vendor/clang/clang-trunk-r307894

98 files changed, 3609 insertions, 2250 deletions

diff --git a/test/Analysis/analyzer-config.c b/test/Analysis/analyzer-config.c
index 70521c63fbad..0d8d34fbf6b0 100644
--- a/test/Analysis/analyzer-config.c
+++ b/test/Analysis/analyzer-config.c
@@ -12,6 +12,8 @@ void foo() {
 
 // CHECK: [config]
 // CHECK-NEXT: cfg-conditional-static-initializers = true
+// CHECK-NEXT: cfg-implicit-dtors = true
+// CHECK-NEXT: cfg-lifetime = false
 // CHECK-NEXT: cfg-temporary-dtors = false
 // CHECK-NEXT: faux-bodies = true
 // CHECK-NEXT: graph-trim-interval = 1000
@@ -27,5 +29,4 @@ void foo() {
 // CHECK-NEXT: region-store-small-struct-limit = 2
 // CHECK-NEXT: widen-loops = false
 // CHECK-NEXT: [stats]
-// CHECK-NEXT: num-entries = 15
-
+// CHECK-NEXT: num-entries = 17
diff --git a/test/Analysis/analyzer-config.cpp b/test/Analysis/analyzer-config.cpp
index 60c03c1e497b..4166a730cf53 100644
--- a/test/Analysis/analyzer-config.cpp
+++ b/test/Analysis/analyzer-config.cpp
@@ -23,6 +23,8 @@ public:
 // CHECK-NEXT: c++-stdlib-inlining = true
 // CHECK-NEXT: c++-template-inlining = true
 // CHECK-NEXT: cfg-conditional-static-initializers = true
+// CHECK-NEXT: cfg-implicit-dtors = true
+// CHECK-NEXT: cfg-lifetime = false
 // CHECK-NEXT: cfg-temporary-dtors = false
 // CHECK-NEXT: faux-bodies = true
 // CHECK-NEXT: graph-trim-interval = 1000
@@ -38,4 +40,4 @@ public:
 // CHECK-NEXT: region-store-small-struct-limit = 2
 // CHECK-NEXT: widen-loops = false
 // CHECK-NEXT: [stats]
-// CHECK-NEXT: num-entries = 20
+// CHECK-NEXT: num-entries = 22
diff --git a/test/Analysis/enum.cpp b/test/Analysis/enum.cpp
index b561e65ac86e..96408473cede 100644
--- a/test/Analysis/enum.cpp
+++ b/test/Analysis/enum.cpp
@@ -37,3 +37,33 @@ bool testNoCrashOnSwitchEnumBool(EnumBool E) {
   }
   return true;
 }
+
+bool testNoCrashOnSwitchEnumBoolConstant() {
+  EnumBool E = EnumBool::F;
+  switch (E) {
+    case EnumBool::F:
+      return false; 
+  }
+  return true; 
+}
+
+typedef __INTPTR_TYPE__ intptr_t;
+bool testNoCrashOnSwitchEnumBoolConstantCastedFromNullptr() {
+  EnumBool E = static_cast<EnumBool>((intptr_t)nullptr);
+  switch (E) {
+  case EnumBool::F:
+    return false;
+  }
+  return true;
+}
+
+bool testNoCrashOnSwitchEnumBoolConstantCastedFromPtr() {
+  int X;
+  intptr_t P = (intptr_t)&X;
+  EnumBool E = static_cast<EnumBool>(P);
+  switch (E) {
+  case EnumBool::F:
+    return false;
+  }
+  return true;
+}
diff --git a/test/Analysis/lifetime-cfg-output.cpp b/test/Analysis/lifetime-cfg-output.cpp
new file mode 100644
index 000000000000..1e6f56df6a47
--- /dev/null
+++ b/test/Analysis/lifetime-cfg-output.cpp
@@ -0,0 +1,783 @@
+// RUN: %clang_cc1 -fcxx-exceptions -fexceptions -analyze -analyzer-checker=debug.DumpCFG -analyzer-config cfg-lifetime=true -analyzer-config cfg-implicit-dtors=false %s > %t 2>&1
+// RUN: FileCheck --input-file=%t %s
+
+extern bool UV;
+class A {
+public:
+  // CHECK:       [B2 (ENTRY)]
+  // CHECK-NEXT:    Succs (1): B1
+  // CHECK:       [B1]
+  // CHECK-NEXT:    1: true
+  // CHECK-NEXT:    2: UV
+  // CHECK-NEXT:    3: [B1.2] = [B1.1]
+  // CHECK-NEXT:    Preds (1): B2
+  // CHECK-NEXT:    Succs (1): B0
+  // CHECK:       [B0 (EXIT)]
+  // CHECK-NEXT:    Preds (1): B1
+  A() {
+    UV = true;
+  }
+  // CHECK:       [B3 (ENTRY)]
+  // CHECK-NEXT:    Succs (1): B2
+  // CHECK:       [B1]
+  // CHECK-NEXT:    1: 0
+  // CHECK-NEXT:    2: this
+  // CHECK-NEXT:    3: [B1.2]->p
+  // CHECK-NEXT:    4: [B1.3] (ImplicitCastExpr, LValueToRValue, int *)
+  // CHECK-NEXT:    5: *[B1.4]
+  // CHECK-NEXT:    6: [B1.5] = [B1.1]
+  // CHECK-NEXT:    Preds (1): B2
+  // CHECK-NEXT:    Succs (1): B0
+  // CHECK:       [B2]
+  // CHECK-NEXT:    1: this
+  // CHECK-NEXT:    2: [B2.1]->p
+  // CHECK-NEXT:    3: [B2.2] (ImplicitCastExpr, LValueToRValue, int *)
+  // CHECK-NEXT:    4: [B2.3] (ImplicitCastExpr, PointerToBoolean, _Bool)
+  // CHECK-NEXT:    T: if [B2.4]
+  // CHECK-NEXT:    Preds (1): B3
+  // CHECK-NEXT:    Succs (2): B1 B0
+  // CHECK:       [B0 (EXIT)]
+  // CHECK-NEXT:    Preds (2): B1 B2
+  ~A() {
+    if (p)
+      *p = 0;
+  }
+  // CHECK:       [B2 (ENTRY)]
+  // CHECK-NEXT:    Succs (1): B1
+  // CHECK:       [B1]
+  // CHECK-NEXT:    1: 1
+  // CHECK-NEXT:    2: return [B1.1];
+  // CHECK-NEXT:    Preds (1): B2
+  // CHECK-NEXT:    Succs (1): B0
+  // CHECK:       [B0 (EXIT)]
+  // CHECK-NEXT:    Preds (1): B1
+  operator int() const { return 1; }
+  int *p;
+};
+
+// CHECK:       [B2 (ENTRY)]
+// CHECK-NEXT:    Succs (1): B1
+// CHECK:       [B1]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A a;
+// CHECK-NEXT:    3: a
+// CHECK-NEXT:    4: [B1.3] (ImplicitCastExpr, NoOp, const class A)
+// CHECK-NEXT:    5: const A &b = a;
+// CHECK-NEXT:    6: A() (CXXConstructExpr, class A)
+// CHECK-NEXT:    7: [B1.6] (BindTemporary)
+// CHECK-NEXT:    8: [B1.7] (ImplicitCastExpr, NoOp, const class A)
+// CHECK-NEXT:    9: [B1.8]
+// CHECK-NEXT:   10: const A &c = A();
+// CHECK-NEXT:   11: [B1.10] (Lifetime ends)
+// CHECK-NEXT:   12: [B1.2] (Lifetime ends)
+// CHECK-NEXT:   13: [B1.5] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B2
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B0 (EXIT)]
+// CHECK-NEXT:    Preds (1): B1
+void test_const_ref() {
+  A a;
+  const A &b = a;
+  const A &c = A();
+}
+
+// CHECK:      [B2 (ENTRY)]
+// CHECK-NEXT:   Succs (1): B1
+// CHECK:       [B1]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A [2])
+// CHECK-NEXT:    2: A a[2];
+// CHECK-NEXT:    3:  (CXXConstructExpr, class A [0])
+// CHECK-NEXT:    4: A b[0];
+// lifetime of a ends when its destructors are run
+// CHECK-NEXT:    5: [B1.2] (Lifetime ends)
+// lifetime of b ends when its storage duration ends
+// CHECK-NEXT:    6: [B1.4] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B2
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:      [B0 (EXIT)]
+// CHECK-NEXT:   Preds (1): B1
+void test_array() {
+  A a[2];
+  A b[0];
+}
+
+// CHECK:      [B2 (ENTRY)]
+// CHECK-NEXT:   Succs (1): B1
+// CHECK:       [B1]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A a;
+// CHECK-NEXT:    3:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    4: A c;
+// CHECK-NEXT:    5:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    6: A d;
+// CHECK-NEXT:    7: [B1.6] (Lifetime ends)
+// CHECK-NEXT:    8: [B1.4] (Lifetime ends)
+// CHECK-NEXT:    9:  (CXXConstructExpr, class A)
+// CHECK-NEXT:   10: A b;
+// CHECK-NEXT:   11: [B1.10] (Lifetime ends)
+// CHECK-NEXT:   12: [B1.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B2
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:      [B0 (EXIT)]
+// CHECK-NEXT:   Preds (1): B1
+void test_scope() {
+  A a;
+  {
+    A c;
+    A d;
+  }
+  A b;
+}
+
+// CHECK:      [B4 (ENTRY)]
+// CHECK-NEXT:   Succs (1): B3
+// CHECK:       [B1]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A c;
+// CHECK-NEXT:    3: [B1.2] (Lifetime ends)
+// CHECK-NEXT:    4: [B3.4] (Lifetime ends)
+// CHECK-NEXT:    5: [B3.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B3
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B2]
+// CHECK-NEXT:    1: return;
+// CHECK-NEXT:    2: [B3.4] (Lifetime ends)
+// CHECK-NEXT:    3: [B3.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B3
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B3]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A a;
+// CHECK-NEXT:    3:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    4: A b;
+// CHECK-NEXT:    5: UV
+// CHECK-NEXT:    6: [B3.5] (ImplicitCastExpr, LValueToRValue, _Bool)
+// CHECK-NEXT:    T: if [B3.6]
+// CHECK-NEXT:    Preds (1): B4
+// CHECK-NEXT:    Succs (2): B2 B1
+
+// CHECK:      [B0 (EXIT)]
+// CHECK-NEXT:   Preds (2): B1 B2
+void test_return() {
+  A a;
+  A b;
+  if (UV)
+    return;
+  A c;
+}
+
+// CHECK:       [B5 (ENTRY)]
+// CHECK-NEXT:    Succs (1): B4
+// CHECK:       [B1]
+// CHECK-NEXT:    1: [B4.6] (Lifetime ends)
+// CHECK-NEXT:    2: [B4.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (2): B2 B3
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B2]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A c;
+// CHECK-NEXT:    3: [B2.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B4
+// CHECK-NEXT:    Succs (1): B1
+// CHECK:       [B3]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A c;
+// CHECK-NEXT:    3: [B3.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B4
+// CHECK-NEXT:    Succs (1): B1
+// CHECK:       [B4]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A a;
+// CHECK-NEXT:    3: a
+// CHECK-NEXT:    4: [B4.3] (ImplicitCastExpr, NoOp, const class A)
+// CHECK-NEXT:    5: [B4.4] (CXXConstructExpr, class A)
+// CHECK-NEXT:    6: A b = a;
+// CHECK-NEXT:    7: b
+// CHECK-NEXT:    8: [B4.7] (ImplicitCastExpr, NoOp, const class A)
+// CHECK-NEXT:    9: [B4.8].operator int
+// CHECK-NEXT:   10: [B4.8]
+// CHECK-NEXT:   11: [B4.10] (ImplicitCastExpr, UserDefinedConversion, int)
+// CHECK-NEXT:   12: [B4.11] (ImplicitCastExpr, IntegralToBoolean, _Bool)
+// CHECK-NEXT:    T: if [B4.12]
+// CHECK-NEXT:    Preds (1): B5
+// CHECK-NEXT:    Succs (2): B3 B2
+// CHECK:       [B0 (EXIT)]
+// CHECK-NEXT:    Preds (1): B1
+void test_if_implicit_scope() {
+  A a;
+  if (A b = a)
+    A c;
+  else
+    A c;
+}
+
+// CHECK:       [B9 (ENTRY)]
+// CHECK-NEXT:    Succs (1): B8
+// CHECK:       [B1]
+// CHECK-NEXT:    1: [B8.6] (Lifetime ends)
+// CHECK-NEXT:    2:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    3: A e;
+// CHECK-NEXT:    4: [B1.3] (Lifetime ends)
+// CHECK-NEXT:    5: [B8.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (2): B2 B5
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B2]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A d;
+// CHECK-NEXT:    3: [B2.2] (Lifetime ends)
+// CHECK-NEXT:    4: [B4.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B4
+// CHECK-NEXT:    Succs (1): B1
+// CHECK:       [B3]
+// CHECK-NEXT:    1: return;
+// CHECK-NEXT:    2: [B4.2] (Lifetime ends)
+// CHECK-NEXT:    3: [B8.6] (Lifetime ends)
+// CHECK-NEXT:    4: [B8.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B4
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B4]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A c;
+// CHECK-NEXT:    3: UV
+// CHECK-NEXT:    4: [B4.3] (ImplicitCastExpr, LValueToRValue, _Bool)
+// CHECK-NEXT:    T: if [B4.4]
+// CHECK-NEXT:    Preds (1): B8
+// CHECK-NEXT:    Succs (2): B3 B2
+// CHECK:       [B5]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A d;
+// CHECK-NEXT:    3: [B5.2] (Lifetime ends)
+// CHECK-NEXT:    4: [B7.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B7
+// CHECK-NEXT:    Succs (1): B1
+// CHECK:       [B6]
+// CHECK-NEXT:    1: return;
+// CHECK-NEXT:    2: [B7.2] (Lifetime ends)
+// CHECK-NEXT:    3: [B8.6] (Lifetime ends)
+// CHECK-NEXT:    4: [B8.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B7
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B7]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A c;
+// CHECK-NEXT:    3: UV
+// CHECK-NEXT:    4: [B7.3] (ImplicitCastExpr, LValueToRValue, _Bool)
+// CHECK-NEXT:    T: if [B7.4]
+// CHECK-NEXT:    Preds (1): B8
+// CHECK-NEXT:    Succs (2): B6 B5
+// CHECK:       [B8]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A a;
+// CHECK-NEXT:    3: a
+// CHECK-NEXT:    4: [B8.3] (ImplicitCastExpr, NoOp, const class A)
+// CHECK-NEXT:    5: [B8.4] (CXXConstructExpr, class A)
+// CHECK-NEXT:    6: A b = a;
+// CHECK-NEXT:    7: b
+// CHECK-NEXT:    8: [B8.7] (ImplicitCastExpr, NoOp, const class A)
+// CHECK-NEXT:    9: [B8.8].operator int
+// CHECK-NEXT:   10: [B8.8]
+// CHECK-NEXT:   11: [B8.10] (ImplicitCastExpr, UserDefinedConversion, int)
+// CHECK-NEXT:   12: [B8.11] (ImplicitCastExpr, IntegralToBoolean, _Bool)
+// CHECK-NEXT:    T: if [B8.12]
+// CHECK-NEXT:    Preds (1): B9
+// CHECK-NEXT:    Succs (2): B7 B4
+// CHECK:       [B0 (EXIT)]
+// CHECK-NEXT:    Preds (3): B1 B3 B6
+void test_if_jumps() {
+  A a;
+  if (A b = a) {
+    A c;
+    if (UV)
+      return;
+    A d;
+  } else {
+    A c;
+    if (UV)
+      return;
+    A d;
+  }
+  A e;
+}
+
+// CHECK:       [B6 (ENTRY)]
+// CHECK-NEXT:    Succs (1): B5
+// CHECK:       [B1]
+// CHECK-NEXT:    1: [B4.4] (Lifetime ends)
+// CHECK-NEXT:    2: [B5.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B4
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B2]
+// CHECK-NEXT:    Preds (1): B3
+// CHECK-NEXT:    Succs (1): B4
+// CHECK:       [B3]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A c;
+// CHECK-NEXT:    3: [B3.2] (Lifetime ends)
+// CHECK-NEXT:    4: [B4.4] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B4
+// CHECK-NEXT:    Succs (1): B2
+// CHECK:       [B4]
+// CHECK-NEXT:    1: a
+// CHECK-NEXT:    2: [B4.1] (ImplicitCastExpr, NoOp, const class A)
+// CHECK-NEXT:    3: [B4.2] (CXXConstructExpr, class A)
+// CHECK-NEXT:    4: A b = a;
+// CHECK-NEXT:    5: b
+// CHECK-NEXT:    6: [B4.5] (ImplicitCastExpr, NoOp, const class A)
+// CHECK-NEXT:    7: [B4.6].operator int
+// CHECK-NEXT:    8: [B4.6]
+// CHECK-NEXT:    9: [B4.8] (ImplicitCastExpr, UserDefinedConversion, int)
+// CHECK-NEXT:   10: [B4.9] (ImplicitCastExpr, IntegralToBoolean, _Bool)
+// CHECK-NEXT:    T: while [B4.10]
+// CHECK-NEXT:    Preds (2): B2 B5
+// CHECK-NEXT:    Succs (2): B3 B1
+// CHECK:       [B5]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A a;
+// CHECK-NEXT:    Preds (1): B6
+// CHECK-NEXT:    Succs (1): B4
+// CHECK:       [B0 (EXIT)]
+// CHECK-NEXT:    Preds (1): B1
+void test_while_implicit_scope() {
+  A a;
+  while (A b = a)
+    A c;
+}
+
+// CHECK:       [B12 (ENTRY)]
+// CHECK-NEXT:    Succs (1): B11
+// CHECK:       [B1]
+// CHECK-NEXT:    1: [B10.4] (Lifetime ends)
+// CHECK-NEXT:    2:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    3: A e;
+// CHECK-NEXT:    4: [B1.3] (Lifetime ends)
+// CHECK-NEXT:    5: [B11.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (2): B8 B10
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B2]
+// CHECK-NEXT:    Preds (2): B3 B6
+// CHECK-NEXT:    Succs (1): B10
+// CHECK:       [B3]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A d;
+// CHECK-NEXT:    3: [B3.2] (Lifetime ends)
+// CHECK-NEXT:    4: [B9.2] (Lifetime ends)
