/Basic/Builtins.def
+++ b/clang/include/clang/Basic/Builtins.def
@@ -1625,50 +1625,50 @@ LANGBUILTIN(__builtin_coro_suspend, "cIb", "n", COR_LANG)
 
 // OpenCL v2.0 s6.13.16, s9.17.3.5 - Pipe functions.
 // We need the generic prototype, since the packet type could be anything.
-LANGBUILTIN(read_pipe, "i.", "tn", OCLC20_LANG)
-LANGBUILTIN(write_pipe, "i.", "tn", OCLC20_LANG)
+LANGBUILTIN(read_pipe, "i.", "tn", OCL_PIPE)
+LANGBUILTIN(write_pipe, "i.", "tn", OCL_PIPE)
 
-LANGBUILTIN(reserve_read_pipe, "i.", "tn", OCLC20_LANG)
-LANGBUILTIN(reserve_write_pipe, "i.", "tn", OCLC20_LANG)
+LANGBUILTIN(reserve_read_pipe, "i.", "tn", OCL_PIPE)
+LANGBUILTIN(reserve_write_pipe, "i.", "tn", OCL_PIPE)
 
-LANGBUILTIN(commit_write_pipe, "v.", "tn", OCLC20_LANG)
-LANGBUILTIN(commit_read_pipe, "v.", "tn", OCLC20_LANG)
+LANGBUILTIN(commit_write_pipe, "v.", "tn", OCL_PIPE)
+LANGBUILTIN(commit_read_pipe, "v.", "tn", OCL_PIPE)
 
-LANGBUILTIN(sub_group_reserve_read_pipe, "i.", "tn", OCLC20_LANG)
-LANGBUILTIN(sub_group_reserve_write_pipe, "i.", "tn", OCLC20_LANG)
+LANGBUILTIN(sub_group_reserve_read_pipe, "i.", "tn", OCL_PIPE)
+LANGBUILTIN(sub_group_reserve_write_pipe, "i.", "tn", OCL_PIPE)
 
-LANGBUILTIN(sub_group_commit_read_pipe, "v.", "tn", OCLC20_LANG)
-LANGBUILTIN(sub_group_commit_write_pipe, "v.", "tn", OCLC20_LANG)
+LANGBUILTIN(sub_group_commit_read_pipe, "v.", "tn", OCL_PIPE)
+LANGBUILTIN(sub_group_commit_write_pipe, "v.", "tn", OCL_PIPE)
 
-LANGBUILTIN(work_group_reserve_read_pipe, "i.", "tn", OCLC20_LANG)
-LANGBUILTIN(work_group_reserve_write_pipe, "i.", "tn", OCLC20_LANG)
+LANGBUILTIN(work_group_reserve_read_pipe, "i.", "tn", OCL_PIPE)
+LANGBUILTIN(work_group_reserve_write_pipe, "i.", "tn", OCL_PIPE)
 
-LANGBUILTIN(work_group_commit_read_pipe, "v.", "tn", OCLC20_LANG)
-LANGBUILTIN(work_group_commit_write_pipe, "v.", "tn", OCLC20_LANG)
+LANGBUILTIN(work_group_commit_read_pipe, "v.", "tn", OCL_PIPE)
+LANGBUILTIN(work_group_commit_write_pipe, "v.", "tn", OCL_PIPE)
 
-LANGBUILTIN(get_pipe_num_packets, "Ui.", "tn", OCLC20_LANG)
-LANGBUILTIN(get_pipe_max_packets, "Ui.", "tn", OCLC20_LANG)
+LANGBUILTIN(get_pipe_num_packets, "Ui.", "tn", OCL_PIPE)
+LANGBUILTIN(get_pipe_max_packets, "Ui.", "tn", OCL_PIPE)
 
 // OpenCL v2.0 s6.13.17 - Enqueue kernel functions.
 // Custom builtin check allows to perform special check of passed block arguments.
-LANGBUILTIN(enqueue_kernel, "i.", "tn", OCLC20_LANG)
-LANGBUILTIN(get_kernel_work_group_size, "Ui.", "tn", OCLC20_LANG)
-LANGBUILTIN(get_kernel_preferred_work_group_size_multiple, "Ui.", "tn", OCLC20_LANG)
-LANGBUILTIN(get_kernel_max_sub_group_size_for_ndrange, "Ui.", "tn", OCLC20_LANG)
-LANGBUILTIN(get_kernel_sub_group_count_for_ndrange, "Ui.", "tn", OCLC20_LANG)
+LANGBUILTIN(enqueue_kernel, "i.", "tn", OCL_DSE)
+LANGBUILTIN(get_kernel_work_group_size, "Ui.", "tn", OCL_DSE)
+LANGBUILTIN(get_kernel_preferred_work_group_size_multiple, "Ui.", "tn", OCL_DSE)
+LANGBUILTIN(get_kernel_max_sub_group_size_for_ndrange, "Ui.", "tn", OCL_DSE)
+LANGBUILTIN(get_kernel_sub_group_count_for_ndrange, "Ui.", "tn", OCL_DSE)
 
 // OpenCL v2.0 s6.13.9 - Address space qualifier functions.
 // FIXME: Pointer parameters of OpenCL builtins should have their address space
 // requirement defined.
-LANGBUILTIN(to_global, "v*v*", "tn", OCLC20_LANG)
-LANGBUILTIN(to_local, "v*v*", "tn", OCLC20_LANG)
-LANGBUILTIN(to_private, "v*v*", "tn", OCLC20_LANG)
+LANGBUILTIN(to_global, "v*v*", "tn", OCL_GAS)
+LANGBUILTIN(to_local, "v*v*", "tn", OCL_GAS)
+LANGBUILTIN(to_private, "v*v*", "tn", OCL_GAS)
 
 // OpenCL half load/store builtin
-LANGBUILTIN(__builtin_store_half, "vdh*", "n", ALL_OCLC_LANGUAGES)
-LANGBUILTIN(__builtin_store_halff, "vfh*", "n", ALL_OCLC_LANGUAGES)
-LANGBUILTIN(__builtin_load_half, "dhC*", "nc", ALL_OCLC_LANGUAGES)
-LANGBUILTIN(__builtin_load_halff, "fhC*", "nc", ALL_OCLC_LANGUAGES)
+LANGBUILTIN(__builtin_store_half, "vdh*", "n", ALL_OCL_LANGUAGES)
+LANGBUILTIN(__builtin_store_halff, "vfh*", "n", ALL_OCL_LANGUAGES)
+LANGBUILTIN(__builtin_load_half, "dhC*", "nc", ALL_OCL_LANGUAGES)
+LANGBUILTIN(__builtin_load_halff, "fhC*", "nc", ALL_OCL_LANGUAGES)
 