+// CHECK-NEXT:    5: [B10.4] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B5
+// CHECK-NEXT:    Succs (1): B2
+// CHECK:       [B4]
+// CHECK-NEXT:    1: return;
+// CHECK-NEXT:    2: [B9.2] (Lifetime ends)
+// CHECK-NEXT:    3: [B10.4] (Lifetime ends)
+// CHECK-NEXT:    4: [B11.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B5
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B5]
+// CHECK-NEXT:    1: UV
+// CHECK-NEXT:    2: [B5.1] (ImplicitCastExpr, LValueToRValue, _Bool)
+// CHECK-NEXT:    T: if [B5.2]
+// CHECK-NEXT:    Preds (1): B7
+// CHECK-NEXT:    Succs (2): B4 B3
+// CHECK:       [B6]
+// CHECK-NEXT:    1: [B9.2] (Lifetime ends)
+// CHECK-NEXT:    2: [B10.4] (Lifetime ends)
+// CHECK-NEXT:    T: continue;
+// CHECK-NEXT:    Preds (1): B7
+// CHECK-NEXT:    Succs (1): B2
+// CHECK:       [B7]
+// CHECK-NEXT:    1: UV
+// CHECK-NEXT:    2: [B7.1] (ImplicitCastExpr, LValueToRValue, _Bool)
+// CHECK-NEXT:    T: if [B7.2]
+// CHECK-NEXT:    Preds (1): B9
+// CHECK-NEXT:    Succs (2): B6 B5
+// CHECK:       [B8]
+// CHECK-NEXT:    1: [B9.2] (Lifetime ends)
+// CHECK-NEXT:    T: break;
+// CHECK-NEXT:    Preds (1): B9
+// CHECK-NEXT:    Succs (1): B1
+// CHECK:       [B9]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A c;
+// CHECK-NEXT:    3: UV
+// CHECK-NEXT:    4: [B9.3] (ImplicitCastExpr, LValueToRValue, _Bool)
+// CHECK-NEXT:    T: if [B9.4]
+// CHECK-NEXT:    Preds (1): B10
+// CHECK-NEXT:    Succs (2): B8 B7
+// CHECK:       [B10]
+// CHECK-NEXT:    1: a
+// CHECK-NEXT:    2: [B10.1] (ImplicitCastExpr, NoOp, const class A)
+// CHECK-NEXT:    3: [B10.2] (CXXConstructExpr, class A)
+// CHECK-NEXT:    4: A b = a;
+// CHECK-NEXT:    5: b
+// CHECK-NEXT:    6: [B10.5] (ImplicitCastExpr, NoOp, const class A)
+// CHECK-NEXT:    7: [B10.6].operator int
+// CHECK-NEXT:    8: [B10.6]
+// CHECK-NEXT:    9: [B10.8] (ImplicitCastExpr, UserDefinedConversion, int)
+// CHECK-NEXT:   10: [B10.9] (ImplicitCastExpr, IntegralToBoolean, _Bool)
+// CHECK-NEXT:    T: while [B10.10]
+// CHECK-NEXT:    Preds (2): B2 B11
+// CHECK-NEXT:    Succs (2): B9 B1
+// CHECK:       [B11]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A a;
+// CHECK-NEXT:    Preds (1): B12
+// CHECK-NEXT:    Succs (1): B10
+// CHECK:       [B0 (EXIT)]
+// CHECK-NEXT:    Preds (2): B1 B4
+void test_while_jumps() {
+  A a;
+  while (A b = a) {
+    A c;
+    if (UV)
+      break;
+    if (UV)
+      continue;
+    if (UV)
+      return;
+    A d;
+  }
+  A e;
+}
+
+// CHECK:       [B12 (ENTRY)]
+// CHECK-NEXT:    Succs (1): B11
+// CHECK:       [B1]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A d;
+// CHECK-NEXT:    3: [B1.2] (Lifetime ends)
+// CHECK-NEXT:    4: [B11.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (2): B8 B2
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B2]
+// CHECK-NEXT:    1: UV
+// CHECK-NEXT:    2: [B2.1] (ImplicitCastExpr, LValueToRValue, _Bool)
+// CHECK-NEXT:    T: do ... while [B2.2]
+// CHECK-NEXT:    Preds (2): B3 B6
+// CHECK-NEXT:    Succs (2): B10 B1
+// CHECK:       [B3]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A c;
+// CHECK-NEXT:    3: [B3.2] (Lifetime ends)
+// CHECK-NEXT:    4: [B9.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B5
+// CHECK-NEXT:    Succs (1): B2
+// CHECK:       [B4]
+// CHECK-NEXT:    1: return;
+// CHECK-NEXT:    2: [B9.2] (Lifetime ends)
+// CHECK-NEXT:    3: [B11.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B5
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B5]
+// CHECK-NEXT:    1: UV
+// CHECK-NEXT:    2: [B5.1] (ImplicitCastExpr, LValueToRValue, _Bool)
+// CHECK-NEXT:    T: if [B5.2]
+// CHECK-NEXT:    Preds (1): B7
+// CHECK-NEXT:    Succs (2): B4 B3
+// CHECK:       [B6]
+// CHECK-NEXT:    1: [B9.2] (Lifetime ends)
+// CHECK-NEXT:    T: continue;
+// CHECK-NEXT:    Preds (1): B7
+// CHECK-NEXT:    Succs (1): B2
+// CHECK:       [B7]
+// CHECK-NEXT:    1: UV
+// CHECK-NEXT:    2: [B7.1] (ImplicitCastExpr, LValueToRValue, _Bool)
+// CHECK-NEXT:    T: if [B7.2]
+// CHECK-NEXT:    Preds (1): B9
+// CHECK-NEXT:    Succs (2): B6 B5
+// CHECK:       [B8]
+// CHECK-NEXT:    1: [B9.2] (Lifetime ends)
+// CHECK-NEXT:    T: break;
+// CHECK-NEXT:    Preds (1): B9
+// CHECK-NEXT:    Succs (1): B1
+// CHECK:       [B9]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A b;
+// CHECK-NEXT:    3: UV
+// CHECK-NEXT:    4: [B9.3] (ImplicitCastExpr, LValueToRValue, _Bool)
+// CHECK-NEXT:    T: if [B9.4]
+// CHECK-NEXT:    Preds (2): B10 B11
+// CHECK-NEXT:    Succs (2): B8 B7
+// CHECK:       [B10]
+// CHECK-NEXT:    Preds (1): B2
+// CHECK-NEXT:    Succs (1): B9
+// CHECK:       [B11]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A a;
+// CHECK-NEXT:    Preds (1): B12
+// CHECK-NEXT:    Succs (1): B9
+// CHECK:       [B0 (EXIT)]
+// CHECK-NEXT:    Preds (2): B1 B4
+void test_do_jumps() {
+  A a;
+  do {
+    A b;
+    if (UV)
+      break;
+    if (UV)
+      continue;
+    if (UV)
+      return;
+    A c;
+  } while (UV);
+  A d;
+}
+
+// CHECK:       [B6 (ENTRY)]
+// CHECK-NEXT:    Succs (1): B5
+// CHECK:       [B1]
+// CHECK-NEXT:    1: [B4.4] (Lifetime ends)
+// CHECK-NEXT:    2: [B5.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B4
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B2]
+// CHECK-NEXT:    Preds (1): B3
+// CHECK-NEXT:    Succs (1): B4
+// CHECK:       [B3]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A c;
+// CHECK-NEXT:    3: [B3.2] (Lifetime ends)
+// CHECK-NEXT:    4: [B4.4] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B4
+// CHECK-NEXT:    Succs (1): B2
+// CHECK:       [B4]
+// CHECK-NEXT:    1: a
+// CHECK-NEXT:    2: [B4.1] (ImplicitCastExpr, NoOp, const class A)
+// CHECK-NEXT:    3: [B4.2] (CXXConstructExpr, class A)
+// CHECK-NEXT:    4: A b = a;
+// CHECK-NEXT:    5: b
+// CHECK-NEXT:    6: [B4.5] (ImplicitCastExpr, NoOp, const class A)
+// CHECK-NEXT:    7: [B4.6].operator int
+// CHECK-NEXT:    8: [B4.6]
+// CHECK-NEXT:    9: [B4.8] (ImplicitCastExpr, UserDefinedConversion, int)
+// CHECK-NEXT:   10: [B4.9] (ImplicitCastExpr, IntegralToBoolean, _Bool)
+// CHECK-NEXT:    T: for (...; [B4.10]; )
+// CHECK-NEXT:    Preds (2): B2 B5
+// CHECK-NEXT:    Succs (2): B3 B1
+// CHECK:       [B5]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A a;
+// CHECK-NEXT:    Preds (1): B6
+// CHECK-NEXT:    Succs (1): B4
+// CHECK:       [B0 (EXIT)]
+// CHECK-NEXT:    Preds (1): B1
+void test_for_implicit_scope() {
+  for (A a; A b = a;)
+    A c;
+}
+
+// CHECK:       [B12 (ENTRY)]
+// CHECK-NEXT:    Succs (1): B11
+// CHECK:       [B1]
+// CHECK-NEXT:    1: [B10.4] (Lifetime ends)
+// CHECK-NEXT:    2: [B11.4] (Lifetime ends)
+// CHECK-NEXT:    3:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    4: A f;
+// CHECK-NEXT:    5: [B1.4] (Lifetime ends)
+// CHECK-NEXT:    6: [B11.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (2): B8 B10
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B2]
+// CHECK-NEXT:    Preds (2): B3 B6
+// CHECK-NEXT:    Succs (1): B10
+// CHECK:       [B3]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A e;
+// CHECK-NEXT:    3: [B3.2] (Lifetime ends)
+// CHECK-NEXT:    4: [B9.2] (Lifetime ends)
+// CHECK-NEXT:    5: [B10.4] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B5
+// CHECK-NEXT:    Succs (1): B2
+// CHECK:       [B4]
+// CHECK-NEXT:    1: return;
+// CHECK-NEXT:    2: [B9.2] (Lifetime ends)
+// CHECK-NEXT:    3: [B10.4] (Lifetime ends)
+// CHECK-NEXT:    4: [B11.4] (Lifetime ends)
+// CHECK-NEXT:    5: [B11.2] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B5
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B5]
+// CHECK-NEXT:    1: UV
+// CHECK-NEXT:    2: [B5.1] (ImplicitCastExpr, LValueToRValue, _Bool)
+// CHECK-NEXT:    T: if [B5.2]
+// CHECK-NEXT:    Preds (1): B7
+// CHECK-NEXT:    Succs (2): B4 B3
+// CHECK:       [B6]
+// CHECK-NEXT:    1: [B9.2] (Lifetime ends)
+// CHECK-NEXT:    T: continue;
+// CHECK-NEXT:    Preds (1): B7
+// CHECK-NEXT:    Succs (1): B2
+// CHECK:       [B7]
+// CHECK-NEXT:    1: UV
+// CHECK-NEXT:    2: [B7.1] (ImplicitCastExpr, LValueToRValue, _Bool)
+// CHECK-NEXT:    T: if [B7.2]
+// CHECK-NEXT:    Preds (1): B9
+// CHECK-NEXT:    Succs (2): B6 B5
+// CHECK:       [B8]
+// CHECK-NEXT:    1: [B9.2] (Lifetime ends)
+// CHECK-NEXT:    T: break;
+// CHECK-NEXT:    Preds (1): B9
+// CHECK-NEXT:    Succs (1): B1
+// CHECK:       [B9]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A d;
+// CHECK-NEXT:    3: UV
+// CHECK-NEXT:    4: [B9.3] (ImplicitCastExpr, LValueToRValue, _Bool)
+// CHECK-NEXT:    T: if [B9.4]
+// CHECK-NEXT:    Preds (1): B10
+// CHECK-NEXT:    Succs (2): B8 B7
+// CHECK:       [B10]
+// CHECK-NEXT:    1: b
+// CHECK-NEXT:    2: [B10.1] (ImplicitCastExpr, NoOp, const class A)
+// CHECK-NEXT:    3: [B10.2] (CXXConstructExpr, class A)
+// CHECK-NEXT:    4: A c = b;
+// CHECK-NEXT:    5: c
+// CHECK-NEXT:    6: [B10.5] (ImplicitCastExpr, NoOp, const class A)
+// CHECK-NEXT:    7: [B10.6].operator int
+// CHECK-NEXT:    8: [B10.6]
+// CHECK-NEXT:    9: [B10.8] (ImplicitCastExpr, UserDefinedConversion, int)
+// CHECK-NEXT:   10: [B10.9] (ImplicitCastExpr, IntegralToBoolean, _Bool)
+// CHECK-NEXT:    T: for (...; [B10.10]; )
+// CHECK-NEXT:    Preds (2): B2 B11
+// CHECK-NEXT:    Succs (2): B9 B1
+// CHECK:       [B11]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A a;
+// CHECK-NEXT:    3:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    4: A b;
+// CHECK-NEXT:    Preds (1): B12
+// CHECK-NEXT:    Succs (1): B10
+// CHECK:       [B0 (EXIT)]
+// CHECK-NEXT:    Preds (2): B1 B4
+void test_for_jumps() {
+  A a;
+  for (A b; A c = b;) {
+    A d;
+    if (UV)
+      break;
+    if (UV)
+      continue;
+    if (UV)
+      return;
+    A e;
+  }
+  A f;
+}
+
+// CHECK:       [B2 (ENTRY)]
+// CHECK-NEXT:    Succs (1): B1
+// CHECK:       [B1]
+// CHECK-NEXT:    1:  (CXXConstructExpr, class A)
+// CHECK-NEXT:    2: A a;
+// CHECK-NEXT:    3: int n;
+// CHECK-NEXT:    4: n
+// CHECK-NEXT:    5: &[B1.4]
+// CHECK-NEXT:    6: a
+// CHECK-NEXT:    7: [B1.6].p
+// CHECK-NEXT:    8: [B1.7] = [B1.5]
+// CHECK-NEXT:    9: [B1.2] (Lifetime ends)
+// CHECK-NEXT:   10: [B1.3] (Lifetime ends)
+// CHECK-NEXT:    Preds (1): B2
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B0 (EXIT)]
+// CHECK-NEXT:    Preds (1): B1
+void test_trivial_vs_non_trivial_order() {
+  A a;
+  int n;
+  a.p = &n;
+}
+
+// CHECK:       [B4 (ENTRY)]
+// CHECK-NEXT:    Succs (1): B3
+// CHECK:       [B1]
+// CHECK-NEXT:   a:
+// CHECK-NEXT:    1: 1
+// CHECK-NEXT:    2: i
+// CHECK-NEXT:    3: [B1.2] = [B1.1]
+// CHECK-NEXT:    4: [B2.1] (Lifetime ends)
+// CHECK-NEXT:    Preds (2): B2 B3
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B2]
+// CHECK-NEXT:    1: int i;
+// CHECK-NEXT:    Succs (1): B1
+// CHECK:       [B3]
+// CHECK-NEXT:    T: goto a;
+// CHECK-NEXT:    Preds (1): B4
+// CHECK-NEXT:    Succs (1): B1
+// CHECK:       [B0 (EXIT)]
+// CHECK-NEXT:    Preds (1): B1
+void goto_past_declaration() {
+  goto a;
+  int i;
+a:
+  i = 1;
+}
+
+// CHECK:       [B4 (ENTRY)]
+// CHECK-NEXT:    Succs (1): B3
+// CHECK:       [B1]
+// CHECK-NEXT:   a:
+// CHECK-NEXT:    1: 1
+// CHECK-NEXT:    2: k
+// CHECK-NEXT:    3: [B1.2] = [B1.1]
+// CHECK-NEXT:    4: [B2.4] (Lifetime ends)
+// CHECK-NEXT:    Preds (2): B2 B3
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B2]
+// CHECK-NEXT:    1: int j;
+// CHECK-NEXT:    2: [B2.1] (Lifetime ends)
+// CHECK-NEXT:    3: [B3.1] (Lifetime ends)
+// CHECK-NEXT:    4: int k;
+// CHECK-NEXT:    Succs (1): B1
+// CHECK:       [B3]
+// CHECK-NEXT:    1: int i;
+// CHECK-NEXT:    2: [B3.1] (Lifetime ends)
+// CHECK-NEXT:    T: goto a;
+// CHECK-NEXT:    Preds (1): B4
+// CHECK-NEXT:    Succs (1): B1
+// CHECK:       [B0 (EXIT)]
+// CHECK-NEXT:    Preds (1): B1
+void goto_past_declaration2() {
+  {
+    int i;
+    goto a;
+    int j;
+  }
+  {
+    int k;
+  a:
+    k = 1;
+  }
+}
+
+struct B {
+  ~B();
+};
+
+// CHECK:       [B4 (ENTRY)]
+// CHECK-NEXT:    Succs (1): B3
+// CHECK:       [B1]
+// CHECK-NEXT:    1: i
+// CHECK-NEXT:    2: [B1.1]++
+// CHECK-NEXT:    3: [B2.2] (Lifetime ends)
+// CHECK-NEXT:    4: [B3.1] (Lifetime ends)
+// CHECK-NEXT:    Succs (1): B0
+// CHECK:       [B2]
+// CHECK-NEXT:   label:
+// CHECK-NEXT:    1:  (CXXConstructExpr, struct B)
+// CHECK-NEXT:    2: B b;
+// CHECK-NEXT:    3: [B2.2] (Lifetime ends)
+// CHECK-NEXT:    T: goto label;
+// CHECK-NEXT:    Preds (2): B3 B2
+// CHECK-NEXT:    Succs (1): B2
+// CHECK:       [B3]
+// CHECK-NEXT:    1: int i;
+// CHECK-NEXT:    Preds (1): B4
+// CHECK-NEXT:    Succs (1): B2
+// CHECK:       [B0 (EXIT)]
+// CHECK-NEXT:    Preds (1): B1
+int backpatched_goto() {
+  int i;
+label:
+  B b;
+  goto label;
+  i++;
+}
diff --git a/test/CXX/except/except.spec/p11.cpp b/test/CXX/except/except.spec/p11.cpp
index 196f84c557ea..b7b9fcb23144 100644
--- a/test/CXX/except/except.spec/p11.cpp
+++ b/test/CXX/except/except.spec/p11.cpp
@@ -1,10 +1,10 @@
 // RUN: %clang_cc1 -std=c++11 -fexceptions -fcxx-exceptions -fsyntax-only -verify %s
 
 // This is the "let the user shoot themselves in the foot" clause.
-void f() noexcept { // expected-note {{non-throwing function declare here}}
+void f() noexcept { // expected-note {{function declared non-throwing here}}
   throw 0; // expected-warning {{has a non-throwing exception specification but}} 
 }
-void g() throw() { // expected-note {{non-throwing function declare here}}
+void g() throw() { // expected-note {{function declared non-throwing here}}
   throw 0; // expected-warning {{has a non-throwing exception specification but}} 
 }
 void h() throw(int) {
diff --git a/test/CXX/modules-ts/basic/basic.def.odr/p4/module.cpp b/test/CXX/modules-ts/basic/basic.def.odr/p4/module.cpp
new file mode 100644
index 000000000000..dc6a3635a8ee
--- /dev/null
+++ b/test/CXX/modules-ts/basic/basic.def.odr/p4/module.cpp
@@ -0,0 +1,46 @@
+// RUN: %clang_cc1 -fmodules-ts %S/module.cppm -triple %itanium_abi_triple -emit-module-interface -o %t
+// RUN: %clang_cc1 -fmodules-ts %s -triple %itanium_abi_triple -fmodule-file=%t -emit-llvm -o - | FileCheck %s --implicit-check-not=unused --implicit-check-not=global_module
+
+// CHECK-DAG: @extern_var_exported = external global
+// FIXME: Should this be 'external global'?