 // Builtins for os_log/os_trace
 BUILTIN(__builtin_os_log_format_buffer_size, "zcC*.", "p:0:nut")
diff --git a/clang/include/clang/Basic/Builtins.h b/clang/include/clang/Basic/Builtins.h
index cdaaee48c32d8..1dabafce54f39 100644
--- a/clang/include/clang/Basic/Builtins.h
+++ b/clang/include/clang/Basic/Builtins.h
@@ -28,20 +28,21 @@ class IdentifierTable;
 class LangOptions;
 
 enum LanguageID {
-  GNU_LANG = 0x1,     // builtin requires GNU mode.
-  C_LANG = 0x2,       // builtin for c only.
-  CXX_LANG = 0x4,     // builtin for cplusplus only.
-  OBJC_LANG = 0x8,    // builtin for objective-c and objective-c++
-  MS_LANG = 0x10,     // builtin requires MS mode.
-  OCLC20_LANG = 0x20, // builtin for OpenCL C 2.0 only.
-  OCLC1X_LANG = 0x40, // builtin for OpenCL C 1.x only.
-  OMP_LANG = 0x80,    // builtin requires OpenMP.
-  CUDA_LANG = 0x100,  // builtin requires CUDA.
-  COR_LANG = 0x200,   // builtin requires use of 'fcoroutine-ts' option.
+  GNU_LANG = 0x1,            // builtin requires GNU mode.
+  C_LANG = 0x2,              // builtin for c only.
+  CXX_LANG = 0x4,            // builtin for cplusplus only.
+  OBJC_LANG = 0x8,           // builtin for objective-c and objective-c++
+  MS_LANG = 0x10,            // builtin requires MS mode.
+  OMP_LANG = 0x20,           // builtin requires OpenMP.
+  CUDA_LANG = 0x40,          // builtin requires CUDA.
+  COR_LANG = 0x80,           // builtin requires use of 'fcoroutine-ts' option.
+  OCL_GAS = 0x100,           // builtin requires OpenCL generic address space.
+  OCL_PIPE = 0x200,          // builtin requires OpenCL pipe.
+  OCL_DSE = 0x400,           // builtin requires OpenCL device side enqueue.
+  ALL_OCL_LANGUAGES = 0x800, // builtin for OCL languages.
   ALL_LANGUAGES = C_LANG | CXX_LANG | OBJC_LANG, // builtin for all languages.
   ALL_GNU_LANGUAGES = ALL_LANGUAGES | GNU_LANG,  // builtin requires GNU mode.
-  ALL_MS_LANGUAGES = ALL_LANGUAGES | MS_LANG,    // builtin requires MS mode.
-  ALL_OCLC_LANGUAGES = OCLC1X_LANG | OCLC20_LANG // builtin for OCLC languages.
+  ALL_MS_LANGUAGES = ALL_LANGUAGES | MS_LANG     // builtin requires MS mode.
 };
 
 namespace Builtin {
diff --git a/clang/lib/Basic/Builtins.cpp b/clang/lib/Basic/Builtins.cpp
index 2b0f4071662c0..6d278e9c4a229 100644
--- a/clang/lib/Basic/Builtins.cpp
+++ b/clang/lib/Basic/Builtins.cpp
@@ -69,22 +69,26 @@ bool Builtin::Context::builtinIsSupported(const Builtin::Info &BuiltinInfo,
   bool MSModeUnsupported =
       !LangOpts.MicrosoftExt && (BuiltinInfo.Langs & MS_LANG);
   bool ObjCUnsupported = !LangOpts.ObjC && BuiltinInfo.Langs == OBJC_LANG;
-  bool OclC1Unsupported = (LangOpts.OpenCLVersion / 100) != 1 &&
-                          (BuiltinInfo.Langs & ALL_OCLC_LANGUAGES ) ==  OCLC1X_LANG;
-  bool OclC2Unsupported =
-      (LangOpts.getOpenCLCompatibleVersion() != 200) &&
-      (BuiltinInfo.Langs & ALL_OCLC_LANGUAGES) == OCLC20_LANG;
-  bool OclCUnsupported = !LangOpts.OpenCL &&
-                         (BuiltinInfo.Langs & ALL_OCLC_LANGUAGES);
+  bool OclCUnsupported =
+      !LangOpts.OpenCL && (BuiltinInfo.Langs & ALL_OCL_LANGUAGES);
+  bool OclGASUnsupported =
+      !LangOpts.OpenCLGenericAddressSpace && (BuiltinInfo.Langs & OCL_GAS);
+  bool OclPipeUnsupported =
+      !LangOpts.OpenCLPipes && (BuiltinInfo.Langs & OCL_PIPE);
+  // Device side enqueue is not supported until OpenCL 2.0. In 2.0 and higher
+  // support is indicated with language option for blocks.
+  bool OclDSEUnsupported =
+      (LangOpts.getOpenCLCompatibleVersion() < 200 || !LangOpts.Blocks) &&
+      (BuiltinInfo.Langs & OCL_DSE);
   bool OpenMPUnsupported = !LangOpts.OpenMP && BuiltinInfo.Langs == OMP_LANG;
   bool CUDAUnsupported = !LangOpts.CUDA && BuiltinInfo.Langs == CUDA_LANG;
   bool CPlusPlusUnsupported =
       !LangOpts.CPlusPlus && BuiltinInfo.Langs == CXX_LANG;
   return !BuiltinsUnsupported && !CorBuiltinsUnsupported &&
-         !MathBuiltinsUnsupported && !OclCUnsupported && !OclC1Unsupported &&
-         !OclC2Unsupported && !OpenMPUnsupported && !GnuModeUnsupported &&
-         !MSModeUnsupported && !ObjCUnsupported && !CPlusPlusUnsupported &&
-         !CUDAUnsupported;
+         !MathBuiltinsUnsupported && !OclCUnsupported && !OclGASUnsupported &&
+         !OclPipeUnsupported && !OclDSEUnsupported && !OpenMPUnsupported &&
+         !GnuModeUnsupported && !MSModeUnsupported && !ObjCUnsupported &&
+         !CPlusPlusUnsupported && !CUDAUnsupported;
 }
 