+// CHECK-DAG: @inline_var_exported = linkonce_odr global
+// CHECK-DAG: @_ZL19static_var_exported = external global
+// CHECK-DAG: @const_var_exported = external constant
+//
+// FIXME: The module name should be mangled into all of these.
+// CHECK-DAG: @extern_var_module_linkage = external global
+// FIXME: Should this be 'external global'?
+// CHECK-DAG: @inline_var_module_linkage = linkonce_odr global
+// CHECK-DAG: @_ZL25static_var_module_linkage = external global
+// CHECK-DAG: @_ZL24const_var_module_linkage = external constant
+
+module Module;
+
+void use() {
+  // CHECK: define linkonce_odr {{.*}}@_Z20used_inline_exportedv
+  used_inline_exported();
+  // CHECK: declare {{.*}}@_Z18noninline_exportedv
+  noninline_exported();
+
+  (void)&extern_var_exported;
+  (void)&inline_var_exported;
+  (void)&static_var_exported; // FIXME: Should not be exported.
+  (void)&const_var_exported;
+
+  // FIXME: This symbol should not be visible here.
+  // CHECK: declare {{.*}}@_ZL26used_static_module_linkagev
+  used_static_module_linkage();
+
+  // FIXME: The module name should be mangled into the name of this function.
+  // CHECK: define linkonce_odr {{.*}}@_Z26used_inline_module_linkagev
+  used_inline_module_linkage();
+
+  // FIXME: The module name should be mangled into the name of this function.
+  // CHECK: declare {{.*}}@_Z24noninline_module_linkagev
+  noninline_module_linkage();
+
+  (void)&extern_var_module_linkage;
+  (void)&inline_var_module_linkage;
+  (void)&static_var_module_linkage; // FIXME: Should not be visible here.
+  (void)&const_var_module_linkage;
+}
diff --git a/test/CXX/modules-ts/basic/basic.def.odr/p4/module.cppm b/test/CXX/modules-ts/basic/basic.def.odr/p4/module.cppm
new file mode 100644
index 000000000000..d452f741a0fb
--- /dev/null
+++ b/test/CXX/modules-ts/basic/basic.def.odr/p4/module.cppm
@@ -0,0 +1,118 @@
+// RUN: %clang_cc1 -fmodules-ts %s -triple %itanium_abi_triple -emit-llvm -o - | FileCheck %s --implicit-check-not unused_inline --implicit-check-not unused_stastic_global_module
+
+// CHECK-DAG: @extern_var_global_module = external global
+// CHECK-DAG: @inline_var_global_module = linkonce_odr global
+// CHECK-DAG: @_ZL24static_var_global_module = internal global
+// CHECK-DAG: @_ZL23const_var_global_module = internal constant
+//
+// For ABI compatibility, these symbols do not include the module name.
+// CHECK-DAG: @extern_var_exported = external global
+// FIXME: Should this be 'weak_odr global'? Presumably it must be, since we
+// can discard this global and its initializer (if any), and other TUs are not
+// permitted to run the initializer for this variable.
+// CHECK-DAG: @inline_var_exported = linkonce_odr global
+// CHECK-DAG: @_ZL19static_var_exported = global
+// CHECK-DAG: @const_var_exported = constant
+//
+// FIXME: The module name should be mangled into all of these.
+// CHECK-DAG: @extern_var_module_linkage = external global
+// FIXME: Should this be 'weak_odr global'? Presumably it must be, since we
+// can discard this global and its initializer (if any), and other TUs are not
+// permitted to run the initializer for this variable.
+// CHECK-DAG: @inline_var_module_linkage = linkonce_odr global
+// CHECK-DAG: @_ZL25static_var_module_linkage = global
+// CHECK-DAG: @_ZL24const_var_module_linkage = constant
+
+static void unused_static_global_module() {}
+static void used_static_global_module() {}
+
+inline void unused_inline_global_module() {}
+inline void used_inline_global_module() {}
+
+extern int extern_var_global_module;
+inline int inline_var_global_module;
+static int static_var_global_module;
+const int const_var_global_module = 3;
+
+// CHECK: define void {{.*}}@_Z23noninline_global_modulev
+void noninline_global_module() {
+  // FIXME: This should be promoted to module linkage and given a
+  // module-mangled name, if it's called from an inline function within
+  // the module interface.
+  // (We should try to avoid this when it's not reachable from outside
+  // the module interface unit.)
+  // CHECK: define internal {{.*}}@_ZL25used_static_global_modulev
+  used_static_global_module();
+  // CHECK: define linkonce_odr {{.*}}@_Z25used_inline_global_modulev
+  used_inline_global_module();
+
+  (void)&extern_var_global_module;
+  (void)&inline_var_global_module;
+  (void)&static_var_global_module;
+  (void)&const_var_global_module;
+}
+
+export module Module;
+
+export {
+  // FIXME: These should be ill-formed: you can't export an internal linkage
+  // symbol, per [dcl.module.interface]p2.
+  // CHECK: define void {{.*}}@_ZL22unused_static_exportedv
+  static void unused_static_exported() {}
+  // CHECK: define void {{.*}}@_ZL20used_static_exportedv
+  static void used_static_exported() {}
+
+  inline void unused_inline_exported() {}
+  inline void used_inline_exported() {}
+
+  extern int extern_var_exported;
+  inline int inline_var_exported;
+  // FIXME: This should be ill-formed: you can't export an internal linkage
+  // symbol.
+  static int static_var_exported;
+  const int const_var_exported = 3;
+
+  // CHECK: define void {{.*}}@_Z18noninline_exportedv
+  void noninline_exported() {
+    used_static_exported();
+    // CHECK: define linkonce_odr {{.*}}@_Z20used_inline_exportedv
+    used_inline_exported();
+
+    (void)&extern_var_exported;
+    (void)&inline_var_exported;
+    (void)&static_var_exported;
+    (void)&const_var_exported;
+  }
+}
+
+// FIXME: Ideally we wouldn't emit this as its name is not visible outside this
+// TU, but this module interface might contain a template that can use this
+// function so we conservatively emit it for now.
+// FIXME: The module name should be mangled into the name of this function.
+// CHECK: define void {{.*}}@_ZL28unused_static_module_linkagev
+static void unused_static_module_linkage() {}
+// FIXME: The module name should be mangled into the name of this function.
+// CHECK: define void {{.*}}@_ZL26used_static_module_linkagev
+static void used_static_module_linkage() {}
+
+inline void unused_inline_module_linkage() {}
+inline void used_inline_module_linkage() {}
+
+extern int extern_var_module_linkage;
+inline int inline_var_module_linkage;
+static int static_var_module_linkage;
+const int const_var_module_linkage = 3;
+
+// FIXME: The module name should be mangled into the name of this function.
+// CHECK: define void {{.*}}@_Z24noninline_module_linkagev
+void noninline_module_linkage() {
+  used_static_module_linkage();
+  // FIXME: The module name should be mangled into the name of this function.
+  // CHECK: define linkonce_odr {{.*}}@_Z26used_inline_module_linkagev
+  used_inline_module_linkage();
+
+  (void)&extern_var_module_linkage;
+  (void)&inline_var_module_linkage;
+  (void)&static_var_module_linkage;
+  (void)&const_var_module_linkage;
+}
diff --git a/test/CXX/modules-ts/basic/basic.def.odr/p4/user.cpp b/test/CXX/modules-ts/basic/basic.def.odr/p4/user.cpp
new file mode 100644
index 000000000000..f6e0238c6b4b
--- /dev/null
+++ b/test/CXX/modules-ts/basic/basic.def.odr/p4/user.cpp
@@ -0,0 +1,25 @@
+// RUN: %clang_cc1 -fmodules-ts %S/module.cppm -triple %itanium_abi_triple -emit-module-interface -o %t
+// RUN: %clang_cc1 -fmodules-ts %s -triple %itanium_abi_triple -fmodule-file=%t -emit-llvm -o - | FileCheck %s --implicit-check-not=unused --implicit-check-not=global_module
+
+// CHECK-DAG: @extern_var_exported = external global
+// FIXME: Should this be 'external global'?
+// CHECK-DAG: @inline_var_exported = linkonce_odr global
+// FIXME: These should be 'extern global' and 'extern constant'.
+// CHECK-DAG: @_ZL19static_var_exported = global
+// CHECK-DAG: @const_var_exported = constant
+
+import Module;
+
+void use() {
+  // CHECK: define linkonce_odr {{.*}}@_Z20used_inline_exportedv
+  used_inline_exported();
+  // CHECK: declare {{.*}}@_Z18noninline_exportedv
+  noninline_exported();
+
+  (void)&extern_var_exported;
+  (void)&inline_var_exported;
+  (void)&static_var_exported;
+  (void)&const_var_exported;
+
+  // Module-linkage declarations are not visible here.
+}
diff --git a/test/CXX/modules-ts/basic/basic.link/p2/module.cpp b/test/CXX/modules-ts/basic/basic.link/p2/module.cpp
new file mode 100644
index 000000000000..3fc6044d8e94
--- /dev/null
+++ b/test/CXX/modules-ts/basic/basic.link/p2/module.cpp
@@ -0,0 +1,17 @@
+// RUN: %clang_cc1 -std=c++1z -fmodules-ts %S/module.cppm -emit-module-interface -o %t
+// RUN: %clang_cc1 -std=c++1z -fmodules-ts -fmodule-file=%t %s -verify
+// expected-no-diagnostics
+module M;
+
+// FIXME: Use of internal linkage entities should be rejected.
+void use_from_module_impl() {
+  external_linkage_fn();
+  module_linkage_fn();
+  internal_linkage_fn();
+  (void)external_linkage_class{};
+  (void)module_linkage_class{};
+  (void)internal_linkage_class{};
+  (void)external_linkage_var;
+  (void)module_linkage_var;
+  (void)internal_linkage_var;
+}
diff --git a/test/CXX/modules-ts/basic/basic.link/p2/module.cppm b/test/CXX/modules-ts/basic/basic.link/p2/module.cppm
new file mode 100644
index 000000000000..bb261700db84
--- /dev/null
+++ b/test/CXX/modules-ts/basic/basic.link/p2/module.cppm
@@ -0,0 +1,29 @@
+// RUN: %clang_cc1 -std=c++1z -fmodules-ts %s -verify
+// expected-no-diagnostics
+export module M;
+
+export int external_linkage_var;
+int module_linkage_var;
+static int internal_linkage_var;
+
+export void external_linkage_fn() {}
+void module_linkage_fn() {}
+static void internal_linkage_fn() {}
+
+export struct external_linkage_class {};
+struct module_linkage_class {};
+namespace {
+  struct internal_linkage_class {};
+}
+
+void use() {
+  external_linkage_fn();
+  module_linkage_fn();
+  internal_linkage_fn();
+  (void)external_linkage_class{};
+  (void)module_linkage_class{};
+  (void)internal_linkage_class{};
+  (void)external_linkage_var;
+  (void)module_linkage_var;
+  (void)internal_linkage_var;
+}
diff --git a/test/CXX/modules-ts/basic/basic.link/p2/other.cpp b/test/CXX/modules-ts/basic/basic.link/p2/other.cpp
new file mode 100644
index 000000000000..8370777e7ed4
--- /dev/null
+++ b/test/CXX/modules-ts/basic/basic.link/p2/other.cpp
@@ -0,0 +1,16 @@
+// RUN: %clang_cc1 -std=c++1z -fmodules-ts %S/module.cppm -emit-module-interface -o %t
+// RUN: %clang_cc1 -std=c++1z -fmodules-ts -fmodule-file=%t %s -verify
+import M;
+
+void use_from_module_impl() {
+  external_linkage_fn();
+  module_linkage_fn(); // expected-error {{undeclared identifier}}
+  internal_linkage_fn(); // expected-error {{undeclared identifier}}
+  (void)external_linkage_class{};
+  (void)module_linkage_class{}; // expected-error {{undeclared identifier}} expected-error 0+{{}}
+  (void)internal_linkage_class{}; // expected-error {{undeclared identifier}} expected-error 0+{{}}
+  // expected-note@module.cppm:9 {{here}}
+  (void)external_linkage_var;
+  (void)module_linkage_var; // expected-error {{undeclared identifier}}
+  (void)internal_linkage_var; // expected-error {{undeclared identifier}}
+}
diff --git a/test/CXX/modules-ts/dcl.dcl/dcl.module/dcl.module.import/p1.cpp b/test/CXX/modules-ts/dcl.dcl/dcl.module/dcl.module.import/p1.cpp
index aaf43d6584a4..aad31b4b46d1 100644
--- a/test/CXX/modules-ts/dcl.dcl/dcl.module/dcl.module.import/p1.cpp
+++ b/test/CXX/modules-ts/dcl.dcl/dcl.module/dcl.module.import/p1.cpp
@@ -1,7 +1,7 @@
 // RUN: rm -rf %t
 // RUN: mkdir -p %t
-// RUN: echo 'export module x; int a, b;' > %t/x.cppm
-// RUN: echo 'export module x.y; int c;' > %t/x.y.cppm
+// RUN: echo 'export module x; export int a, b;' > %t/x.cppm
+// RUN: echo 'export module x.y; export int c;' > %t/x.y.cppm
 //
 // RUN: %clang_cc1 -std=c++1z -fmodules-ts -emit-module-interface %t/x.cppm -o %t/x.pcm
 // RUN: %clang_cc1 -std=c++1z -fmodules-ts -emit-module-interface -fmodule-file=%t/x.pcm %t/x.y.cppm -o %t/x.y.pcm
diff --git a/test/CXX/modules-ts/dcl.dcl/dcl.module/p5.cpp b/test/CXX/modules-ts/dcl.dcl/dcl.module/p5.cpp
new file mode 100644
index 000000000000..734b89173df9
--- /dev/null
+++ b/test/CXX/modules-ts/dcl.dcl/dcl.module/p5.cpp
@@ -0,0 +1,33 @@
+// RUN: rm -f %t
+// RUN: %clang_cc1 -std=c++1z -fmodules-ts -emit-module-interface %s -o %t -DINTERFACE
+// RUN: %clang_cc1 -std=c++1z -fmodules-ts -fmodule-file=%t %s -verify -DIMPLEMENTATION
+// RUN: %clang_cc1 -std=c++1z -fmodules-ts -fmodule-file=%t %s -verify -DEARLY_IMPLEMENTATION
+// RUN: %clang_cc1 -std=c++1z -fmodules-ts -fmodule-file=%t %s -verify -DUSER
+
+// expected-no-diagnostics
+
+#ifdef USER
+import Foo;
+#endif
+
+#ifdef EARLY_IMPLEMENTATION
+module Foo;
+#endif
+
+template<typename T> struct type_template {
+  typedef T type;
+  void f(type);
+};
+
+template<typename T> void type_template<T>::f(type) {}
+
+template<int = 0, typename = int, template<typename> class = type_template>
+struct default_template_args {};
+
+#ifdef INTERFACE
+export module Foo;
+#endif
+
+#ifdef IMPLEMENTATION
+module Foo;
+#endif
diff --git a/test/CodeGen/aarch64-neon-intrinsics.c b/test/CodeGen/aarch64-neon-intrinsics.c
index cbc2e72fcbac..bcb680c4b518 100644
--- a/test/CodeGen/aarch64-neon-intrinsics.c
+++ b/test/CodeGen/aarch64-neon-intrinsics.c
@@ -9037,9 +9037,10 @@ int64x2_t test_vld1q_s64(int64_t const *a) {
 
 // CHECK-LABEL: @test_vld1q_f16(
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP1:%.*]] = bitcast i8* [[TMP0]] to <8 x half>*
-// CHECK:   [[TMP2:%.*]] = load <8 x half>, <8 x half>* [[TMP1]]
-// CHECK:   ret <8 x half> [[TMP2]]
+// CHECK:   [[TMP1:%.*]] = bitcast i8* [[TMP0]] to <8 x i16>*
+// CHECK:   [[TMP2:%.*]] = load <8 x i16>, <8 x i16>* [[TMP1]]
+// CHECK:   [[TMP3:%.*]] = bitcast <8 x i16> [[TMP2]] to <8 x half>
+// CHECK:   ret <8 x half> [[TMP3]]
 float16x8_t test_vld1q_f16(float16_t const *a) {
   return vld1q_f16(a);
 }
@@ -9151,9 +9152,10 @@ int64x1_t test_vld1_s64(int64_t const *a) {
 
 // CHECK-LABEL: @test_vld1_f16(
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP1:%.*]] = bitcast i8* [[TMP0]] to <4 x half>*
-// CHECK:   [[TMP2:%.*]] = load <4 x half>, <4 x half>* [[TMP1]]
-// CHECK:   ret <4 x half> [[TMP2]]
+// CHECK:   [[TMP1:%.*]] = bitcast i8* [[TMP0]] to <4 x i16>*
+// CHECK:   [[TMP2:%.*]] = load <4 x i16>, <4 x i16>* [[TMP1]]
+// CHECK:   [[TMP3:%.*]] = bitcast <4 x i16> [[TMP2]] to <4 x half>
+// CHECK:   ret <4 x half> [[TMP3]]
 float16x4_t test_vld1_f16(float16_t const *a) {
   return vld1_f16(a);
 }
@@ -9340,10 +9342,10 @@ int64x2x2_t test_vld2q_s64(int64_t const *a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x8x2_t, align 16
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x8x2_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to <8 x half>*
-// CHECK:   [[VLD2:%.*]] = call { <8 x half>, <8 x half> } @llvm.aarch64.neon.ld2.v8f16.p0v8f16(<8 x half>* [[TMP2]])
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <8 x half>, <8 x half> }*
-// CHECK:   store { <8 x half>, <8 x half> } [[VLD2]], { <8 x half>, <8 x half> }* [[TMP3]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to <8 x i16>*
+// CHECK:   [[VLD2:%.*]] = call { <8 x i16>, <8 x i16> } @llvm.aarch64.neon.ld2.v8i16.p0v8i16(<8 x i16>* [[TMP2]])
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <8 x i16>, <8 x i16> }*
+// CHECK:   store { <8 x i16>, <8 x i16> } [[VLD2]], { <8 x i16>, <8 x i16> }* [[TMP3]]
 // CHECK:   [[TMP4:%.*]] = bitcast %struct.float16x8x2_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x8x2_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false)
@@ -9571,10 +9573,10 @@ int64x1x2_t test_vld2_s64(int64_t const *a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x4x2_t, align 8
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x4x2_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to <4 x half>*
-// CHECK:   [[VLD2:%.*]] = call { <4 x half>, <4 x half> } @llvm.aarch64.neon.ld2.v4f16.p0v4f16(<4 x half>* [[TMP2]])
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <4 x half>, <4 x half> }*
-// CHECK:   store { <4 x half>, <4 x half> } [[VLD2]], { <4 x half>, <4 x half> }* [[TMP3]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to <4 x i16>*
+// CHECK:   [[VLD2:%.*]] = call { <4 x i16>, <4 x i16> } @llvm.aarch64.neon.ld2.v4i16.p0v4i16(<4 x i16>* [[TMP2]])
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <4 x i16>, <4 x i16> }*
+// CHECK:   store { <4 x i16>, <4 x i16> } [[VLD2]], { <4 x i16>, <4 x i16> }* [[TMP3]]
 // CHECK:   [[TMP4:%.*]] = bitcast %struct.float16x4x2_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x4x2_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false)
@@ -9802,10 +9804,10 @@ int64x2x3_t test_vld3q_s64(int64_t const *a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x8x3_t, align 16
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x8x3_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to <8 x half>*
-// CHECK:   [[VLD3:%.*]] = call { <8 x half>, <8 x half>, <8 x half> } @llvm.aarch64.neon.ld3.v8f16.p0v8f16(<8 x half>* [[TMP2]])
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <8 x half>, <8 x half>, <8 x half> }*
-// CHECK:   store { <8 x half>, <8 x half>, <8 x half> } [[VLD3]], { <8 x half>, <8 x half>, <8 x half> }* [[TMP3]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to <8 x i16>*
+// CHECK:   [[VLD3:%.*]] = call { <8 x i16>, <8 x i16>, <8 x i16> } @llvm.aarch64.neon.ld3.v8i16.p0v8i16(<8 x i16>* [[TMP2]])
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <8 x i16>, <8 x i16>, <8 x i16> }*
+// CHECK:   store { <8 x i16>, <8 x i16>, <8 x i16> } [[VLD3]], { <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP3]]
 // CHECK:   [[TMP4:%.*]] = bitcast %struct.float16x8x3_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x8x3_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false)
@@ -10033,10 +10035,10 @@ int64x1x3_t test_vld3_s64(int64_t const *a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x4x3_t, align 8
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x4x3_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to <4 x half>*
-// CHECK:   [[VLD3:%.*]] = call { <4 x half>, <4 x half>, <4 x half> } @llvm.aarch64.neon.ld3.v4f16.p0v4f16(<4 x half>* [[TMP2]])
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <4 x half>, <4 x half>, <4 x half> }*
-// CHECK:   store { <4 x half>, <4 x half>, <4 x half> } [[VLD3]], { <4 x half>, <4 x half>, <4 x half> }* [[TMP3]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to <4 x i16>*
+// CHECK:   [[VLD3:%.*]] = call { <4 x i16>, <4 x i16>, <4 x i16> } @llvm.aarch64.neon.ld3.v4i16.p0v4i16(<4 x i16>* [[TMP2]])
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <4 x i16>, <4 x i16>, <4 x i16> }*
+// CHECK:   store { <4 x i16>, <4 x i16>, <4 x i16> } [[VLD3]], { <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP3]]
 // CHECK:   [[TMP4:%.*]] = bitcast %struct.float16x4x3_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x4x3_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false)
@@ -10264,10 +10266,10 @@ int64x2x4_t test_vld4q_s64(int64_t const *a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x8x4_t, align 16
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x8x4_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to <8 x half>*
-// CHECK:   [[VLD4:%.*]] = call { <8 x half>, <8 x half>, <8 x half>, <8 x half> } @llvm.aarch64.neon.ld4.v8f16.p0v8f16(<8 x half>* [[TMP2]])
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <8 x half>, <8 x half>, <8 x half>, <8 x half> }*
-// CHECK:   store { <8 x half>, <8 x half>, <8 x half>, <8 x half> } [[VLD4]], { <8 x half>, <8 x half>, <8 x half>, <8 x half> }* [[TMP3]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to <8 x i16>*
+// CHECK:   [[VLD4:%.*]] = call { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> } @llvm.aarch64.neon.ld4.v8i16.p0v8i16(<8 x i16>* [[TMP2]])
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> }*
+// CHECK:   store { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> } [[VLD4]], { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP3]]
 // CHECK:   [[TMP4:%.*]] = bitcast %struct.float16x8x4_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x8x4_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false)
@@ -10495,10 +10497,10 @@ int64x1x4_t test_vld4_s64(int64_t const *a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x4x4_t, align 8
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x4x4_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to <4 x half>*
-// CHECK:   [[VLD4:%.*]] = call { <4 x half>, <4 x half>, <4 x half>, <4 x half> } @llvm.aarch64.neon.ld4.v4f16.p0v4f16(<4 x half>* [[TMP2]])
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <4 x half>, <4 x half>, <4 x half>, <4 x half> }*
-// CHECK:   store { <4 x half>, <4 x half>, <4 x half>, <4 x half> } [[VLD4]], { <4 x half>, <4 x half>, <4 x half>, <4 x half> }* [[TMP3]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to <4 x i16>*
+// CHECK:   [[VLD4:%.*]] = call { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> } @llvm.aarch64.neon.ld4.v4i16.p0v4i16(<4 x i16>* [[TMP2]])
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> }*
+// CHECK:   store { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> } [[VLD4]], { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP3]]
 // CHECK:   [[TMP4:%.*]] = bitcast %struct.float16x4x4_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x4x4_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false)
@@ -10664,9 +10666,9 @@ void test_vst1q_s64(int64_t *a, int64x2_t b) {
 // CHECK-LABEL: @test_vst1q_f16(
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast <8 x half> %b to <16 x i8>
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP0]] to <8 x half>*
-// CHECK:   [[TMP3:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x half>
-// CHECK:   store <8 x half> [[TMP3]], <8 x half>* [[TMP2]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP0]] to <8 x i16>*
+// CHECK:   [[TMP3:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x i16>
+// CHECK:   store <8 x i16> [[TMP3]], <8 x i16>* [[TMP2]]
 // CHECK:   ret void
 void test_vst1q_f16(float16_t *a, float16x8_t b) {
   vst1q_f16(a, b);
@@ -10798,9 +10800,9 @@ void test_vst1_s64(int64_t *a, int64x1_t b) {
 // CHECK-LABEL: @test_vst1_f16(
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast <4 x half> %b to <8 x i8>
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP0]] to <4 x half>*
-// CHECK:   [[TMP3:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x half>
-// CHECK:   store <4 x half> [[TMP3]], <4 x half>* [[TMP2]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP0]] to <4 x i16>*
+// CHECK:   [[TMP3:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x i16>
+// CHECK:   store <4 x i16> [[TMP3]], <4 x i16>* [[TMP2]]
 // CHECK:   ret void
 void test_vst1_f16(float16_t *a, float16x4_t b) {
   vst1_f16(a, b);
@@ -11054,9 +11056,9 @@ void test_vst2q_s64(int64_t *a, int64x2x2_t b) {
 // CHECK:   [[ARRAYIDX2:%.*]] = getelementptr inbounds [2 x <8 x half>], [2 x <8 x half>]* [[VAL1]], i64 0, i64 1
 // CHECK:   [[TMP5:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX2]], align 16
 // CHECK:   [[TMP6:%.*]] = bitcast <8 x half> [[TMP5]] to <16 x i8>
-// CHECK:   [[TMP7:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x half>
-// CHECK:   [[TMP8:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x half>
-// CHECK:   call void @llvm.aarch64.neon.st2.v8f16.p0i8(<8 x half> [[TMP7]], <8 x half> [[TMP8]], i8* [[TMP2]])
+// CHECK:   [[TMP7:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
+// CHECK:   [[TMP8:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x i16>
+// CHECK:   call void @llvm.aarch64.neon.st2.v8i16.p0i8(<8 x i16> [[TMP7]], <8 x i16> [[TMP8]], i8* [[TMP2]])
 // CHECK:   ret void
 void test_vst2q_f16(float16_t *a, float16x8x2_t b) {
   vst2q_f16(a, b);
@@ -11364,9 +11366,9 @@ void test_vst2_s64(int64_t *a, int64x1x2_t b) {
 // CHECK:   [[ARRAYIDX2:%.*]] = getelementptr inbounds [2 x <4 x half>], [2 x <4 x half>]* [[VAL1]], i64 0, i64 1
 // CHECK:   [[TMP5:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX2]], align 8
 // CHECK:   [[TMP6:%.*]] = bitcast <4 x half> [[TMP5]] to <8 x i8>
-// CHECK:   [[TMP7:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x half>
-// CHECK:   [[TMP8:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x half>
-// CHECK:   call void @llvm.aarch64.neon.st2.v4f16.p0i8(<4 x half> [[TMP7]], <4 x half> [[TMP8]], i8* [[TMP2]])
+// CHECK:   [[TMP7:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
+// CHECK:   [[TMP8:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x i16>
+// CHECK:   call void @llvm.aarch64.neon.st2.v4i16.p0i8(<4 x i16> [[TMP7]], <4 x i16> [[TMP8]], i8* [[TMP2]])
 // CHECK:   ret void
 void test_vst2_f16(float16_t *a, float16x4x2_t b) {
   vst2_f16(a, b);
@@ -11714,10 +11716,10 @@ void test_vst3q_s64(int64_t *a, int64x2x3_t b) {
 // CHECK:   [[ARRAYIDX4:%.*]] = getelementptr inbounds [3 x <8 x half>], [3 x <8 x half>]* [[VAL3]], i64 0, i64 2
 // CHECK:   [[TMP7:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX4]], align 16
 // CHECK:   [[TMP8:%.*]] = bitcast <8 x half> [[TMP7]] to <16 x i8>
-// CHECK:   [[TMP9:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x half>
-// CHECK:   [[TMP10:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x half>
-// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP8]] to <8 x half>
-// CHECK:   call void @llvm.aarch64.neon.st3.v8f16.p0i8(<8 x half> [[TMP9]], <8 x half> [[TMP10]], <8 x half> [[TMP11]], i8* [[TMP2]])
+// CHECK:   [[TMP9:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
+// CHECK:   [[TMP10:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x i16>
+// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP8]] to <8 x i16>
+// CHECK:   call void @llvm.aarch64.neon.st3.v8i16.p0i8(<8 x i16> [[TMP9]], <8 x i16> [[TMP10]], <8 x i16> [[TMP11]], i8* [[TMP2]])
 // CHECK:   ret void
 void test_vst3q_f16(float16_t *a, float16x8x3_t b) {
   vst3q_f16(a, b);
@@ -12083,10 +12085,10 @@ void test_vst3_s64(int64_t *a, int64x1x3_t b) {
 // CHECK:   [[ARRAYIDX4:%.*]] = getelementptr inbounds [3 x <4 x half>], [3 x <4 x half>]* [[VAL3]], i64 0, i64 2
 // CHECK:   [[TMP7:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX4]], align 8
 // CHECK:   [[TMP8:%.*]] = bitcast <4 x half> [[TMP7]] to <8 x i8>
-// CHECK:   [[TMP9:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x half>
-// CHECK:   [[TMP10:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x half>
-// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP8]] to <4 x half>
-// CHECK:   call void @llvm.aarch64.neon.st3.v4f16.p0i8(<4 x half> [[TMP9]], <4 x half> [[TMP10]], <4 x half> [[TMP11]], i8* [[TMP2]])
+// CHECK:   [[TMP9:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
+// CHECK:   [[TMP10:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x i16>
+// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP8]] to <4 x i16>
+// CHECK:   call void @llvm.aarch64.neon.st3.v4i16.p0i8(<4 x i16> [[TMP9]], <4 x i16> [[TMP10]], <4 x i16> [[TMP11]], i8* [[TMP2]])
 // CHECK:   ret void
 void test_vst3_f16(float16_t *a, float16x4x3_t b) {
   vst3_f16(a, b);
@@ -12492,11 +12494,11 @@ void test_vst4q_s64(int64_t *a, int64x2x4_t b) {
 // CHECK:   [[ARRAYIDX6:%.*]] = getelementptr inbounds [4 x <8 x half>], [4 x <8 x half>]* [[VAL5]], i64 0, i64 3
 // CHECK:   [[TMP9:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX6]], align 16
 // CHECK:   [[TMP10:%.*]] = bitcast <8 x half> [[TMP9]] to <16 x i8>
-// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x half>
-// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x half>
-// CHECK:   [[TMP13:%.*]] = bitcast <16 x i8> [[TMP8]] to <8 x half>
-// CHECK:   [[TMP14:%.*]] = bitcast <16 x i8> [[TMP10]] to <8 x half>
-// CHECK:   call void @llvm.aarch64.neon.st4.v8f16.p0i8(<8 x half> [[TMP11]], <8 x half> [[TMP12]], <8 x half> [[TMP13]], <8 x half> [[TMP14]], i8* [[TMP2]])
+// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
+// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x i16>
+// CHECK:   [[TMP13:%.*]] = bitcast <16 x i8> [[TMP8]] to <8 x i16>
+// CHECK:   [[TMP14:%.*]] = bitcast <16 x i8> [[TMP10]] to <8 x i16>
+// CHECK:   call void @llvm.aarch64.neon.st4.v8i16.p0i8(<8 x i16> [[TMP11]], <8 x i16> [[TMP12]], <8 x i16> [[TMP13]], <8 x i16> [[TMP14]], i8* [[TMP2]])
 // CHECK:   ret void
 void test_vst4q_f16(float16_t *a, float16x8x4_t b) {
   vst4q_f16(a, b);
@@ -12920,11 +12922,11 @@ void test_vst4_s64(int64_t *a, int64x1x4_t b) {
 // CHECK:   [[ARRAYIDX6:%.*]] = getelementptr inbounds [4 x <4 x half>], [4 x <4 x half>]* [[VAL5]], i64 0, i64 3
 // CHECK:   [[TMP9:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX6]], align 8
 // CHECK:   [[TMP10:%.*]] = bitcast <4 x half> [[TMP9]] to <8 x i8>
-// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x half>
-// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x half>
-// CHECK:   [[TMP13:%.*]] = bitcast <8 x i8> [[TMP8]] to <4 x half>
-// CHECK:   [[TMP14:%.*]] = bitcast <8 x i8> [[TMP10]] to <4 x half>
-// CHECK:   call void @llvm.aarch64.neon.st4.v4f16.p0i8(<4 x half> [[TMP11]], <4 x half> [[TMP12]], <4 x half> [[TMP13]], <4 x half> [[TMP14]], i8* [[TMP2]])
+// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
+// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x i16>
+// CHECK:   [[TMP13:%.*]] = bitcast <8 x i8> [[TMP8]] to <4 x i16>
+// CHECK:   [[TMP14:%.*]] = bitcast <8 x i8> [[TMP10]] to <4 x i16>
+// CHECK:   call void @llvm.aarch64.neon.st4.v4i16.p0i8(<4 x i16> [[TMP11]], <4 x i16> [[TMP12]], <4 x i16> [[TMP13]], <4 x i16> [[TMP14]], i8* [[TMP2]])
 // CHECK:   ret void
 void test_vst4_f16(float16_t *a, float16x4x4_t b) {
   vst4_f16(a, b);
@@ -13206,10 +13208,10 @@ int64x2x2_t test_vld1q_s64_x2(int64_t const *a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x8x2_t, align 16
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x8x2_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to half*
-// CHECK:   [[VLD1XN:%.*]] = call { <8 x half>, <8 x half> } @llvm.aarch64.neon.ld1x2.v8f16.p0f16(half* [[TMP2]])
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <8 x half>, <8 x half> }*
-// CHECK:   store { <8 x half>, <8 x half> } [[VLD1XN]], { <8 x half>, <8 x half> }* [[TMP3]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to i16*