 /// initializeBuiltins - Mark the identifiers for all the builtins with their
diff --git a/clang/lib/Driver/ToolChains/MSVC.cpp b/clang/lib/Driver/ToolChains/MSVC.cpp
index 18cef288f018a..9f4751167ac12 100644
--- a/clang/lib/Driver/ToolChains/MSVC.cpp
+++ b/clang/lib/Driver/ToolChains/MSVC.cpp
@@ -1393,8 +1393,8 @@ VersionTuple MSVCToolChain::computeMSVCVersion(const Driver *D,
   if (MSVT.empty() &&
       Args.hasFlag(options::OPT_fms_extensions, options::OPT_fno_ms_extensions,
                    IsWindowsMSVC)) {
-    // -fms-compatibility-version=19.14 is default, aka 2017, 15.7
-    MSVT = VersionTuple(19, 14);
+    // -fms-compatibility-version=19.20 is default, aka 2019, 16.x
+    MSVT = VersionTuple(19, 20);
   }
   return MSVT;
 }
diff --git a/clang/lib/Sema/SemaChecking.cpp b/clang/lib/Sema/SemaChecking.cpp
index dfbf4cdc89cb9..69dcc3aaaaf37 100644
--- a/clang/lib/Sema/SemaChecking.cpp
+++ b/clang/lib/Sema/SemaChecking.cpp
@@ -1042,9 +1042,15 @@ static bool checkOpenCLBlockArgs(Sema &S, Expr *BlockArg) {
 }
 
 static bool checkOpenCLSubgroupExt(Sema &S, CallExpr *Call) {
-  if (!S.getOpenCLOptions().isSupported("cl_khr_subgroups", S.getLangOpts())) {
+  // OpenCL device can support extension but not the feature as extension
+  // requires subgroup independent forward progress, but subgroup independent
+  // forward progress is optional in OpenCL C 3.0 __opencl_c_subgroups feature.
+  if (!S.getOpenCLOptions().isSupported("cl_khr_subgroups", S.getLangOpts()) &&
+      !S.getOpenCLOptions().isSupported("__opencl_c_subgroups",
+                                        S.getLangOpts())) {
     S.Diag(Call->getBeginLoc(), diag::err_opencl_requires_extension)
-        << 1 << Call->getDirectCallee() << "cl_khr_subgroups";
+        << 1 << Call->getDirectCallee()
+        << "cl_khr_subgroups or __opencl_c_subgroups";
     return true;
   }
   return false;
diff --git a/clang/test/CodeGenOpenCL/cl20-device-side-enqueue.cl b/clang/test/CodeGenOpenCL/cl20-device-side-enqueue.cl
index fdbc0aaa6e263..01750804884b3 100644
--- a/clang/test/CodeGenOpenCL/cl20-device-side-enqueue.cl
+++ b/clang/test/CodeGenOpenCL/cl20-device-side-enqueue.cl
@@ -1,6 +1,9 @@
 // RUN: %clang_cc1 -disable-noundef-analysis %s -cl-std=CL2.0 -ffake-address-space-map -O0 -emit-llvm -o - -triple "spir-unknown-unknown" | FileCheck %s --check-prefix=COMMON --check-prefix=B32
 // RUN: %clang_cc1 -disable-noundef-analysis %s -cl-std=CL2.0 -ffake-address-space-map -O0 -emit-llvm -o - -triple "spir64-unknown-unknown" | FileCheck %s --check-prefix=COMMON --check-prefix=B64
 // RUN: %clang_cc1 -disable-noundef-analysis %s -cl-std=CL2.0 -ffake-address-space-map -O1 -emit-llvm -o - -triple "spir64-unknown-unknown" | FileCheck %s --check-prefix=CHECK-LIFETIMES
+// RUN: %clang_cc1 -disable-noundef-analysis %s -cl-std=CL3.0 -ffake-address-space-map -O0 -emit-llvm -o - -triple "spir-unknown-unknown" | FileCheck %s --check-prefix=COMMON --check-prefix=B32
+// RUN: %clang_cc1 -disable-noundef-analysis %s -cl-std=CL3.0 -ffake-address-space-map -O0 -emit-llvm -o - -triple "spir64-unknown-unknown" | FileCheck %s --check-prefix=COMMON --check-prefix=B64
+// RUN: %clang_cc1 -disable-noundef-analysis %s -cl-std=CL3.0 -ffake-address-space-map -O1 -emit-llvm -o - -triple "spir64-unknown-unknown" | FileCheck %s --check-prefix=CHECK-LIFETIMES
 
 #pragma OPENCL EXTENSION cl_khr_subgroups : enable
 
diff --git a/clang/test/CodeGenOpenCL/pipe_types.cl b/clang/test/CodeGenOpenCL/pipe_types.cl
index 57da95fec0d70..9644aff1a6248 100644
--- a/clang/test/CodeGenOpenCL/pipe_types.cl
+++ b/clang/test/CodeGenOpenCL/pipe_types.cl
@@ -1,4 +1,4 @@
-// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -emit-llvm -O0 -cl-std=CL2.0 -DTEST_STRUCT -o - %s | FileCheck --check-prefixes=CHECK,CHECK-STRUCT %s
+// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -emit-llvm -O0 -cl-std=CL2.0 -o - %s | FileCheck --check-prefixes=CHECK,CHECK-STRUCT %s
 // RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -emit-llvm -O0 -cl-std=CL3.0 -cl-ext=-all,+__opencl_c_pipes,+__opencl_c_generic_address_space,+__opencl_c_program_scope_global_variables -o - %s | FileCheck --check-prefixes=CHECK %s
 // RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -emit-llvm -O0 -cl-std=CL3.0 -cl-ext=-all,+__opencl_c_pipes,+__opencl_c_generic_address_space -o - %s | FileCheck --check-prefixes=CHECK %s
 // RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -emit-llvm -O0 -cl-std=clc++2021 -cl-ext=-all,+__opencl_c_pipes,+__opencl_c_generic_address_space,+__opencl_c_program_scope_global_variables -o - %s | FileCheck --check-prefixes=CHECK %s
@@ -36,8 +36,6 @@ kernel void test6(MyPipe p) {
 // CHECK: define{{.*}} spir_kernel void @test6(%opencl.pipe_ro_t* %p)
 }
 