+// CHECK:   [[VLD1XN:%.*]] = call { <8 x i16>, <8 x i16> } @llvm.aarch64.neon.ld1x2.v8i16.p0i16(i16* [[TMP2]])
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <8 x i16>, <8 x i16> }*
+// CHECK:   store { <8 x i16>, <8 x i16> } [[VLD1XN]], { <8 x i16>, <8 x i16> }* [[TMP3]]
 // CHECK:   [[TMP4:%.*]] = bitcast %struct.float16x8x2_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x8x2_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false)
@@ -13452,10 +13454,10 @@ int64x1x2_t test_vld1_s64_x2(int64_t const *a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x4x2_t, align 8
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x4x2_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to half*
-// CHECK:   [[VLD1XN:%.*]] = call { <4 x half>, <4 x half> } @llvm.aarch64.neon.ld1x2.v4f16.p0f16(half* [[TMP2]])
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <4 x half>, <4 x half> }*
-// CHECK:   store { <4 x half>, <4 x half> } [[VLD1XN]], { <4 x half>, <4 x half> }* [[TMP3]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to i16*
+// CHECK:   [[VLD1XN:%.*]] = call { <4 x i16>, <4 x i16> } @llvm.aarch64.neon.ld1x2.v4i16.p0i16(i16* [[TMP2]])
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <4 x i16>, <4 x i16> }*
+// CHECK:   store { <4 x i16>, <4 x i16> } [[VLD1XN]], { <4 x i16>, <4 x i16> }* [[TMP3]]
 // CHECK:   [[TMP4:%.*]] = bitcast %struct.float16x4x2_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x4x2_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false)
@@ -13698,10 +13700,10 @@ int64x2x3_t test_vld1q_s64_x3(int64_t const *a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x8x3_t, align 16
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x8x3_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to half*
-// CHECK:   [[VLD1XN:%.*]] = call { <8 x half>, <8 x half>, <8 x half> } @llvm.aarch64.neon.ld1x3.v8f16.p0f16(half* [[TMP2]])
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <8 x half>, <8 x half>, <8 x half> }*
-// CHECK:   store { <8 x half>, <8 x half>, <8 x half> } [[VLD1XN]], { <8 x half>, <8 x half>, <8 x half> }* [[TMP3]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to i16*
+// CHECK:   [[VLD1XN:%.*]] = call { <8 x i16>, <8 x i16>, <8 x i16> } @llvm.aarch64.neon.ld1x3.v8i16.p0i16(i16* [[TMP2]])
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <8 x i16>, <8 x i16>, <8 x i16> }*
+// CHECK:   store { <8 x i16>, <8 x i16>, <8 x i16> } [[VLD1XN]], { <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP3]]
 // CHECK:   [[TMP4:%.*]] = bitcast %struct.float16x8x3_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x8x3_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false)
@@ -13944,10 +13946,10 @@ int64x1x3_t test_vld1_s64_x3(int64_t const *a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x4x3_t, align 8
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x4x3_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to half*
-// CHECK:   [[VLD1XN:%.*]] = call { <4 x half>, <4 x half>, <4 x half> } @llvm.aarch64.neon.ld1x3.v4f16.p0f16(half* [[TMP2]])
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <4 x half>, <4 x half>, <4 x half> }*
-// CHECK:   store { <4 x half>, <4 x half>, <4 x half> } [[VLD1XN]], { <4 x half>, <4 x half>, <4 x half> }* [[TMP3]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to i16*
+// CHECK:   [[VLD1XN:%.*]] = call { <4 x i16>, <4 x i16>, <4 x i16> } @llvm.aarch64.neon.ld1x3.v4i16.p0i16(i16* [[TMP2]])
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <4 x i16>, <4 x i16>, <4 x i16> }*
+// CHECK:   store { <4 x i16>, <4 x i16>, <4 x i16> } [[VLD1XN]], { <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP3]]
 // CHECK:   [[TMP4:%.*]] = bitcast %struct.float16x4x3_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x4x3_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false)
@@ -14190,10 +14192,10 @@ int64x2x4_t test_vld1q_s64_x4(int64_t const *a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x8x4_t, align 16
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x8x4_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to half*
-// CHECK:   [[VLD1XN:%.*]] = call { <8 x half>, <8 x half>, <8 x half>, <8 x half> } @llvm.aarch64.neon.ld1x4.v8f16.p0f16(half* [[TMP2]])
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <8 x half>, <8 x half>, <8 x half>, <8 x half> }*
-// CHECK:   store { <8 x half>, <8 x half>, <8 x half>, <8 x half> } [[VLD1XN]], { <8 x half>, <8 x half>, <8 x half>, <8 x half> }* [[TMP3]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to i16*
+// CHECK:   [[VLD1XN:%.*]] = call { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> } @llvm.aarch64.neon.ld1x4.v8i16.p0i16(i16* [[TMP2]])
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> }*
+// CHECK:   store { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> } [[VLD1XN]], { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP3]]
 // CHECK:   [[TMP4:%.*]] = bitcast %struct.float16x8x4_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x8x4_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false)
@@ -14436,10 +14438,10 @@ int64x1x4_t test_vld1_s64_x4(int64_t const *a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x4x4_t, align 8
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x4x4_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to half*
-// CHECK:   [[VLD1XN:%.*]] = call { <4 x half>, <4 x half>, <4 x half>, <4 x half> } @llvm.aarch64.neon.ld1x4.v4f16.p0f16(half* [[TMP2]])
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <4 x half>, <4 x half>, <4 x half>, <4 x half> }*
-// CHECK:   store { <4 x half>, <4 x half>, <4 x half>, <4 x half> } [[VLD1XN]], { <4 x half>, <4 x half>, <4 x half>, <4 x half> }* [[TMP3]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to i16*
+// CHECK:   [[VLD1XN:%.*]] = call { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> } @llvm.aarch64.neon.ld1x4.v4i16.p0i16(i16* [[TMP2]])
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> }*
+// CHECK:   store { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> } [[VLD1XN]], { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP3]]
 // CHECK:   [[TMP4:%.*]] = bitcast %struct.float16x4x4_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x4x4_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false)
@@ -14750,10 +14752,10 @@ void test_vst1q_s64_x2(int64_t *a, int64x2x2_t b) {
 // CHECK:   [[ARRAYIDX2:%.*]] = getelementptr inbounds [2 x <8 x half>], [2 x <8 x half>]* [[VAL1]], i64 0, i64 1
 // CHECK:   [[TMP5:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX2]], align 16
 // CHECK:   [[TMP6:%.*]] = bitcast <8 x half> [[TMP5]] to <16 x i8>
-// CHECK:   [[TMP7:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x half>
-// CHECK:   [[TMP8:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x half>
-// CHECK:   [[TMP9:%.*]] = bitcast i8* [[TMP2]] to half*
-// CHECK:   call void @llvm.aarch64.neon.st1x2.v8f16.p0f16(<8 x half> [[TMP7]], <8 x half> [[TMP8]], half* [[TMP9]])
+// CHECK:   [[TMP7:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
+// CHECK:   [[TMP8:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x i16>
+// CHECK:   [[TMP9:%.*]] = bitcast i8* [[TMP2]] to i16*
+// CHECK:   call void @llvm.aarch64.neon.st1x2.v8i16.p0i16(<8 x i16> [[TMP7]], <8 x i16> [[TMP8]], i16* [[TMP9]])
 // CHECK:   ret void
 void test_vst1q_f16_x2(float16_t *a, float16x8x2_t b) {
   vst1q_f16_x2(a, b);
@@ -15096,10 +15098,10 @@ void test_vst1_s64_x2(int64_t *a, int64x1x2_t b) {
 // CHECK:   [[ARRAYIDX2:%.*]] = getelementptr inbounds [2 x <4 x half>], [2 x <4 x half>]* [[VAL1]], i64 0, i64 1
 // CHECK:   [[TMP5:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX2]], align 8
 // CHECK:   [[TMP6:%.*]] = bitcast <4 x half> [[TMP5]] to <8 x i8>
-// CHECK:   [[TMP7:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x half>
-// CHECK:   [[TMP8:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x half>
-// CHECK:   [[TMP9:%.*]] = bitcast i8* [[TMP2]] to half*
-// CHECK:   call void @llvm.aarch64.neon.st1x2.v4f16.p0f16(<4 x half> [[TMP7]], <4 x half> [[TMP8]], half* [[TMP9]])
+// CHECK:   [[TMP7:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
+// CHECK:   [[TMP8:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x i16>
+// CHECK:   [[TMP9:%.*]] = bitcast i8* [[TMP2]] to i16*
+// CHECK:   call void @llvm.aarch64.neon.st1x2.v4i16.p0i16(<4 x i16> [[TMP7]], <4 x i16> [[TMP8]], i16* [[TMP9]])
 // CHECK:   ret void
 void test_vst1_f16_x2(float16_t *a, float16x4x2_t b) {
   vst1_f16_x2(a, b);
@@ -15482,11 +15484,11 @@ void test_vst1q_s64_x3(int64_t *a, int64x2x3_t b) {
 // CHECK:   [[ARRAYIDX4:%.*]] = getelementptr inbounds [3 x <8 x half>], [3 x <8 x half>]* [[VAL3]], i64 0, i64 2
 // CHECK:   [[TMP7:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX4]], align 16
 // CHECK:   [[TMP8:%.*]] = bitcast <8 x half> [[TMP7]] to <16 x i8>
-// CHECK:   [[TMP9:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x half>
-// CHECK:   [[TMP10:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x half>
-// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP8]] to <8 x half>
-// CHECK:   [[TMP12:%.*]] = bitcast i8* [[TMP2]] to half*
-// CHECK:   call void @llvm.aarch64.neon.st1x3.v8f16.p0f16(<8 x half> [[TMP9]], <8 x half> [[TMP10]], <8 x half> [[TMP11]], half* [[TMP12]])
+// CHECK:   [[TMP9:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
+// CHECK:   [[TMP10:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x i16>
+// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP8]] to <8 x i16>
+// CHECK:   [[TMP12:%.*]] = bitcast i8* [[TMP2]] to i16*
+// CHECK:   call void @llvm.aarch64.neon.st1x3.v8i16.p0i16(<8 x i16> [[TMP9]], <8 x i16> [[TMP10]], <8 x i16> [[TMP11]], i16* [[TMP12]])
 // CHECK:   ret void
 void test_vst1q_f16_x3(float16_t *a, float16x8x3_t b) {
   vst1q_f16_x3(a, b);
@@ -15892,11 +15894,11 @@ void test_vst1_s64_x3(int64_t *a, int64x1x3_t b) {
 // CHECK:   [[ARRAYIDX4:%.*]] = getelementptr inbounds [3 x <4 x half>], [3 x <4 x half>]* [[VAL3]], i64 0, i64 2
 // CHECK:   [[TMP7:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX4]], align 8
 // CHECK:   [[TMP8:%.*]] = bitcast <4 x half> [[TMP7]] to <8 x i8>
-// CHECK:   [[TMP9:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x half>
-// CHECK:   [[TMP10:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x half>
-// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP8]] to <4 x half>
-// CHECK:   [[TMP12:%.*]] = bitcast i8* [[TMP2]] to half*
-// CHECK:   call void @llvm.aarch64.neon.st1x3.v4f16.p0f16(<4 x half> [[TMP9]], <4 x half> [[TMP10]], <4 x half> [[TMP11]], half* [[TMP12]])
+// CHECK:   [[TMP9:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
+// CHECK:   [[TMP10:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x i16>
+// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP8]] to <4 x i16>
+// CHECK:   [[TMP12:%.*]] = bitcast i8* [[TMP2]] to i16*
+// CHECK:   call void @llvm.aarch64.neon.st1x3.v4i16.p0i16(<4 x i16> [[TMP9]], <4 x i16> [[TMP10]], <4 x i16> [[TMP11]], i16* [[TMP12]])
 // CHECK:   ret void
 void test_vst1_f16_x3(float16_t *a, float16x4x3_t b) {
   vst1_f16_x3(a, b);
@@ -16342,12 +16344,12 @@ void test_vst1q_s64_x4(int64_t *a, int64x2x4_t b) {
 // CHECK:   [[ARRAYIDX6:%.*]] = getelementptr inbounds [4 x <8 x half>], [4 x <8 x half>]* [[VAL5]], i64 0, i64 3
 // CHECK:   [[TMP9:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX6]], align 16
 // CHECK:   [[TMP10:%.*]] = bitcast <8 x half> [[TMP9]] to <16 x i8>
-// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x half>
-// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x half>
-// CHECK:   [[TMP13:%.*]] = bitcast <16 x i8> [[TMP8]] to <8 x half>
-// CHECK:   [[TMP14:%.*]] = bitcast <16 x i8> [[TMP10]] to <8 x half>
-// CHECK:   [[TMP15:%.*]] = bitcast i8* [[TMP2]] to half*
-// CHECK:   call void @llvm.aarch64.neon.st1x4.v8f16.p0f16(<8 x half> [[TMP11]], <8 x half> [[TMP12]], <8 x half> [[TMP13]], <8 x half> [[TMP14]], half* [[TMP15]])
+// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
+// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x i16>
+// CHECK:   [[TMP13:%.*]] = bitcast <16 x i8> [[TMP8]] to <8 x i16>
+// CHECK:   [[TMP14:%.*]] = bitcast <16 x i8> [[TMP10]] to <8 x i16>
+// CHECK:   [[TMP15:%.*]] = bitcast i8* [[TMP2]] to i16*
+// CHECK:   call void @llvm.aarch64.neon.st1x4.v8i16.p0i16(<8 x i16> [[TMP11]], <8 x i16> [[TMP12]], <8 x i16> [[TMP13]], <8 x i16> [[TMP14]], i16* [[TMP15]])
 // CHECK:   ret void
 void test_vst1q_f16_x4(float16_t *a, float16x8x4_t b) {
   vst1q_f16_x4(a, b);
@@ -16816,12 +16818,12 @@ void test_vst1_s64_x4(int64_t *a, int64x1x4_t b) {
 // CHECK:   [[ARRAYIDX6:%.*]] = getelementptr inbounds [4 x <4 x half>], [4 x <4 x half>]* [[VAL5]], i64 0, i64 3
 // CHECK:   [[TMP9:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX6]], align 8
 // CHECK:   [[TMP10:%.*]] = bitcast <4 x half> [[TMP9]] to <8 x i8>
-// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x half>
-// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x half>
-// CHECK:   [[TMP13:%.*]] = bitcast <8 x i8> [[TMP8]] to <4 x half>
-// CHECK:   [[TMP14:%.*]] = bitcast <8 x i8> [[TMP10]] to <4 x half>
-// CHECK:   [[TMP15:%.*]] = bitcast i8* [[TMP2]] to half*
-// CHECK:   call void @llvm.aarch64.neon.st1x4.v4f16.p0f16(<4 x half> [[TMP11]], <4 x half> [[TMP12]], <4 x half> [[TMP13]], <4 x half> [[TMP14]], half* [[TMP15]])
+// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
+// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x i16>
+// CHECK:   [[TMP13:%.*]] = bitcast <8 x i8> [[TMP8]] to <4 x i16>
+// CHECK:   [[TMP14:%.*]] = bitcast <8 x i8> [[TMP10]] to <4 x i16>
+// CHECK:   [[TMP15:%.*]] = bitcast i8* [[TMP2]] to i16*
+// CHECK:   call void @llvm.aarch64.neon.st1x4.v4i16.p0i16(<4 x i16> [[TMP11]], <4 x i16> [[TMP12]], <4 x i16> [[TMP13]], <4 x i16> [[TMP14]], i16* [[TMP15]])
 // CHECK:   ret void
 void test_vst1_f16_x4(float16_t *a, float16x4x4_t b) {
   vst1_f16_x4(a, b);
diff --git a/test/CodeGen/aarch64-neon-ldst-one.c b/test/CodeGen/aarch64-neon-ldst-one.c
index a3c5b140a0d2..9bd9ab1cb61b 100644
--- a/test/CodeGen/aarch64-neon-ldst-one.c
+++ b/test/CodeGen/aarch64-neon-ldst-one.c
@@ -90,11 +90,12 @@ int64x2_t test_vld1q_dup_s64(int64_t  *a) {
 
 // CHECK-LABEL: define <8 x half> @test_vld1q_dup_f16(half* %a) #0 {
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP1:%.*]] = bitcast i8* [[TMP0]] to half*
-// CHECK:   [[TMP2:%.*]] = load half, half* [[TMP1]]
-// CHECK:   [[TMP3:%.*]] = insertelement <8 x half> undef, half [[TMP2]], i32 0
-// CHECK:   [[LANE:%.*]] = shufflevector <8 x half> [[TMP3]], <8 x half> [[TMP3]], <8 x i32> zeroinitializer
-// CHECK:   ret <8 x half> [[LANE]]
+// CHECK:   [[TMP1:%.*]] = bitcast i8* [[TMP0]] to i16*
+// CHECK:   [[TMP2:%.*]] = load i16, i16* [[TMP1]]
+// CHECK:   [[TMP3:%.*]] = insertelement <8 x i16> undef, i16 [[TMP2]], i32 0
+// CHECK:   [[LANE:%.*]] = shufflevector <8 x i16> [[TMP3]], <8 x i16> [[TMP3]], <8 x i32> zeroinitializer
+// CHECK:   [[TMP4:%.*]] = bitcast <8 x i16> [[LANE]] to <8 x half>
+// CHECK:   ret <8 x half> [[TMP4]]
 float16x8_t test_vld1q_dup_f16(float16_t  *a) {
   return vld1q_dup_f16(a);
 }
@@ -238,11 +239,12 @@ int64x1_t test_vld1_dup_s64(int64_t  *a) {
 
 // CHECK-LABEL: define <4 x half> @test_vld1_dup_f16(half* %a) #0 {
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP1:%.*]] = bitcast i8* [[TMP0]] to half*
-// CHECK:   [[TMP2:%.*]] = load half, half* [[TMP1]]
-// CHECK:   [[TMP3:%.*]] = insertelement <4 x half> undef, half [[TMP2]], i32 0
-// CHECK:   [[LANE:%.*]] = shufflevector <4 x half> [[TMP3]], <4 x half> [[TMP3]], <4 x i32> zeroinitializer
-// CHECK:   ret <4 x half> [[LANE]]
+// CHECK:   [[TMP1:%.*]] = bitcast i8* [[TMP0]] to i16*
+// CHECK:   [[TMP2:%.*]] = load i16, i16* [[TMP1]]
+// CHECK:   [[TMP3:%.*]] = insertelement <4 x i16> undef, i16 [[TMP2]], i32 0
+// CHECK:   [[LANE:%.*]] = shufflevector <4 x i16> [[TMP3]], <4 x i16> [[TMP3]], <4 x i32> zeroinitializer
+// CHECK:   [[TMP4:%.*]] = bitcast <4 x i16> [[LANE]] to <4 x half>
+// CHECK:   ret <4 x half> [[TMP4]]
 float16x4_t test_vld1_dup_f16(float16_t  *a) {
   return vld1_dup_f16(a);
 }
@@ -445,10 +447,10 @@ int64x2x2_t test_vld2q_dup_s64(int64_t  *a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x8x2_t, align 16
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x8x2_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to half*
-// CHECK:   [[VLD2:%.*]] = call { <8 x half>, <8 x half> } @llvm.aarch64.neon.ld2r.v8f16.p0f16(half* [[TMP2]])
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <8 x half>, <8 x half> }*
-// CHECK:   store { <8 x half>, <8 x half> } [[VLD2]], { <8 x half>, <8 x half> }* [[TMP3]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to i16*
+// CHECK:   [[VLD2:%.*]] = call { <8 x i16>, <8 x i16> } @llvm.aarch64.neon.ld2r.v8i16.p0i16(i16* [[TMP2]])
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <8 x i16>, <8 x i16> }*
+// CHECK:   store { <8 x i16>, <8 x i16> } [[VLD2]], { <8 x i16>, <8 x i16> }* [[TMP3]]
 // CHECK:   [[TMP4:%.*]] = bitcast %struct.float16x8x2_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x8x2_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false)
@@ -691,10 +693,10 @@ int64x1x2_t test_vld2_dup_s64(int64_t  *a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x4x2_t, align 8
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x4x2_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to half*
-// CHECK:   [[VLD2:%.*]] = call { <4 x half>, <4 x half> } @llvm.aarch64.neon.ld2r.v4f16.p0f16(half* [[TMP2]])
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <4 x half>, <4 x half> }*
-// CHECK:   store { <4 x half>, <4 x half> } [[VLD2]], { <4 x half>, <4 x half> }* [[TMP3]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to i16*
+// CHECK:   [[VLD2:%.*]] = call { <4 x i16>, <4 x i16> } @llvm.aarch64.neon.ld2r.v4i16.p0i16(i16* [[TMP2]])
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <4 x i16>, <4 x i16> }*
+// CHECK:   store { <4 x i16>, <4 x i16> } [[VLD2]], { <4 x i16>, <4 x i16> }* [[TMP3]]
 // CHECK:   [[TMP4:%.*]] = bitcast %struct.float16x4x2_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x4x2_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false)
@@ -945,10 +947,10 @@ int64x2x3_t test_vld3q_dup_s64(int64_t  *a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x8x3_t, align 16
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x8x3_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to half*
-// CHECK:   [[VLD3:%.*]] = call { <8 x half>, <8 x half>, <8 x half> } @llvm.aarch64.neon.ld3r.v8f16.p0f16(half* [[TMP2]])
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <8 x half>, <8 x half>, <8 x half> }*
-// CHECK:   store { <8 x half>, <8 x half>, <8 x half> } [[VLD3]], { <8 x half>, <8 x half>, <8 x half> }* [[TMP3]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to i16*
+// CHECK:   [[VLD3:%.*]] = call { <8 x i16>, <8 x i16>, <8 x i16> } @llvm.aarch64.neon.ld3r.v8i16.p0i16(i16* [[TMP2]])
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <8 x i16>, <8 x i16>, <8 x i16> }*
+// CHECK:   store { <8 x i16>, <8 x i16>, <8 x i16> } [[VLD3]], { <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP3]]
 // CHECK:   [[TMP4:%.*]] = bitcast %struct.float16x8x3_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x8x3_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false)
@@ -1205,10 +1207,10 @@ int64x1x3_t test_vld3_dup_s64(int64_t  *a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x4x3_t, align 8
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x4x3_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to half*
-// CHECK:   [[VLD3:%.*]] = call { <4 x half>, <4 x half>, <4 x half> } @llvm.aarch64.neon.ld3r.v4f16.p0f16(half* [[TMP2]])
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <4 x half>, <4 x half>, <4 x half> }*
-// CHECK:   store { <4 x half>, <4 x half>, <4 x half> } [[VLD3]], { <4 x half>, <4 x half>, <4 x half> }* [[TMP3]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to i16*
+// CHECK:   [[VLD3:%.*]] = call { <4 x i16>, <4 x i16>, <4 x i16> } @llvm.aarch64.neon.ld3r.v4i16.p0i16(i16* [[TMP2]])
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <4 x i16>, <4 x i16>, <4 x i16> }*
+// CHECK:   store { <4 x i16>, <4 x i16>, <4 x i16> } [[VLD3]], { <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP3]]
 // CHECK:   [[TMP4:%.*]] = bitcast %struct.float16x4x3_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x4x3_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false)
@@ -1457,10 +1459,10 @@ int64x2x4_t test_vld4q_dup_s64(int64_t  *a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x8x4_t, align 16
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x8x4_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to half*
-// CHECK:   [[VLD4:%.*]] = call { <8 x half>, <8 x half>, <8 x half>, <8 x half> } @llvm.aarch64.neon.ld4r.v8f16.p0f16(half* [[TMP2]])
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <8 x half>, <8 x half>, <8 x half>, <8 x half> }*
-// CHECK:   store { <8 x half>, <8 x half>, <8 x half>, <8 x half> } [[VLD4]], { <8 x half>, <8 x half>, <8 x half>, <8 x half> }* [[TMP3]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to i16*
+// CHECK:   [[VLD4:%.*]] = call { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> } @llvm.aarch64.neon.ld4r.v8i16.p0i16(i16* [[TMP2]])
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> }*
+// CHECK:   store { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> } [[VLD4]], { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP3]]
 // CHECK:   [[TMP4:%.*]] = bitcast %struct.float16x8x4_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x8x4_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false)
@@ -1703,10 +1705,10 @@ int64x1x4_t test_vld4_dup_s64(int64_t  *a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x4x4_t, align 8
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x4x4_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to half*
-// CHECK:   [[VLD4:%.*]] = call { <4 x half>, <4 x half>, <4 x half>, <4 x half> } @llvm.aarch64.neon.ld4r.v4f16.p0f16(half* [[TMP2]])
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <4 x half>, <4 x half>, <4 x half>, <4 x half> }*
-// CHECK:   store { <4 x half>, <4 x half>, <4 x half>, <4 x half> } [[VLD4]], { <4 x half>, <4 x half>, <4 x half>, <4 x half> }* [[TMP3]]
+// CHECK:   [[TMP2:%.*]] = bitcast i8* [[TMP1]] to i16*
+// CHECK:   [[VLD4:%.*]] = call { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> } @llvm.aarch64.neon.ld4r.v4i16.p0i16(i16* [[TMP2]])
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> }*
+// CHECK:   store { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> } [[VLD4]], { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP3]]
 // CHECK:   [[TMP4:%.*]] = bitcast %struct.float16x4x4_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x4x4_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false)
@@ -1895,11 +1897,12 @@ int64x2_t test_vld1q_lane_s64(int64_t  *a, int64x2_t b) {
 // CHECK-LABEL: define <8 x half> @test_vld1q_lane_f16(half* %a, <8 x half> %b) #0 {
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast <8 x half> %b to <16 x i8>
-// CHECK:   [[TMP2:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x half>
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to half*
-// CHECK:   [[TMP4:%.*]] = load half, half* [[TMP3]]
-// CHECK:   [[VLD1_LANE:%.*]] = insertelement <8 x half> [[TMP2]], half [[TMP4]], i32 7
-// CHECK:   ret <8 x half> [[VLD1_LANE]]
+// CHECK:   [[TMP2:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x i16>
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to i16*
+// CHECK:   [[TMP4:%.*]] = load i16, i16* [[TMP3]]
+// CHECK:   [[VLD1_LANE:%.*]] = insertelement <8 x i16> [[TMP2]], i16 [[TMP4]], i32 7
+// CHECK:   [[TMP5:%.*]] = bitcast <8 x i16> [[VLD1_LANE]] to <8 x half>
+// CHECK:   ret <8 x half> [[TMP5]]
 float16x8_t test_vld1q_lane_f16(float16_t  *a, float16x8_t b) {
   return vld1q_lane_f16(a, b, 7);
 }
@@ -2051,11 +2054,12 @@ int64x1_t test_vld1_lane_s64(int64_t  *a, int64x1_t b) {
 // CHECK-LABEL: define <4 x half> @test_vld1_lane_f16(half* %a, <4 x half> %b) #0 {
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast <4 x half> %b to <8 x i8>
-// CHECK:   [[TMP2:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x half>
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to half*
-// CHECK:   [[TMP4:%.*]] = load half, half* [[TMP3]]
-// CHECK:   [[VLD1_LANE:%.*]] = insertelement <4 x half> [[TMP2]], half [[TMP4]], i32 3
-// CHECK:   ret <4 x half> [[VLD1_LANE]]
+// CHECK:   [[TMP2:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x i16>
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to i16*
+// CHECK:   [[TMP4:%.*]] = load i16, i16* [[TMP3]]
+// CHECK:   [[VLD1_LANE:%.*]] = insertelement <4 x i16> [[TMP2]], i16 [[TMP4]], i32 3
+// CHECK:   [[TMP5:%.*]] = bitcast <4 x i16> [[VLD1_LANE]] to <4 x half>
+// CHECK:   ret <4 x half> [[TMP5]]
 float16x4_t test_vld1_lane_f16(float16_t  *a, float16x4_t b) {
   return vld1_lane_f16(a, b, 3);
 }
@@ -2491,11 +2495,11 @@ int64x2x2_t test_vld2q_lane_s64(int64_t  *a, int64x2x2_t b) {
 // CHECK:   [[ARRAYIDX2:%.*]] = getelementptr inbounds [2 x <8 x half>], [2 x <8 x half>]* [[VAL1]], i64 0, i64 1
 // CHECK:   [[TMP6:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX2]], align 16
 // CHECK:   [[TMP7:%.*]] = bitcast <8 x half> [[TMP6]] to <16 x i8>
-// CHECK:   [[TMP8:%.*]] = bitcast <16 x i8> [[TMP5]] to <8 x half>
-// CHECK:   [[TMP9:%.*]] = bitcast <16 x i8> [[TMP7]] to <8 x half>
-// CHECK:   [[VLD2_LANE:%.*]] = call { <8 x half>, <8 x half> } @llvm.aarch64.neon.ld2lane.v8f16.p0i8(<8 x half> [[TMP8]], <8 x half> [[TMP9]], i64 7, i8* [[TMP3]])
-// CHECK:   [[TMP10:%.*]] = bitcast i8* [[TMP2]] to { <8 x half>, <8 x half> }*
-// CHECK:   store { <8 x half>, <8 x half> } [[VLD2_LANE]], { <8 x half>, <8 x half> }* [[TMP10]]
+// CHECK:   [[TMP8:%.*]] = bitcast <16 x i8> [[TMP5]] to <8 x i16>
+// CHECK:   [[TMP9:%.*]] = bitcast <16 x i8> [[TMP7]] to <8 x i16>
+// CHECK:   [[VLD2_LANE:%.*]] = call { <8 x i16>, <8 x i16> } @llvm.aarch64.neon.ld2lane.v8i16.p0i8(<8 x i16> [[TMP8]], <8 x i16> [[TMP9]], i64 7, i8* [[TMP3]])
+// CHECK:   [[TMP10:%.*]] = bitcast i8* [[TMP2]] to { <8 x i16>, <8 x i16> }*
+// CHECK:   store { <8 x i16>, <8 x i16> } [[VLD2_LANE]], { <8 x i16>, <8 x i16> }* [[TMP10]]
 // CHECK:   [[TMP11:%.*]] = bitcast %struct.float16x8x2_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP12:%.*]] = bitcast %struct.float16x8x2_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP11]], i8* [[TMP12]], i64 32, i32 16, i1 false)
@@ -2923,11 +2927,11 @@ int64x1x2_t test_vld2_lane_s64(int64_t  *a, int64x1x2_t b) {
 // CHECK:   [[ARRAYIDX2:%.*]] = getelementptr inbounds [2 x <4 x half>], [2 x <4 x half>]* [[VAL1]], i64 0, i64 1
 // CHECK:   [[TMP6:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX2]], align 8
 // CHECK:   [[TMP7:%.*]] = bitcast <4 x half> [[TMP6]] to <8 x i8>
-// CHECK:   [[TMP8:%.*]] = bitcast <8 x i8> [[TMP5]] to <4 x half>
-// CHECK:   [[TMP9:%.*]] = bitcast <8 x i8> [[TMP7]] to <4 x half>
-// CHECK:   [[VLD2_LANE:%.*]] = call { <4 x half>, <4 x half> } @llvm.aarch64.neon.ld2lane.v4f16.p0i8(<4 x half> [[TMP8]], <4 x half> [[TMP9]], i64 3, i8* [[TMP3]])
-// CHECK:   [[TMP10:%.*]] = bitcast i8* [[TMP2]] to { <4 x half>, <4 x half> }*
-// CHECK:   store { <4 x half>, <4 x half> } [[VLD2_LANE]], { <4 x half>, <4 x half> }* [[TMP10]]
+// CHECK:   [[TMP8:%.*]] = bitcast <8 x i8> [[TMP5]] to <4 x i16>
+// CHECK:   [[TMP9:%.*]] = bitcast <8 x i8> [[TMP7]] to <4 x i16>
+// CHECK:   [[VLD2_LANE:%.*]] = call { <4 x i16>, <4 x i16> } @llvm.aarch64.neon.ld2lane.v4i16.p0i8(<4 x i16> [[TMP8]], <4 x i16> [[TMP9]], i64 3, i8* [[TMP3]])
+// CHECK:   [[TMP10:%.*]] = bitcast i8* [[TMP2]] to { <4 x i16>, <4 x i16> }*
+// CHECK:   store { <4 x i16>, <4 x i16> } [[VLD2_LANE]], { <4 x i16>, <4 x i16> }* [[TMP10]]
 // CHECK:   [[TMP11:%.*]] = bitcast %struct.float16x4x2_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP12:%.*]] = bitcast %struct.float16x4x2_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP11]], i8* [[TMP12]], i64 16, i32 8, i1 false)
@@ -3360,12 +3364,12 @@ int64x2x3_t test_vld3q_lane_s64(int64_t  *a, int64x2x3_t b) {
 // CHECK:   [[ARRAYIDX4:%.*]] = getelementptr inbounds [3 x <8 x half>], [3 x <8 x half>]* [[VAL3]], i64 0, i64 2
 // CHECK:   [[TMP8:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX4]], align 16
 // CHECK:   [[TMP9:%.*]] = bitcast <8 x half> [[TMP8]] to <16 x i8>
-// CHECK:   [[TMP10:%.*]] = bitcast <16 x i8> [[TMP5]] to <8 x half>
-// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP7]] to <8 x half>
-// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP9]] to <8 x half>
-// CHECK:   [[VLD3_LANE:%.*]] = call { <8 x half>, <8 x half>, <8 x half> } @llvm.aarch64.neon.ld3lane.v8f16.p0i8(<8 x half> [[TMP10]], <8 x half> [[TMP11]], <8 x half> [[TMP12]], i64 7, i8* [[TMP3]])
-// CHECK:   [[TMP13:%.*]] = bitcast i8* [[TMP2]] to { <8 x half>, <8 x half>, <8 x half> }*
-// CHECK:   store { <8 x half>, <8 x half>, <8 x half> } [[VLD3_LANE]], { <8 x half>, <8 x half>, <8 x half> }* [[TMP13]]
+// CHECK:   [[TMP10:%.*]] = bitcast <16 x i8> [[TMP5]] to <8 x i16>
+// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP7]] to <8 x i16>
+// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP9]] to <8 x i16>
+// CHECK:   [[VLD3_LANE:%.*]] = call { <8 x i16>, <8 x i16>, <8 x i16> } @llvm.aarch64.neon.ld3lane.v8i16.p0i8(<8 x i16> [[TMP10]], <8 x i16> [[TMP11]], <8 x i16> [[TMP12]], i64 7, i8* [[TMP3]])
+// CHECK:   [[TMP13:%.*]] = bitcast i8* [[TMP2]] to { <8 x i16>, <8 x i16>, <8 x i16> }*
+// CHECK:   store { <8 x i16>, <8 x i16>, <8 x i16> } [[VLD3_LANE]], { <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP13]]
 // CHECK:   [[TMP14:%.*]] = bitcast %struct.float16x8x3_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP15:%.*]] = bitcast %struct.float16x8x3_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP14]], i8* [[TMP15]], i64 48, i32 16, i1 false)