-#ifdef TEST_STRUCT
-// FIXME: not supported for OpenCL C 3.0 as language built-ins not supported yet
 struct Person {
   const char *Name;
   bool isFemale;
@@ -52,4 +50,3 @@ void test_reserved_read_pipe(global struct Person *SDst,
   read_pipe (SPipe, SDst);
   // CHECK-STRUCT: call i32 @__read_pipe_2(%opencl.pipe_ro_t* %{{.*}}, i8* %{{.*}}, i32 16, i32 8)
 }
-#endif // TEST_STRUCT
diff --git a/clang/test/CodeGenOpenCL/to_addr_builtin.cl b/clang/test/CodeGenOpenCL/to_addr_builtin.cl
index b5b197079b74b..03097f60f7931 100644
--- a/clang/test/CodeGenOpenCL/to_addr_builtin.cl
+++ b/clang/test/CodeGenOpenCL/to_addr_builtin.cl
@@ -1,4 +1,6 @@
 // RUN: %clang_cc1 -triple spir-unknown-unknown -emit-llvm -O0 -cl-std=clc++ -o - %s | FileCheck %s
+// RUN: %clang_cc1 -triple spir-unknown-unknown -emit-llvm -O0 -cl-std=cl2.0 -o - %s | FileCheck %s
+// RUN: %clang_cc1 -triple spir-unknown-unknown -emit-llvm -O0 -cl-std=cl3.0 -o - %s | FileCheck %s
 
 // CHECK: %[[A:.*]] = type { float, float, float }
 typedef struct {
diff --git a/clang/test/Driver/cl-options.c b/clang/test/Driver/cl-options.c
index 733e733de738e..8d240e6e6e5c6 100644
--- a/clang/test/Driver/cl-options.c
+++ b/clang/test/Driver/cl-options.c
@@ -757,7 +757,7 @@
 
 // Validate that the default triple is used when run an empty tools dir is specified
 // RUN: %clang_cl -vctoolsdir "" -### -- %s 2>&1 | FileCheck %s --check-prefix VCTOOLSDIR
-// VCTOOLSDIR: "-triple" "{{[a-zA-Z0-9_-]*}}-pc-windows-msvc19.14.0"
+// VCTOOLSDIR: "-triple" "{{[a-zA-Z0-9_-]*}}-pc-windows-msvc19.20.0"
 
 // Validate that built-in include paths are based on the supplied path
 // RUN: %clang_cl --target=aarch64-pc-windows-msvc -vctoolsdir "/fake" -winsdkdir "/foo" -winsdkversion 10.0.12345.0 -### -- %s 2>&1 | FileCheck %s --check-prefix FAKEDIR
diff --git a/clang/test/SemaOpenCL/cl20-device-side-enqueue.cl b/clang/test/SemaOpenCL/cl20-device-side-enqueue.cl
index 5cfdfb3cdc5ca..344297ebbe0ca 100644
--- a/clang/test/SemaOpenCL/cl20-device-side-enqueue.cl
+++ b/clang/test/SemaOpenCL/cl20-device-side-enqueue.cl
@@ -3,6 +3,11 @@
 // RUN: %clang_cc1 %s -cl-std=CL2.0 -triple "spir-unknown-unknown" -verify -pedantic -fsyntax-only -DB32 -DQUALS="const volatile"
 // RUN: %clang_cc1 %s -cl-std=CL2.0 -triple "spir64-unknown-unknown" -verify -pedantic -fsyntax-only -Wconversion -DWCONV -DQUALS=
 // RUN: %clang_cc1 %s -cl-std=CL2.0 -triple "spir64-unknown-unknown" -verify -pedantic -fsyntax-only -Wconversion -DWCONV -DQUALS="const volatile"
+// RUN: %clang_cc1 %s -cl-std=CL3.0 -triple "spir-unknown-unknown" -verify -pedantic -fsyntax-only -DB32 -DQUALS=
+// RUN: %clang_cc1 %s -cl-std=CL3.0 -triple "spir-unknown-unknown" -verify -pedantic -fsyntax-only -DB32 -DQUALS= -cl-ext=-cl_khr_subgroups,-__opencl_c_subgroups
+// RUN: %clang_cc1 %s -cl-std=CL3.0 -triple "spir-unknown-unknown" -verify -pedantic -fsyntax-only -DB32 -DQUALS="const volatile"
+// RUN: %clang_cc1 %s -cl-std=CL3.0 -triple "spir64-unknown-unknown" -verify -pedantic -fsyntax-only -Wconversion -DWCONV -DQUALS=
+// RUN: %clang_cc1 %s -cl-std=CL3.0 -triple "spir64-unknown-unknown" -verify -pedantic -fsyntax-only -Wconversion -DWCONV -DQUALS="const volatile"
 
 typedef struct {int a;} ndrange_t;
 // Diagnostic tests for different overloads of enqueue_kernel from Table 6.13.17.1 of OpenCL 2.0 Spec.
@@ -236,12 +241,12 @@ kernel void bar(global unsigned int *buf)
 kernel void foo1(global unsigned int *buf)
 {
   ndrange_t n;
-  buf[0] = get_kernel_max_sub_group_size_for_ndrange(n, ^(){}); // expected-error {{use of declaration 'get_kernel_max_sub_group_size_for_ndrange' requires cl_khr_subgroups support}}
+  buf[0] = get_kernel_max_sub_group_size_for_ndrange(n, ^(){}); // expected-error {{use of declaration 'get_kernel_max_sub_group_size_for_ndrange' requires cl_khr_subgroups or __opencl_c_subgroups support}}
 }
 
 kernel void bar1(global unsigned int *buf)
 {
   ndrange_t n;
-  buf[0] = get_kernel_sub_group_count_for_ndrange(n, ^(){}); // expected-error {{use of declaration 'get_kernel_sub_group_count_for_ndrange' requires cl_khr_subgroups support}}
+  buf[0] = get_kernel_sub_group_count_for_ndrange(n, ^(){}); // expected-error {{use of declaration 'get_kernel_sub_group_count_for_ndrange' requires cl_khr_subgroups or __opencl_c_subgroups support}}
 }
 #endif // ifdef cl_khr_subgroups
diff --git a/clang/test/SemaOpenCL/clang-builtin-version.cl b/clang/test/SemaOpenCL/clang-builtin-version.cl
index 3e270e64c69d1..7111db1784c60 100644
--- a/clang/test/SemaOpenCL/clang-builtin-version.cl
+++ b/clang/test/SemaOpenCL/clang-builtin-version.cl
@@ -1,6 +1,8 @@
 // RUN: %clang_cc1 %s -fblocks -verify -pedantic -fsyntax-only -ferror-limit 100
+// RUN: %clang_cc1 %s -fblocks -verify -pedantic -fsyntax-only -ferror-limit 100 -cl-std=CL1.2
+// RUN: %clang_cc1 %s -fblocks -verify -pedantic -fsyntax-only -ferror-limit 100 -cl-std=CL3.0 -cl-ext=-__opencl_c_device_enqueue,-__opencl_c_generic_address_space,-__opencl_c_pipes
 