@@ -3885,12 +3889,12 @@ int64x1x3_t test_vld3_lane_s64(int64_t  *a, int64x1x3_t b) {
 // CHECK:   [[ARRAYIDX4:%.*]] = getelementptr inbounds [3 x <4 x half>], [3 x <4 x half>]* [[VAL3]], i64 0, i64 2
 // CHECK:   [[TMP8:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX4]], align 8
 // CHECK:   [[TMP9:%.*]] = bitcast <4 x half> [[TMP8]] to <8 x i8>
-// CHECK:   [[TMP10:%.*]] = bitcast <8 x i8> [[TMP5]] to <4 x half>
-// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP7]] to <4 x half>
-// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP9]] to <4 x half>
-// CHECK:   [[VLD3_LANE:%.*]] = call { <4 x half>, <4 x half>, <4 x half> } @llvm.aarch64.neon.ld3lane.v4f16.p0i8(<4 x half> [[TMP10]], <4 x half> [[TMP11]], <4 x half> [[TMP12]], i64 3, i8* [[TMP3]])
-// CHECK:   [[TMP13:%.*]] = bitcast i8* [[TMP2]] to { <4 x half>, <4 x half>, <4 x half> }*
-// CHECK:   store { <4 x half>, <4 x half>, <4 x half> } [[VLD3_LANE]], { <4 x half>, <4 x half>, <4 x half> }* [[TMP13]]
+// CHECK:   [[TMP10:%.*]] = bitcast <8 x i8> [[TMP5]] to <4 x i16>
+// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP7]] to <4 x i16>
+// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP9]] to <4 x i16>
+// CHECK:   [[VLD3_LANE:%.*]] = call { <4 x i16>, <4 x i16>, <4 x i16> } @llvm.aarch64.neon.ld3lane.v4i16.p0i8(<4 x i16> [[TMP10]], <4 x i16> [[TMP11]], <4 x i16> [[TMP12]], i64 3, i8* [[TMP3]])
+// CHECK:   [[TMP13:%.*]] = bitcast i8* [[TMP2]] to { <4 x i16>, <4 x i16>, <4 x i16> }*
+// CHECK:   store { <4 x i16>, <4 x i16>, <4 x i16> } [[VLD3_LANE]], { <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP13]]
 // CHECK:   [[TMP14:%.*]] = bitcast %struct.float16x4x3_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP15:%.*]] = bitcast %struct.float16x4x3_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP14]], i8* [[TMP15]], i64 24, i32 8, i1 false)
@@ -4450,13 +4454,13 @@ int64x2x4_t test_vld4q_lane_s64(int64_t  *a, int64x2x4_t b) {
 // CHECK:   [[ARRAYIDX6:%.*]] = getelementptr inbounds [4 x <8 x half>], [4 x <8 x half>]* [[VAL5]], i64 0, i64 3
 // CHECK:   [[TMP10:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX6]], align 16
 // CHECK:   [[TMP11:%.*]] = bitcast <8 x half> [[TMP10]] to <16 x i8>
-// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP5]] to <8 x half>
-// CHECK:   [[TMP13:%.*]] = bitcast <16 x i8> [[TMP7]] to <8 x half>
-// CHECK:   [[TMP14:%.*]] = bitcast <16 x i8> [[TMP9]] to <8 x half>
-// CHECK:   [[TMP15:%.*]] = bitcast <16 x i8> [[TMP11]] to <8 x half>
-// CHECK:   [[VLD4_LANE:%.*]] = call { <8 x half>, <8 x half>, <8 x half>, <8 x half> } @llvm.aarch64.neon.ld4lane.v8f16.p0i8(<8 x half> [[TMP12]], <8 x half> [[TMP13]], <8 x half> [[TMP14]], <8 x half> [[TMP15]], i64 7, i8* [[TMP3]])
-// CHECK:   [[TMP16:%.*]] = bitcast i8* [[TMP2]] to { <8 x half>, <8 x half>, <8 x half>, <8 x half> }*
-// CHECK:   store { <8 x half>, <8 x half>, <8 x half>, <8 x half> } [[VLD4_LANE]], { <8 x half>, <8 x half>, <8 x half>, <8 x half> }* [[TMP16]]
+// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP5]] to <8 x i16>
+// CHECK:   [[TMP13:%.*]] = bitcast <16 x i8> [[TMP7]] to <8 x i16>
+// CHECK:   [[TMP14:%.*]] = bitcast <16 x i8> [[TMP9]] to <8 x i16>
+// CHECK:   [[TMP15:%.*]] = bitcast <16 x i8> [[TMP11]] to <8 x i16>
+// CHECK:   [[VLD4_LANE:%.*]] = call { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> } @llvm.aarch64.neon.ld4lane.v8i16.p0i8(<8 x i16> [[TMP12]], <8 x i16> [[TMP13]], <8 x i16> [[TMP14]], <8 x i16> [[TMP15]], i64 7, i8* [[TMP3]])
+// CHECK:   [[TMP16:%.*]] = bitcast i8* [[TMP2]] to { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> }*
+// CHECK:   store { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> } [[VLD4_LANE]], { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16> }* [[TMP16]]
 // CHECK:   [[TMP17:%.*]] = bitcast %struct.float16x8x4_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP18:%.*]] = bitcast %struct.float16x8x4_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP17]], i8* [[TMP18]], i64 64, i32 16, i1 false)
@@ -5039,13 +5043,13 @@ int64x1x4_t test_vld4_lane_s64(int64_t  *a, int64x1x4_t b) {
 // CHECK:   [[ARRAYIDX6:%.*]] = getelementptr inbounds [4 x <4 x half>], [4 x <4 x half>]* [[VAL5]], i64 0, i64 3
 // CHECK:   [[TMP10:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX6]], align 8
 // CHECK:   [[TMP11:%.*]] = bitcast <4 x half> [[TMP10]] to <8 x i8>
-// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP5]] to <4 x half>
-// CHECK:   [[TMP13:%.*]] = bitcast <8 x i8> [[TMP7]] to <4 x half>
-// CHECK:   [[TMP14:%.*]] = bitcast <8 x i8> [[TMP9]] to <4 x half>
-// CHECK:   [[TMP15:%.*]] = bitcast <8 x i8> [[TMP11]] to <4 x half>
-// CHECK:   [[VLD4_LANE:%.*]] = call { <4 x half>, <4 x half>, <4 x half>, <4 x half> } @llvm.aarch64.neon.ld4lane.v4f16.p0i8(<4 x half> [[TMP12]], <4 x half> [[TMP13]], <4 x half> [[TMP14]], <4 x half> [[TMP15]], i64 3, i8* [[TMP3]])
-// CHECK:   [[TMP16:%.*]] = bitcast i8* [[TMP2]] to { <4 x half>, <4 x half>, <4 x half>, <4 x half> }*
-// CHECK:   store { <4 x half>, <4 x half>, <4 x half>, <4 x half> } [[VLD4_LANE]], { <4 x half>, <4 x half>, <4 x half>, <4 x half> }* [[TMP16]]
+// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP5]] to <4 x i16>
+// CHECK:   [[TMP13:%.*]] = bitcast <8 x i8> [[TMP7]] to <4 x i16>
+// CHECK:   [[TMP14:%.*]] = bitcast <8 x i8> [[TMP9]] to <4 x i16>
+// CHECK:   [[TMP15:%.*]] = bitcast <8 x i8> [[TMP11]] to <4 x i16>
+// CHECK:   [[VLD4_LANE:%.*]] = call { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> } @llvm.aarch64.neon.ld4lane.v4i16.p0i8(<4 x i16> [[TMP12]], <4 x i16> [[TMP13]], <4 x i16> [[TMP14]], <4 x i16> [[TMP15]], i64 3, i8* [[TMP3]])
+// CHECK:   [[TMP16:%.*]] = bitcast i8* [[TMP2]] to { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> }*
+// CHECK:   store { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> } [[VLD4_LANE]], { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16> }* [[TMP16]]
 // CHECK:   [[TMP17:%.*]] = bitcast %struct.float16x4x4_t* [[RETVAL]] to i8*
 // CHECK:   [[TMP18:%.*]] = bitcast %struct.float16x4x4_t* [[__RET]] to i8*
 // CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP17]], i8* [[TMP18]], i64 32, i32 8, i1 false)
@@ -5357,10 +5361,10 @@ void test_vst1q_lane_s64(int64_t  *a, int64x2_t b) {
 // CHECK-LABEL: define void @test_vst1q_lane_f16(half* %a, <8 x half> %b) #0 {
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast <8 x half> %b to <16 x i8>
-// CHECK:   [[TMP2:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x half>
-// CHECK:   [[TMP3:%.*]] = extractelement <8 x half> [[TMP2]], i32 7
-// CHECK:   [[TMP4:%.*]] = bitcast i8* [[TMP0]] to half*
-// CHECK:   store half [[TMP3]], half* [[TMP4]]
+// CHECK:   [[TMP2:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x i16>
+// CHECK:   [[TMP3:%.*]] = extractelement <8 x i16> [[TMP2]], i32 7
+// CHECK:   [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i16*
+// CHECK:   store i16 [[TMP3]], i16* [[TMP4]]
 // CHECK:   ret void
 void test_vst1q_lane_f16(float16_t  *a, float16x8_t b) {
   vst1q_lane_f16(a, b, 7);
@@ -5513,10 +5517,10 @@ void test_vst1_lane_s64(int64_t  *a, int64x1_t b) {
 // CHECK-LABEL: define void @test_vst1_lane_f16(half* %a, <4 x half> %b) #0 {
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast <4 x half> %b to <8 x i8>
-// CHECK:   [[TMP2:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x half>
-// CHECK:   [[TMP3:%.*]] = extractelement <4 x half> [[TMP2]], i32 3
-// CHECK:   [[TMP4:%.*]] = bitcast i8* [[TMP0]] to half*
-// CHECK:   store half [[TMP3]], half* [[TMP4]]
+// CHECK:   [[TMP2:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x i16>
+// CHECK:   [[TMP3:%.*]] = extractelement <4 x i16> [[TMP2]], i32 3
+// CHECK:   [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i16*
+// CHECK:   store i16 [[TMP3]], i16* [[TMP4]]
 // CHECK:   ret void
 void test_vst1_lane_f16(float16_t  *a, float16x4_t b) {
   vst1_lane_f16(a, b, 3);
@@ -5785,9 +5789,9 @@ void test_vst2q_lane_s64(int64_t  *a, int64x2x2_t b) {
 // CHECK:   [[ARRAYIDX2:%.*]] = getelementptr inbounds [2 x <8 x half>], [2 x <8 x half>]* [[VAL1]], i64 0, i64 1
 // CHECK:   [[TMP5:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX2]], align 16
 // CHECK:   [[TMP6:%.*]] = bitcast <8 x half> [[TMP5]] to <16 x i8>
-// CHECK:   [[TMP7:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x half>
-// CHECK:   [[TMP8:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x half>
-// CHECK:   call void @llvm.aarch64.neon.st2lane.v8f16.p0i8(<8 x half> [[TMP7]], <8 x half> [[TMP8]], i64 7, i8* [[TMP2]])
+// CHECK:   [[TMP7:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
+// CHECK:   [[TMP8:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x i16>
+// CHECK:   call void @llvm.aarch64.neon.st2lane.v8i16.p0i8(<8 x i16> [[TMP7]], <8 x i16> [[TMP8]], i64 7, i8* [[TMP2]])
 // CHECK:   ret void
 void test_vst2q_lane_f16(float16_t  *a, float16x8x2_t b) {
   vst2q_lane_f16(a, b, 7);
@@ -6120,9 +6124,9 @@ void test_vst2_lane_s64(int64_t  *a, int64x1x2_t b) {
 // CHECK:   [[ARRAYIDX2:%.*]] = getelementptr inbounds [2 x <4 x half>], [2 x <4 x half>]* [[VAL1]], i64 0, i64 1
 // CHECK:   [[TMP5:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX2]], align 8
 // CHECK:   [[TMP6:%.*]] = bitcast <4 x half> [[TMP5]] to <8 x i8>
-// CHECK:   [[TMP7:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x half>
-// CHECK:   [[TMP8:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x half>
-// CHECK:   call void @llvm.aarch64.neon.st2lane.v4f16.p0i8(<4 x half> [[TMP7]], <4 x half> [[TMP8]], i64 3, i8* [[TMP2]])
+// CHECK:   [[TMP7:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
+// CHECK:   [[TMP8:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x i16>
+// CHECK:   call void @llvm.aarch64.neon.st2lane.v4i16.p0i8(<4 x i16> [[TMP7]], <4 x i16> [[TMP8]], i64 3, i8* [[TMP2]])
 // CHECK:   ret void
 void test_vst2_lane_f16(float16_t  *a, float16x4x2_t b) {
   vst2_lane_f16(a, b, 3);
@@ -6495,10 +6499,10 @@ void test_vst3q_lane_s64(int64_t  *a, int64x2x3_t b) {
 // CHECK:   [[ARRAYIDX4:%.*]] = getelementptr inbounds [3 x <8 x half>], [3 x <8 x half>]* [[VAL3]], i64 0, i64 2
 // CHECK:   [[TMP7:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX4]], align 16
 // CHECK:   [[TMP8:%.*]] = bitcast <8 x half> [[TMP7]] to <16 x i8>
-// CHECK:   [[TMP9:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x half>
-// CHECK:   [[TMP10:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x half>
-// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP8]] to <8 x half>
-// CHECK:   call void @llvm.aarch64.neon.st3lane.v8f16.p0i8(<8 x half> [[TMP9]], <8 x half> [[TMP10]], <8 x half> [[TMP11]], i64 7, i8* [[TMP2]])
+// CHECK:   [[TMP9:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
+// CHECK:   [[TMP10:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x i16>
+// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP8]] to <8 x i16>
+// CHECK:   call void @llvm.aarch64.neon.st3lane.v8i16.p0i8(<8 x i16> [[TMP9]], <8 x i16> [[TMP10]], <8 x i16> [[TMP11]], i64 7, i8* [[TMP2]])
 // CHECK:   ret void
 void test_vst3q_lane_f16(float16_t  *a, float16x8x3_t b) {
   vst3q_lane_f16(a, b, 7);
@@ -6894,10 +6898,10 @@ void test_vst3_lane_s64(int64_t  *a, int64x1x3_t b) {
 // CHECK:   [[ARRAYIDX4:%.*]] = getelementptr inbounds [3 x <4 x half>], [3 x <4 x half>]* [[VAL3]], i64 0, i64 2
 // CHECK:   [[TMP7:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX4]], align 8
 // CHECK:   [[TMP8:%.*]] = bitcast <4 x half> [[TMP7]] to <8 x i8>
-// CHECK:   [[TMP9:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x half>
-// CHECK:   [[TMP10:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x half>
-// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP8]] to <4 x half>
-// CHECK:   call void @llvm.aarch64.neon.st3lane.v4f16.p0i8(<4 x half> [[TMP9]], <4 x half> [[TMP10]], <4 x half> [[TMP11]], i64 3, i8* [[TMP2]])
+// CHECK:   [[TMP9:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
+// CHECK:   [[TMP10:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x i16>
+// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP8]] to <4 x i16>
+// CHECK:   call void @llvm.aarch64.neon.st3lane.v4i16.p0i8(<4 x i16> [[TMP9]], <4 x i16> [[TMP10]], <4 x i16> [[TMP11]], i64 3, i8* [[TMP2]])
 // CHECK:   ret void
 void test_vst3_lane_f16(float16_t  *a, float16x4x3_t b) {
   vst3_lane_f16(a, b, 3);
@@ -7333,11 +7337,11 @@ void test_vst4q_lane_s64(int64_t  *a, int64x2x4_t b) {
 // CHECK:   [[ARRAYIDX6:%.*]] = getelementptr inbounds [4 x <8 x half>], [4 x <8 x half>]* [[VAL5]], i64 0, i64 3
 // CHECK:   [[TMP9:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX6]], align 16
 // CHECK:   [[TMP10:%.*]] = bitcast <8 x half> [[TMP9]] to <16 x i8>
-// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x half>
-// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x half>
-// CHECK:   [[TMP13:%.*]] = bitcast <16 x i8> [[TMP8]] to <8 x half>
-// CHECK:   [[TMP14:%.*]] = bitcast <16 x i8> [[TMP10]] to <8 x half>
-// CHECK:   call void @llvm.aarch64.neon.st4lane.v8f16.p0i8(<8 x half> [[TMP11]], <8 x half> [[TMP12]], <8 x half> [[TMP13]], <8 x half> [[TMP14]], i64 7, i8* [[TMP2]])
+// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
+// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x i16>
+// CHECK:   [[TMP13:%.*]] = bitcast <16 x i8> [[TMP8]] to <8 x i16>
+// CHECK:   [[TMP14:%.*]] = bitcast <16 x i8> [[TMP10]] to <8 x i16>
+// CHECK:   call void @llvm.aarch64.neon.st4lane.v8i16.p0i8(<8 x i16> [[TMP11]], <8 x i16> [[TMP12]], <8 x i16> [[TMP13]], <8 x i16> [[TMP14]], i64 7, i8* [[TMP2]])
 // CHECK:   ret void
 void test_vst4q_lane_f16(float16_t  *a, float16x8x4_t b) {
   vst4q_lane_f16(a, b, 7);
@@ -7796,11 +7800,11 @@ void test_vst4_lane_s64(int64_t  *a, int64x1x4_t b) {
 // CHECK:   [[ARRAYIDX6:%.*]] = getelementptr inbounds [4 x <4 x half>], [4 x <4 x half>]* [[VAL5]], i64 0, i64 3
 // CHECK:   [[TMP9:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX6]], align 8
 // CHECK:   [[TMP10:%.*]] = bitcast <4 x half> [[TMP9]] to <8 x i8>
-// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x half>
-// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x half>
-// CHECK:   [[TMP13:%.*]] = bitcast <8 x i8> [[TMP8]] to <4 x half>
-// CHECK:   [[TMP14:%.*]] = bitcast <8 x i8> [[TMP10]] to <4 x half>
-// CHECK:   call void @llvm.aarch64.neon.st4lane.v4f16.p0i8(<4 x half> [[TMP11]], <4 x half> [[TMP12]], <4 x half> [[TMP13]], <4 x half> [[TMP14]], i64 3, i8* [[TMP2]])
+// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
+// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x i16>
+// CHECK:   [[TMP13:%.*]] = bitcast <8 x i8> [[TMP8]] to <4 x i16>
+// CHECK:   [[TMP14:%.*]] = bitcast <8 x i8> [[TMP10]] to <4 x i16>
+// CHECK:   call void @llvm.aarch64.neon.st4lane.v4i16.p0i8(<4 x i16> [[TMP11]], <4 x i16> [[TMP12]], <4 x i16> [[TMP13]], <4 x i16> [[TMP14]], i64 3, i8* [[TMP2]])
 // CHECK:   ret void