-// Confirm CL2.0 Clang builtins are not available in earlier versions
+// Confirm CL2.0 Clang builtins are not available in earlier versions and in OpenCL C 3.0 without required features.
 
 kernel void dse_builtins() {
   int tmp;
@@ -13,6 +15,11 @@ kernel void dse_builtins() {
   size = get_kernel_preferred_work_group_size_multiple(^(void) { // expected-error{{implicit declaration of function 'get_kernel_preferred_work_group_size_multiple' is invalid in OpenCL}}
     return;
   });
+#if (__OPENCL_C_VERSION__ >= CL_VERSION_3_0) && !defined(__opencl_c_device_enqueue)
+// expected-error@-10{{support disabled - compile with -fblocks or for OpenCL C 2.0 or OpenCL C 3.0 with __opencl_c_device_enqueue feature}}
+// expected-error@-8{{support disabled - compile with -fblocks or for OpenCL C 2.0 or OpenCL C 3.0 with __opencl_c_device_enqueue feature}}
+// expected-error@-6{{support disabled - compile with -fblocks or for OpenCL C 2.0 or OpenCL C 3.0 with __opencl_c_device_enqueue feature}}
+#endif
 }
 
 void pipe_builtins() {
diff --git a/clang/test/SemaOpenCL/to_addr_builtin.cl b/clang/test/SemaOpenCL/to_addr_builtin.cl
index 8f4a20ac7662a..f5df7abc3b4f2 100644
--- a/clang/test/SemaOpenCL/to_addr_builtin.cl
+++ b/clang/test/SemaOpenCL/to_addr_builtin.cl
@@ -1,5 +1,7 @@
 // RUN: %clang_cc1 -verify -fsyntax-only %s
 // RUN: %clang_cc1 -Wconversion -verify -fsyntax-only -cl-std=CL2.0 %s
+// RUN: %clang_cc1 -Wconversion -verify -fsyntax-only -cl-std=CL3.0 -cl-ext=-all %s
+// RUN: %clang_cc1 -Wconversion -verify -fsyntax-only -cl-std=CL3.0 %s
 
 void test(void) {
   global int *glob;
@@ -11,7 +13,7 @@ void test(void) {
   const_int_ty *con_typedef;
 
   glob = to_global(glob, loc);
-#if __OPENCL_C_VERSION__ < CL_VERSION_2_0
+#if (__OPENCL_C_VERSION__ < CL_VERSION_2_0) || (__OPENCL_C_VERSION__ >= CL_VERSION_3_0 && !defined(__opencl_c_generic_address_space))
   // expected-error@-2{{implicit declaration of function 'to_global' is invalid in OpenCL}}
   // expected-warning@-3{{incompatible integer to pointer conversion assigning to '__global int *__private' from 'int'}}
 #else
@@ -20,28 +22,28 @@ void test(void) {
 
   int x;
   glob = to_global(x);
-#if __OPENCL_C_VERSION__ < CL_VERSION_2_0
+#if (__OPENCL_C_VERSION__ < CL_VERSION_2_0) || (__OPENCL_C_VERSION__ >= CL_VERSION_3_0 && !defined(__opencl_c_generic_address_space))
   // expected-warning@-2{{incompatible integer to pointer conversion assigning to '__global int *__private' from 'int'}}
 #else
   // expected-error@-4{{invalid argument x to function: 'to_global', expecting a generic pointer argument}}
 #endif
 
   glob = to_global(con);
-#if __OPENCL_C_VERSION__ < CL_VERSION_2_0
+#if (__OPENCL_C_VERSION__ < CL_VERSION_2_0) || (__OPENCL_C_VERSION__ >= CL_VERSION_3_0 && !defined(__opencl_c_generic_address_space))
   // expected-warning@-2{{incompatible integer to pointer conversion assigning to '__global int *__private' from 'int'}}
 #else
   // expected-error@-4{{invalid argument con to function: 'to_global', expecting a generic pointer argument}}
 #endif
 
   glob = to_global(con_typedef);
-#if __OPENCL_C_VERSION__ < CL_VERSION_2_0
+#if (__OPENCL_C_VERSION__ < CL_VERSION_2_0) || (__OPENCL_C_VERSION__ >= CL_VERSION_3_0 && !defined(__opencl_c_generic_address_space))
   // expected-warning@-2{{incompatible integer to pointer conversion assigning to '__global int *__private' from 'int'}}
 #else
   // expected-error@-4{{invalid argument con_typedef to function: 'to_global', expecting a generic pointer argument}}
 #endif
 
   loc = to_global(glob);
-#if __OPENCL_C_VERSION__ < CL_VERSION_2_0
+#if (__OPENCL_C_VERSION__ < CL_VERSION_2_0) || (__OPENCL_C_VERSION__ >= CL_VERSION_3_0 && !defined(__opencl_c_generic_address_space))
   // expected-warning@-2{{incompatible integer to pointer conversion assigning to '__local int *__private' from 'int'}}
 #else
   // expected-error@-4{{assigning '__global int *' to '__local int *__private' changes address space of pointer}}
@@ -49,7 +51,7 @@ void test(void) {
 #endif
 
   loc = to_private(glob);
-#if __OPENCL_C_VERSION__ < CL_VERSION_2_0
+#if (__OPENCL_C_VERSION__ < CL_VERSION_2_0) || (__OPENCL_C_VERSION__ >= CL_VERSION_3_0 && !defined(__opencl_c_generic_address_space))
   // expected-error@-2{{implicit declaration of function 'to_private' is invalid in OpenCL}}
   // expected-warning@-3{{incompatible integer to pointer conversion assigning to '__local int *__private' from 'int'}}
 #else
@@ -58,17 +60,17 @@ void test(void) {
 #endif
 
   loc = to_local(glob);
-#if __OPENCL_C_VERSION__ < CL_VERSION_2_0
+#if (__OPENCL_C_VERSION__ < CL_VERSION_2_0) || (__OPENCL_C_VERSION__ >= CL_VERSION_3_0 && !defined(__opencl_c_generic_address_space))
   // expected-error@-2{{implicit declaration of function 'to_local' is invalid in OpenCL}}
   // expected-warning@-3{{incompatible integer to pointer conversion assigning to '__local int *__private' from 'int'}}
   // expected-note@-4{{did you mean 'to_global'}}
-  // expected-note@13{{'to_global' declared here}}
+  // expected-note@15{{'to_global' declared here}}
 #else
   // expected-warning@-7{{passing non-generic address space pointer to to_local may cause dynamic conversion affecting performance}}
 #endif
 
   priv = to_global(glob);
-#if __OPENCL_C_VERSION__ < CL_VERSION_2_0
+#if (__OPENCL_C_VERSION__ < CL_VERSION_2_0) || (__OPENCL_C_VERSION__ >= CL_VERSION_3_0 && !defined(__opencl_c_generic_address_space))
   // expected-warning@-2{{incompatible integer to pointer conversion assigning to '__private int *__private' from 'int'}}
 #else
   // expected-error@-4{{assigning '__global int *' to '__private int *__private' changes address space of pointer}}
@@ -76,7 +78,7 @@ void test(void) {
 #endif
 
   priv = to_private(glob);
-#if __OPENCL_C_VERSION__ < CL_VERSION_2_0
+#if (__OPENCL_C_VERSION__ < CL_VERSION_2_0) || (__OPENCL_C_VERSION__ >= CL_VERSION_3_0 && !defined(__opencl_c_generic_address_space))
   // expected-warning@-2{{incompatible integer to pointer conversion assigning to '__private int *__private' from 'int'}}
 #else
   // expected-warning@-4{{passing non-generic address space pointer to to_private may cause dynamic conversion affecting performance}}
@@ -84,7 +86,7 @@ void test(void) {
 
 
   priv = to_local(glob);
-#if __OPENCL_C_VERSION__ < CL_VERSION_2_0
+#if (__OPENCL_C_VERSION__ < CL_VERSION_2_0) || (__OPENCL_C_VERSION__ >= CL_VERSION_3_0 && !defined(__opencl_c_generic_address_space))
   // expected-warning@-2{{incompatible integer to pointer conversion assigning to '__private int *__private' from 'int'}}
 #else
   // expected-error@-4{{assigning '__local int *' to '__private int *__private' changes address space of pointer}}
@@ -92,14 +94,14 @@ void test(void) {
 #endif
 
   glob = to_global(glob);
-#if __OPENCL_C_VERSION__ < CL_VERSION_2_0
+#if (__OPENCL_C_VERSION__ < CL_VERSION_2_0) || (__OPENCL_C_VERSION__ >= CL_VERSION_3_0 && !defined(__opencl_c_generic_address_space))
   // expected-warning@-2{{incompatible integer to pointer conversion assigning to '__global int *__private' from 'int'}}
 #else
   // expected-warning@-4{{passing non-generic address space pointer to to_global may cause dynamic conversion affecting performance}}
 #endif
 
   glob = to_private(glob);
-#if __OPENCL_C_VERSION__ < CL_VERSION_2_0
+#if (__OPENCL_C_VERSION__ < CL_VERSION_2_0) || (__OPENCL_C_VERSION__ >= CL_VERSION_3_0 && !defined(__opencl_c_generic_address_space))
   // expected-warning@-2{{incompatible integer to pointer conversion assigning to '__global int *__private' from 'int'}}
 #else
   // expected-error@-4{{assigning '__private int *' to '__global int *__private' changes address space of pointer}}
@@ -107,7 +109,7 @@ void test(void) {
 #endif
 
   glob = to_local(glob);
-#if __OPENCL_C_VERSION__ < CL_VERSION_2_0
+#if (__OPENCL_C_VERSION__ < CL_VERSION_2_0) || (__OPENCL_C_VERSION__ >= CL_VERSION_3_0 && !defined(__opencl_c_generic_address_space))
   // expected-warning@-2{{incompatible integer to pointer conversion assigning to '__global int *__private' from 'int'}}
 #else
   // expected-error@-4{{assigning '__local int *' to '__global int *__private' changes address space of pointer}}
@@ -115,7 +117,7 @@ void test(void) {
 #endif
 
   global char *glob_c = to_global(loc);
-#if __OPENCL_C_VERSION__ < CL_VERSION_2_0
+#if (__OPENCL_C_VERSION__ < CL_VERSION_2_0) || (__OPENCL_C_VERSION__ >= CL_VERSION_3_0 && !defined(__opencl_c_generic_address_space))
   // expected-warning@-2{{incompatible integer to pointer conversion initializing '__global char *__private' with an expression of type 'int'}}
 #else
   // expected-warning@-4{{incompatible pointer types initializing '__global char *__private' with an expression of type '__global int *'}}
@@ -123,7 +125,7 @@ void test(void) {
 #endif
 
   glob_wrong_ty = to_global(glob);
-#if __OPENCL_C_VERSION__ < CL_VERSION_2_0
+#if (__OPENCL_C_VERSION__ < CL_VERSION_2_0) || (__OPENCL_C_VERSION__ >= CL_VERSION_3_0 && !defined(__opencl_c_generic_address_space))
   // expected-warning@-2{{incompatible integer to pointer conversion assigning to '__global float *__private' from 'int'}}
 #else
   // expected-warning@-4{{incompatible pointer types assigning to '__global float *__private' from '__global int *'}}
diff --git a/compiler-rt/lib/tsan/rtl/tsan_interceptors_posix.cpp b/compiler-rt/lib/tsan/rtl/tsan_interceptors_posix.cpp
index 999be79c6418b..ea99c3843075a 100644
--- a/compiler-rt/lib/tsan/rtl/tsan_interceptors_posix.cpp
+++ b/compiler-rt/lib/tsan/rtl/tsan_interceptors_posix.cpp
@@ -1554,16 +1554,16 @@ TSAN_INTERCEPTOR(int, __fxstat, int version, int fd, void *buf) {
 #endif
 
 TSAN_INTERCEPTOR(int, fstat, int fd, void *buf) {
-#if SANITIZER_FREEBSD || SANITIZER_MAC || SANITIZER_ANDROID || SANITIZER_NETBSD
-  SCOPED_TSAN_INTERCEPTOR(fstat, fd, buf);
+#if SANITIZER_GLIBC
+  SCOPED_TSAN_INTERCEPTOR(__fxstat, 0, fd, buf);
   if (fd > 0)
     FdAccess(thr, pc, fd);
-  return REAL(fstat)(fd, buf);
+  return REAL(__fxstat)(0, fd, buf);
 #else
-  SCOPED_TSAN_INTERCEPTOR(__fxstat, 0, fd, buf);
+  SCOPED_TSAN_INTERCEPTOR(fstat, fd, buf);
   if (fd > 0)
     FdAccess(thr, pc, fd);
-  return REAL(__fxstat)(0, fd, buf);
+  return REAL(fstat)(fd, buf);
 #endif
 }
 
diff --git a/libcxx/include/__threading_support b/libcxx/include/__threading_support
index bf85d5f5d9f05..2db571d84c8f7 100644
--- a/libcxx/include/__threading_support
+++ b/libcxx/include/__threading_support
@@ -254,7 +254,6 @@ int __libcpp_tls_set(__libcpp_tls_key __key, void *__p);
 
 #if defined(_LIBCPP_HAS_THREAD_API_PTHREAD)
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_recursive_mutex_init(__libcpp_recursive_mutex_t *__m)
 {
   pthread_mutexattr_t attr;
@@ -279,88 +278,74 @@ int __libcpp_recursive_mutex_init(__libcpp_recursive_mutex_t *__m)
   return 0;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_recursive_mutex_lock(__libcpp_recursive_mutex_t *__m)
 {
   return pthread_mutex_lock(__m);
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 bool __libcpp_recursive_mutex_trylock(__libcpp_recursive_mutex_t *__m)
 {
   return pthread_mutex_trylock(__m) == 0;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_recursive_mutex_unlock(__libcpp_recursive_mutex_t *__m)
 {
   return pthread_mutex_unlock(__m);
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_recursive_mutex_destroy(__libcpp_recursive_mutex_t *__m)
 {
   return pthread_mutex_destroy(__m);
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_mutex_lock(__libcpp_mutex_t *__m)
 {
   return pthread_mutex_lock(__m);
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 bool __libcpp_mutex_trylock(__libcpp_mutex_t *__m)
 {
   return pthread_mutex_trylock(__m) == 0;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_mutex_unlock(__libcpp_mutex_t *__m)
 {
   return pthread_mutex_unlock(__m);
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_mutex_destroy(__libcpp_mutex_t *__m)
 {
   return pthread_mutex_destroy(__m);
 }
 
 // Condition Variable
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_condvar_signal(__libcpp_condvar_t *__cv)
 {
   return pthread_cond_signal(__cv);
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_condvar_broadcast(__libcpp_condvar_t *__cv)
 {
   return pthread_cond_broadcast(__cv);
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_condvar_wait(__libcpp_condvar_t *__cv, __libcpp_mutex_t *__m)
 {
   return pthread_cond_wait(__cv, __m);
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_condvar_timedwait(__libcpp_condvar_t *__cv, __libcpp_mutex_t *__m,
                                __libcpp_timespec_t *__ts)
 {
   return pthread_cond_timedwait(__cv, __m, __ts);
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_condvar_destroy(__libcpp_condvar_t *__cv)
 {
   return pthread_cond_destroy(__cv);
 }
 
 // Execute once
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_execute_once(__libcpp_exec_once_flag *flag,
                           void (*init_routine)()) {
   return pthread_once(flag, init_routine);
@@ -368,40 +353,34 @@ int __libcpp_execute_once(__libcpp_exec_once_flag *flag,
 
 // Thread id
 // Returns non-zero if the thread ids are equal, otherwise 0
-_LIBCPP_HIDE_FROM_ABI inline
 bool __libcpp_thread_id_equal(__libcpp_thread_id t1, __libcpp_thread_id t2)
 {
   return t1 == t2;
 }
 
 // Returns non-zero if t1 < t2, otherwise 0
-_LIBCPP_HIDE_FROM_ABI inline
 bool __libcpp_thread_id_less(__libcpp_thread_id t1, __libcpp_thread_id t2)
 {
   return t1 < t2;
 }
 
 // Thread
-_LIBCPP_HIDE_FROM_ABI inline
 bool __libcpp_thread_isnull(const __libcpp_thread_t *__t) {
   return __libcpp_thread_get_id(__t) == 0;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_thread_create(__libcpp_thread_t *__t, void *(*__func)(void *),
                            void *__arg)
 {
   return pthread_create(__t, nullptr, __func, __arg);
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 __libcpp_thread_id __libcpp_thread_get_current_id()
 {
   const __libcpp_thread_t thread = pthread_self();
   return __libcpp_thread_get_id(&thread);
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 __libcpp_thread_id __libcpp_thread_get_id(const __libcpp_thread_t *__t)
 {
 #if defined(__MVS__)
@@ -411,25 +390,21 @@ __libcpp_thread_id __libcpp_thread_get_id(const __libcpp_thread_t *__t)
 #endif
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_thread_join(__libcpp_thread_t *__t)
 {
   return pthread_join(*__t, nullptr);
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_thread_detach(__libcpp_thread_t *__t)
 {
   return pthread_detach(*__t);
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 void __libcpp_thread_yield()
 {
   sched_yield();
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 void __libcpp_thread_sleep_for(const chrono::nanoseconds& __ns)
 {
    __libcpp_timespec_t __ts = _VSTD::__convert_to_timespec<__libcpp_timespec_t>(__ns);
@@ -437,19 +412,16 @@ void __libcpp_thread_sleep_for(const chrono::nanoseconds& __ns)
 }
 
 // Thread local storage
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_tls_create(__libcpp_tls_key *__key, void (*__at_exit)(void *))
 {
   return pthread_key_create(__key, __at_exit);
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 void *__libcpp_tls_get(__libcpp_tls_key __key)
 {
   return pthread_getspecific(__key);
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_tls_set(__libcpp_tls_key __key, void *__p)
 {
     return pthread_setspecific(__key, __p);
@@ -457,56 +429,47 @@ int __libcpp_tls_set(__libcpp_tls_key __key, void *__p)
 
 #elif defined(_LIBCPP_HAS_THREAD_API_C11)
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_recursive_mutex_init(__libcpp_recursive_mutex_t *__m)
 {
   return mtx_init(__m, mtx_plain | mtx_recursive) == thrd_success ? 0 : EINVAL;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_recursive_mutex_lock(__libcpp_recursive_mutex_t *__m)
 {
   return mtx_lock(__m) == thrd_success ? 0 : EINVAL;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 bool __libcpp_recursive_mutex_trylock(__libcpp_recursive_mutex_t *__m)
 {
   return mtx_trylock(__m) == thrd_success;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_recursive_mutex_unlock(__libcpp_recursive_mutex_t *__m)
 {
   return mtx_unlock(__m) == thrd_success ? 0 : EINVAL;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_recursive_mutex_destroy(__libcpp_recursive_mutex_t *__m)
 {
   mtx_destroy(__m);
   return 0;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_mutex_lock(__libcpp_mutex_t *__m)
 {
   return mtx_lock(__m) == thrd_success ? 0 : EINVAL;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 bool __libcpp_mutex_trylock(__libcpp_mutex_t *__m)
 {
   return mtx_trylock(__m) == thrd_success;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_mutex_unlock(__libcpp_mutex_t *__m)
 {
   return mtx_unlock(__m) == thrd_success ? 0 : EINVAL;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_mutex_destroy(__libcpp_mutex_t *__m)
 {
   mtx_destroy(__m);
@@ -514,25 +477,21 @@ int __libcpp_mutex_destroy(__libcpp_mutex_t *__m)
 }
 
 // Condition Variable
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_condvar_signal(__libcpp_condvar_t *__cv)
 {
   return cnd_signal(__cv) == thrd_success ? 0 : EINVAL;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_condvar_broadcast(__libcpp_condvar_t *__cv)
 {
   return cnd_broadcast(__cv) == thrd_success ? 0 : EINVAL;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_condvar_wait(__libcpp_condvar_t *__cv, __libcpp_mutex_t *__m)
 {
   return cnd_wait(__cv, __m) == thrd_success ? 0 : EINVAL;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_condvar_timedwait(__libcpp_condvar_t *__cv, __libcpp_mutex_t *__m,
                                timespec *__ts)
 {
@@ -540,7 +499,6 @@ int __libcpp_condvar_timedwait(__libcpp_condvar_t *__cv, __libcpp_mutex_t *__m,
   return __ec == thrd_timedout ? ETIMEDOUT : __ec;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_condvar_destroy(__libcpp_condvar_t *__cv)
 {
   cnd_destroy(__cv);
@@ -548,7 +506,6 @@ int __libcpp_condvar_destroy(__libcpp_condvar_t *__cv)
 }
 
 // Execute once
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_execute_once(__libcpp_exec_once_flag *flag,
                           void (*init_routine)(void)) {
   ::call_once(flag, init_routine);
@@ -557,26 +514,22 @@ int __libcpp_execute_once(__libcpp_exec_once_flag *flag,
 
 // Thread id
 // Returns non-zero if the thread ids are equal, otherwise 0
-_LIBCPP_HIDE_FROM_ABI inline
 bool __libcpp_thread_id_equal(__libcpp_thread_id t1, __libcpp_thread_id t2)
 {
   return thrd_equal(t1, t2) != 0;
 }
 
 // Returns non-zero if t1 < t2, otherwise 0
-_LIBCPP_HIDE_FROM_ABI inline
 bool __libcpp_thread_id_less(__libcpp_thread_id t1, __libcpp_thread_id t2)
 {
   return t1 < t2;
 }
 
 // Thread
-_LIBCPP_HIDE_FROM_ABI inline
 bool __libcpp_thread_isnull(const __libcpp_thread_t *__t) {
   return __libcpp_thread_get_id(__t) == 0;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_thread_create(__libcpp_thread_t *__t, void *(*__func)(void *),
                            void *__arg)
 {
@@ -584,37 +537,31 @@ int __libcpp_thread_create(__libcpp_thread_t *__t, void *(*__func)(void *),
   return __ec == thrd_nomem ? ENOMEM : __ec;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 __libcpp_thread_id __libcpp_thread_get_current_id()
 {
   return thrd_current();
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 __libcpp_thread_id __libcpp_thread_get_id(const __libcpp_thread_t *__t)
 {
   return *__t;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_thread_join(__libcpp_thread_t *__t)
 {
   return thrd_join(*__t, nullptr) == thrd_success ? 0 : EINVAL;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_thread_detach(__libcpp_thread_t *__t)
 {
   return thrd_detach(*__t) == thrd_success ? 0 : EINVAL;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 void __libcpp_thread_yield()
 {
   thrd_yield();
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 void __libcpp_thread_sleep_for(const chrono::nanoseconds& __ns)
 {
    __libcpp_timespec_t __ts = _VSTD::__convert_to_timespec<__libcpp_timespec_t>(__ns);
@@ -622,19 +569,16 @@ void __libcpp_thread_sleep_for(const chrono::nanoseconds& __ns)
 }
 
 // Thread local storage
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_tls_create(__libcpp_tls_key *__key, void (*__at_exit)(void *))
 {
   return tss_create(__key, __at_exit) == thrd_success ? 0 : EINVAL;
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 void *__libcpp_tls_get(__libcpp_tls_key __key)
 {
   return tss_get(__key);
 }
 
-_LIBCPP_HIDE_FROM_ABI inline
 int __libcpp_tls_set(__libcpp_tls_key __key, void *__p)
 {
   return tss_set(__key, __p) == thrd_success ? 0 : EINVAL;
diff --git a/llvm/docs/GettingStarted.rst b/llvm/docs/GettingStarted.rst
index 7d699d87491bb..3341d56171fa1 100644
--- a/llvm/docs/GettingStarted.rst
+++ b/llvm/docs/GettingStarted.rst
@@ -1222,7 +1222,7 @@ following options with cmake:
    compiling more than one project, separate the items with a semicolon. Should
    you run into issues with the semicolon, try surrounding it with single quotes.
 
-* -DLLVM_ENABLE_RUNTIMES
+ * -DLLVM_ENABLE_RUNTIMES
    Set this equal to the runtimes you wish to compile (e.g. libcxx, libcxxabi, etc.)
    If compiling more than one runtime, separate the items with a semicolon. Should
    you run into issues with the semicolon, try surrounding it with single quotes.
diff --git a/llvm/include/llvm/Config/llvm-config.h.cmake b/llvm/include/llvm/Config/llvm-config.h.cmake
index 839a1c7c1984f..e8a8c20c7c7ad 100644
--- a/llvm/include/llvm/Config/llvm-config.h.cmake
+++ b/llvm/include/llvm/Config/llvm-config.h.cmake
@@ -109,4 +109,7 @@
 /* Define if building LLVM with BUILD_SHARED_LIBS */
 #cmakedefine LLVM_BUILD_SHARED_LIBS
 
+/* Define if building LLVM with LLVM_FORCE_USE_OLD_TOOLCHAIN_LIBS */
+#cmakedefine LLVM_FORCE_USE_OLD_TOOLCHAIN $(LLVM_FORCE_USE_OLD_TOOLCHAIN)
+
 #endif
diff --git a/llvm/include/llvm/Support/Compiler.h b/llvm/include/llvm/Support/Compiler.h
index f3317049524f3..80b2dfaec4273 100644
--- a/llvm/include/llvm/Support/Compiler.h
+++ b/llvm/include/llvm/Support/Compiler.h
@@ -90,9 +90,11 @@
 #define LLVM_MSC_PREREQ(version) (_MSC_VER >= (version))
 
 // We require at least VS 2019.
+#if !defined(LLVM_FORCE_USE_OLD_TOOLCHAIN)
 #if !LLVM_MSC_PREREQ(1920)
 #error LLVM requires at least VS 2019.
 #endif
+#endif
 
 #else
 #define LLVM_MSC_PREREQ(version) 0