 void test_vst4_lane_f16(float16_t  *a, float16x4x4_t b) {
   vst4_lane_f16(a, b, 3);
diff --git a/test/CodeGen/aarch64-v8.2a-neon-intrinsics.c b/test/CodeGen/aarch64-v8.2a-neon-intrinsics.c
deleted file mode 100644
index 3f61238b64fb..000000000000
--- a/test/CodeGen/aarch64-v8.2a-neon-intrinsics.c
+++ /dev/null
@@ -1,1633 +0,0 @@
-// RUN: %clang_cc1 -triple arm64-none-linux-gnu -target-feature +neon -target-feature +fullfp16 -target-feature +v8.2a\
-// RUN: -fallow-half-arguments-and-returns -S -disable-O0-optnone -emit-llvm -o - %s \
-// RUN: | opt -S -mem2reg \
-// RUN: | FileCheck %s
-
-// REQUIRES: aarch64-registered-target
-
-#include <arm_neon.h>
-
-// CHECK-LABEL: test_vabs_f16
-// CHECK:  [[ABS:%.*]] =  call <4 x half> @llvm.fabs.v4f16(<4 x half> %a)
-// CHECK:  ret <4 x half> [[ABS]]
-float16x4_t test_vabs_f16(float16x4_t a) {
-  return vabs_f16(a);
-}
-
-// CHECK-LABEL: test_vabsq_f16
-// CHECK:  [[ABS:%.*]] = call <8 x half> @llvm.fabs.v8f16(<8 x half> %a)
-// CHECK:  ret <8 x half> [[ABS]]
-float16x8_t test_vabsq_f16(float16x8_t a) {
-  return vabsq_f16(a);
-}
-
-// CHECK-LABEL: test_vceqz_f16
-// CHECK:  [[TMP1:%.*]] = fcmp oeq <4 x half> %a, zeroinitializer
-// CHECK:  [[TMP2:%.*]] = sext <4 x i1> [[TMP1]] to <4 x i16>
-// CHECK:  ret <4 x i16> [[TMP2]]
-uint16x4_t test_vceqz_f16(float16x4_t a) {
-  return vceqz_f16(a);
-}
-
-// CHECK-LABEL: test_vceqzq_f16
-// CHECK:  [[TMP1:%.*]] = fcmp oeq <8 x half> %a, zeroinitializer
-// CHECK:  [[TMP2:%.*]] = sext <8 x i1> [[TMP1:%.*]] to <8 x i16>
-// CHECK:  ret <8 x i16> [[TMP2]]
-uint16x8_t test_vceqzq_f16(float16x8_t a) {
-  return vceqzq_f16(a);
-}
-
-// CHECK-LABEL: test_vcgez_f16
-// CHECK:  [[TMP1:%.*]] = fcmp oge <4 x half> %a, zeroinitializer
-// CHECK:  [[TMP2:%.*]] = sext <4 x i1> [[TMP1]] to <4 x i16>
-// CHECK:  ret <4 x i16> [[TMP2]]
-uint16x4_t test_vcgez_f16(float16x4_t a) {
-  return vcgez_f16(a);
-}
-
-// CHECK-LABEL: test_vcgezq_f16
-// CHECK:  [[TMP1:%.*]] = fcmp oge <8 x half> %a, zeroinitializer
-// CHECK:  [[TMP2:%.*]] = sext <8 x i1> [[TMP1:%.*]] to <8 x i16>
-// CHECK:  ret <8 x i16> [[TMP2]]
-uint16x8_t test_vcgezq_f16(float16x8_t a) {
-  return vcgezq_f16(a);
-}
-
-// CHECK-LABEL: test_vcgtz_f16
-// CHECK:  [[TMP1:%.*]] = fcmp ogt <4 x half> %a, zeroinitializer
-// CHECK:  [[TMP2:%.*]] = sext <4 x i1> [[TMP1]] to <4 x i16>
-// CHECK:  ret <4 x i16> [[TMP2]]
-uint16x4_t test_vcgtz_f16(float16x4_t a) {
-  return vcgtz_f16(a);
-}
-
-// CHECK-LABEL: test_vcgtzq_f16
-// CHECK:  [[TMP1:%.*]] = fcmp ogt <8 x half> %a, zeroinitializer
-// CHECK:  [[TMP2:%.*]] = sext <8 x i1> [[TMP1:%.*]] to <8 x i16>
-// CHECK:  ret <8 x i16> [[TMP2]]
-uint16x8_t test_vcgtzq_f16(float16x8_t a) {
-  return vcgtzq_f16(a);
-}
-
-// CHECK-LABEL: test_vclez_f16
-// CHECK:  [[TMP1:%.*]] = fcmp ole <4 x half> %a, zeroinitializer
-// CHECK:  [[TMP2:%.*]] = sext <4 x i1> [[TMP1]] to <4 x i16>
-// CHECK:  ret <4 x i16> [[TMP2]]
-uint16x4_t test_vclez_f16(float16x4_t a) {
-  return vclez_f16(a);
-}
-
-// CHECK-LABEL: test_vclezq_f16
-// CHECK:  [[TMP1:%.*]] = fcmp ole <8 x half> %a, zeroinitializer
-// CHECK:  [[TMP2:%.*]] = sext <8 x i1> [[TMP1:%.*]] to <8 x i16>
-// CHECK:  ret <8 x i16> [[TMP2]]
-uint16x8_t test_vclezq_f16(float16x8_t a) {
-  return vclezq_f16(a);
-}
-
-// CHECK-LABEL: test_vcltz_f16
-// CHECK:  [[TMP1:%.*]] = fcmp olt <4 x half> %a, zeroinitializer
-// CHECK:  [[TMP2:%.*]] = sext <4 x i1> [[TMP1]] to <4 x i16>
-// CHECK:  ret <4 x i16> [[TMP2]]
-uint16x4_t test_vcltz_f16(float16x4_t a) {
-  return vcltz_f16(a);
-}
-
-// CHECK-LABEL: test_vcltzq_f16
-// CHECK:  [[TMP1:%.*]] = fcmp olt <8 x half> %a, zeroinitializer
-// CHECK:  [[TMP2:%.*]] = sext <8 x i1> [[TMP1:%.*]] to <8 x i16>
-// CHECK:  ret <8 x i16> [[TMP2]]
-uint16x8_t test_vcltzq_f16(float16x8_t a) {
-  return vcltzq_f16(a);
-}
-
-// CHECK-LABEL: test_vcvt_f16_s16
-// CHECK:  [[VCVT:%.*]] = sitofp <4 x i16> %a to <4 x half>
-// CHECK:  ret <4 x half> [[VCVT]]
-float16x4_t test_vcvt_f16_s16 (int16x4_t a) {
-  return vcvt_f16_s16(a);
-}
-
-// CHECK-LABEL: test_vcvtq_f16_s16
-// CHECK:  [[VCVT:%.*]] = sitofp <8 x i16> %a to <8 x half>
-// CHECK:  ret <8 x half> [[VCVT]]
-float16x8_t test_vcvtq_f16_s16 (int16x8_t a) {
-  return vcvtq_f16_s16(a);
-}
-
-// CHECK-LABEL: test_vcvt_f16_u16
-// CHECK:  [[VCVT:%.*]] = uitofp <4 x i16> %a to <4 x half>
-// CHECK:  ret <4 x half> [[VCVT]]
-float16x4_t test_vcvt_f16_u16 (uint16x4_t a) {
-  return vcvt_f16_u16(a);
-}
-
-// CHECK-LABEL: test_vcvtq_f16_u16
-// CHECK:  [[VCVT:%.*]] = uitofp <8 x i16> %a to <8 x half>
-// CHECK:  ret <8 x half> [[VCVT]]
-float16x8_t test_vcvtq_f16_u16 (uint16x8_t a) {
-  return vcvtq_f16_u16(a);
-}
-
-// CHECK-LABEL: test_vcvt_s16_f16
-// CHECK:  [[VCVT:%.*]] = fptosi <4 x half> %a to <4 x i16>
-// CHECK:  ret <4 x i16> [[VCVT]]
-int16x4_t test_vcvt_s16_f16 (float16x4_t a) {
-  return vcvt_s16_f16(a);
-}
-
-// CHECK-LABEL: test_vcvtq_s16_f16
-// CHECK:  [[VCVT:%.*]] = fptosi <8 x half> %a to <8 x i16>
-// CHECK:  ret <8 x i16> [[VCVT]]
-int16x8_t test_vcvtq_s16_f16 (float16x8_t a) {
-  return vcvtq_s16_f16(a);
-}
-
-// CHECK-LABEL: test_vcvt_u16_f16
-// CHECK:  [[VCVT:%.*]] = fptoui <4 x half> %a to <4 x i16>
-// CHECK:  ret <4 x i16> [[VCVT]]
-int16x4_t test_vcvt_u16_f16 (float16x4_t a) {
-  return vcvt_u16_f16(a);
-}
-
-// CHECK-LABEL: test_vcvtq_u16_f16
-// CHECK:  [[VCVT:%.*]] = fptoui <8 x half> %a to <8 x i16>
-// CHECK:  ret <8 x i16> [[VCVT]]
-int16x8_t test_vcvtq_u16_f16 (float16x8_t a) {
-  return vcvtq_u16_f16(a);
-}
-
-// CHECK-LABEL: test_vcvta_s16_f16
-// CHECK:  [[VCVT:%.*]] = call <4 x i16> @llvm.aarch64.neon.fcvtas.v4i16.v4f16(<4 x half> %a)
-// CHECK:  ret <4 x i16> [[VCVT]]
-int16x4_t test_vcvta_s16_f16 (float16x4_t a) {
-  return vcvta_s16_f16(a);
-}
-
-// CHECK-LABEL: test_vcvtaq_s16_f16
-// CHECK:  [[VCVT:%.*]] = call <8 x i16> @llvm.aarch64.neon.fcvtas.v8i16.v8f16(<8 x half> %a)
-// CHECK:  ret <8 x i16> [[VCVT]]
-int16x8_t test_vcvtaq_s16_f16 (float16x8_t a) {
-  return vcvtaq_s16_f16(a);
-}
-
-// CHECK-LABEL: test_vcvtm_s16_f16
-// CHECK:  [[VCVT:%.*]] = call <4 x i16> @llvm.aarch64.neon.fcvtms.v4i16.v4f16(<4 x half> %a)
-// CHECK:  ret <4 x i16> [[VCVT]]
-int16x4_t test_vcvtm_s16_f16 (float16x4_t a) {
-  return vcvtm_s16_f16(a);
-}
-
-// CHECK-LABEL: test_vcvtmq_s16_f16
-// CHECK:  [[VCVT:%.*]] = call <8 x i16> @llvm.aarch64.neon.fcvtms.v8i16.v8f16(<8 x half> %a)
-// CHECK:  ret <8 x i16> [[VCVT]]
-int16x8_t test_vcvtmq_s16_f16 (float16x8_t a) {
-  return vcvtmq_s16_f16(a);
-}
-
-// CHECK-LABEL: test_vcvtm_u16_f16
-// CHECK:  [[VCVT:%.*]] = call <4 x i16> @llvm.aarch64.neon.fcvtmu.v4i16.v4f16(<4 x half> %a)
-// CHECK:  ret <4 x i16> [[VCVT]]
-uint16x4_t test_vcvtm_u16_f16 (float16x4_t a) {
-  return vcvtm_u16_f16(a);
-}
-
-// CHECK-LABEL: test_vcvtmq_u16_f16
-// CHECK:  [[VCVT:%.*]] = call <8 x i16> @llvm.aarch64.neon.fcvtmu.v8i16.v8f16(<8 x half> %a)
-// CHECK:  ret <8 x i16> [[VCVT]]
-uint16x8_t test_vcvtmq_u16_f16 (float16x8_t a) {
-  return vcvtmq_u16_f16(a);
-}
-
-// CHECK-LABEL: test_vcvtn_s16_f16
-// CHECK:  [[VCVT:%.*]] = call <4 x i16> @llvm.aarch64.neon.fcvtns.v4i16.v4f16(<4 x half> %a)
-// CHECK:  ret <4 x i16> [[VCVT]]
-int16x4_t test_vcvtn_s16_f16 (float16x4_t a) {
-  return vcvtn_s16_f16(a);
-}
-
-// CHECK-LABEL: test_vcvtnq_s16_f16
-// CHECK:  [[VCVT:%.*]] = call <8 x i16> @llvm.aarch64.neon.fcvtns.v8i16.v8f16(<8 x half> %a)
-// CHECK:  ret <8 x i16> [[VCVT]]
-int16x8_t test_vcvtnq_s16_f16 (float16x8_t a) {
-  return vcvtnq_s16_f16(a);
-}
-
-// CHECK-LABEL: test_vcvtn_u16_f16
-// CHECK:  [[VCVT:%.*]] = call <4 x i16> @llvm.aarch64.neon.fcvtnu.v4i16.v4f16(<4 x half> %a)
-// CHECK:  ret <4 x i16> [[VCVT]]
-uint16x4_t test_vcvtn_u16_f16 (float16x4_t a) {
-  return vcvtn_u16_f16(a);
-}
-
-// CHECK-LABEL: test_vcvtnq_u16_f16
-// CHECK:  [[VCVT:%.*]] = call <8 x i16> @llvm.aarch64.neon.fcvtnu.v8i16.v8f16(<8 x half> %a)
-// CHECK:  ret <8 x i16> [[VCVT]]
-uint16x8_t test_vcvtnq_u16_f16 (float16x8_t a) {
-  return vcvtnq_u16_f16(a);
-}
-
-// CHECK-LABEL: test_vcvtp_s16_f16
-// CHECK:  [[VCVT:%.*]] = call <4 x i16> @llvm.aarch64.neon.fcvtps.v4i16.v4f16(<4 x half> %a)
-// CHECK:  ret <4 x i16> [[VCVT]]
-int16x4_t test_vcvtp_s16_f16 (float16x4_t a) {
-  return vcvtp_s16_f16(a);
-}
-
-// CHECK-LABEL: test_vcvtpq_s16_f16
-// CHECK:  [[VCVT:%.*]] = call <8 x i16> @llvm.aarch64.neon.fcvtps.v8i16.v8f16(<8 x half> %a)
-// CHECK:  ret <8 x i16> [[VCVT]]
-int16x8_t test_vcvtpq_s16_f16 (float16x8_t a) {
-  return vcvtpq_s16_f16(a);
-}
-
-// CHECK-LABEL: test_vcvtp_u16_f16
-// CHECK:  [[VCVT:%.*]] = call <4 x i16> @llvm.aarch64.neon.fcvtpu.v4i16.v4f16(<4 x half> %a)
-// CHECK:  ret <4 x i16> [[VCVT]]
-uint16x4_t test_vcvtp_u16_f16 (float16x4_t a) {
-  return vcvtp_u16_f16(a);
-}
-
-// CHECK-LABEL: test_vcvtpq_u16_f16
-// CHECK:  [[VCVT:%.*]] = call <8 x i16> @llvm.aarch64.neon.fcvtpu.v8i16.v8f16(<8 x half> %a)
-// CHECK:  ret <8 x i16> [[VCVT]]
-uint16x8_t test_vcvtpq_u16_f16 (float16x8_t a) {
-  return vcvtpq_u16_f16(a);
-}
-
-// FIXME: Fix the zero constant when fp16 non-storage-only type becomes available.
-// CHECK-LABEL: test_vneg_f16
-// CHECK:  [[NEG:%.*]] = fsub <4 x half> <half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000>, %a
-// CHECK:  ret <4 x half> [[NEG]]
-float16x4_t test_vneg_f16(float16x4_t a) {
-  return vneg_f16(a);
-}
-
-// CHECK-LABEL: test_vnegq_f16
-// CHECK:  [[NEG:%.*]] = fsub <8 x half> <half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000>, %a
-// CHECK:  ret <8 x half> [[NEG]]
-float16x8_t test_vnegq_f16(float16x8_t a) {
-  return vnegq_f16(a);
-}
-
-// CHECK-LABEL: test_vrecpe_f16
-// CHECK:  [[RCP:%.*]] = call <4 x half> @llvm.aarch64.neon.frecpe.v4f16(<4 x half> %a)
-// CHECK:  ret <4 x half> [[RCP]]
-float16x4_t test_vrecpe_f16(float16x4_t a) {
-  return vrecpe_f16(a);
-}
-
-// CHECK-LABEL: test_vrecpeq_f16
-// CHECK:  [[RCP:%.*]] = call <8 x half> @llvm.aarch64.neon.frecpe.v8f16(<8 x half> %a)
-// CHECK:  ret <8 x half> [[RCP]]
-float16x8_t test_vrecpeq_f16(float16x8_t a) {
-  return vrecpeq_f16(a);
-}
-
-// CHECK-LABEL: test_vrnd_f16
-// CHECK:  [[RND:%.*]] =  call <4 x half> @llvm.trunc.v4f16(<4 x half> %a)
-// CHECK:  ret <4 x half> [[RND]]
-float16x4_t test_vrnd_f16(float16x4_t a) {
-  return vrnd_f16(a);
-}
-
-// CHECK-LABEL: test_vrndq_f16
-// CHECK:  [[RND:%.*]] =  call <8 x half> @llvm.trunc.v8f16(<8 x half> %a)
-// CHECK:  ret <8 x half> [[RND]]
-float16x8_t test_vrndq_f16(float16x8_t a) {
-  return vrndq_f16(a);
-}
-
-// CHECK-LABEL: test_vrnda_f16
-// CHECK:  [[RND:%.*]] =  call <4 x half> @llvm.round.v4f16(<4 x half> %a)
-// CHECK:  ret <4 x half> [[RND]]
-float16x4_t test_vrnda_f16(float16x4_t a) {
-  return vrnda_f16(a);
-}
-
-// CHECK-LABEL: test_vrndaq_f16
-// CHECK:  [[RND:%.*]] =  call <8 x half> @llvm.round.v8f16(<8 x half> %a)
-// CHECK:  ret <8 x half> [[RND]]
-float16x8_t test_vrndaq_f16(float16x8_t a) {
-  return vrndaq_f16(a);
-}
-
-// CHECK-LABEL: test_vrndi_f16
-// CHECK:  [[RND:%.*]] =  call <4 x half> @llvm.nearbyint.v4f16(<4 x half> %a)
-// CHECK:  ret <4 x half> [[RND]]
-float16x4_t test_vrndi_f16(float16x4_t a) {
-  return vrndi_f16(a);
-}
-
-// CHECK-LABEL: test_vrndiq_f16
-// CHECK:  [[RND:%.*]] =  call <8 x half> @llvm.nearbyint.v8f16(<8 x half> %a)
-// CHECK:  ret <8 x half> [[RND]]
-float16x8_t test_vrndiq_f16(float16x8_t a) {
-  return vrndiq_f16(a);
-}
-
-// CHECK-LABEL: test_vrndm_f16
-// CHECK:  [[RND:%.*]] =  call <4 x half> @llvm.floor.v4f16(<4 x half> %a)
-// CHECK:  ret <4 x half> [[RND]]
-float16x4_t test_vrndm_f16(float16x4_t a) {
-  return vrndm_f16(a);
-}
-
-// CHECK-LABEL: test_vrndmq_f16
-// CHECK:  [[RND:%.*]] =  call <8 x half> @llvm.floor.v8f16(<8 x half> %a)
-// CHECK:  ret <8 x half> [[RND]]
-float16x8_t test_vrndmq_f16(float16x8_t a) {
-  return vrndmq_f16(a);
-}
-
-// CHECK-LABEL: test_vrndn_f16
-// CHECK:  [[RND:%.*]] =  call <4 x half> @llvm.aarch64.neon.frintn.v4f16(<4 x half> %a)
-// CHECK:  ret <4 x half> [[RND]]
-float16x4_t test_vrndn_f16(float16x4_t a) {
-  return vrndn_f16(a);
-}
-
-// CHECK-LABEL: test_vrndnq_f16
-// CHECK:  [[RND:%.*]] =  call <8 x half> @llvm.aarch64.neon.frintn.v8f16(<8 x half> %a)
-// CHECK:  ret <8 x half> [[RND]]
-float16x8_t test_vrndnq_f16(float16x8_t a) {
-  return vrndnq_f16(a);
-}
-
-// CHECK-LABEL: test_vrndp_f16
-// CHECK:  [[RND:%.*]] =  call <4 x half> @llvm.ceil.v4f16(<4 x half> %a)
-// CHECK:  ret <4 x half> [[RND]]
-float16x4_t test_vrndp_f16(float16x4_t a) {
-  return vrndp_f16(a);
-}
-
-// CHECK-LABEL: test_vrndpq_f16
-// CHECK:  [[RND:%.*]] =  call <8 x half> @llvm.ceil.v8f16(<8 x half> %a)
-// CHECK:  ret <8 x half> [[RND]]
-float16x8_t test_vrndpq_f16(float16x8_t a) {
-  return vrndpq_f16(a);
-}
-
-// CHECK-LABEL: test_vrndx_f16
-// CHECK:  [[RND:%.*]] =  call <4 x half> @llvm.rint.v4f16(<4 x half> %a)
-// CHECK:  ret <4 x half> [[RND]]
-float16x4_t test_vrndx_f16(float16x4_t a) {
-  return vrndx_f16(a);
-}
-
-// CHECK-LABEL: test_vrndxq_f16
-// CHECK:  [[RND:%.*]] =  call <8 x half> @llvm.rint.v8f16(<8 x half> %a)
-// CHECK:  ret <8 x half> [[RND]]
-float16x8_t test_vrndxq_f16(float16x8_t a) {
-  return vrndxq_f16(a);
-}
-
-// CHECK-LABEL: test_vrsqrte_f16
-// CHECK:  [[RND:%.*]] = call <4 x half> @llvm.aarch64.neon.frsqrte.v4f16(<4 x half> %a)
-// CHECK:  ret <4 x half> [[RND]]
-float16x4_t test_vrsqrte_f16(float16x4_t a) {
-  return vrsqrte_f16(a);
-}
-
-// CHECK-LABEL: test_vrsqrteq_f16
-// CHECK:  [[RND:%.*]] = call <8 x half> @llvm.aarch64.neon.frsqrte.v8f16(<8 x half> %a)
-// CHECK:  ret <8 x half> [[RND]]
-float16x8_t test_vrsqrteq_f16(float16x8_t a) {
-  return vrsqrteq_f16(a);
-}
-
-// CHECK-LABEL: test_vsqrt_f16
-// CHECK:  [[SQR:%.*]] = call <4 x half> @llvm.sqrt.v4f16(<4 x half> %a)
-// CHECK:  ret <4 x half> [[SQR]]
-float16x4_t test_vsqrt_f16(float16x4_t a) {
-  return vsqrt_f16(a);
-}
-
-// CHECK-LABEL: test_vsqrtq_f16
-// CHECK:  [[SQR:%.*]] = call <8 x half> @llvm.sqrt.v8f16(<8 x half> %a)
-// CHECK:  ret <8 x half> [[SQR]]
-float16x8_t test_vsqrtq_f16(float16x8_t a) {
-  return vsqrtq_f16(a);
-}
-
-// CHECK-LABEL: test_vadd_f16
-// CHECK:  [[ADD:%.*]] = fadd <4 x half> %a, %b
-// CHECK:  ret <4 x half> [[ADD]]
-float16x4_t test_vadd_f16(float16x4_t a, float16x4_t b) {
-  return vadd_f16(a, b);
-}
-
-// CHECK-LABEL: test_vaddq_f16
-// CHECK:  [[ADD:%.*]] = fadd <8 x half> %a, %b
-// CHECK:  ret <8 x half> [[ADD]]
-float16x8_t test_vaddq_f16(float16x8_t a, float16x8_t b) {
-  return vaddq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vabd_f16
-// CHECK:  [[ABD:%.*]] = call <4 x half> @llvm.aarch64.neon.fabd.v4f16(<4 x half> %a, <4 x half> %b)
-// CHECK:  ret <4 x half> [[ABD]]
-float16x4_t test_vabd_f16(float16x4_t a, float16x4_t b) {
-  return vabd_f16(a, b);
-}
-
-// CHECK-LABEL: test_vabdq_f16
-// CHECK:  [[ABD:%.*]] = call <8 x half> @llvm.aarch64.neon.fabd.v8f16(<8 x half> %a, <8 x half> %b)
-// CHECK:  ret <8 x half> [[ABD]]
-float16x8_t test_vabdq_f16(float16x8_t a, float16x8_t b) {
-  return vabdq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vcage_f16
-// CHECK:  [[ABS:%.*]] = call <4 x i16> @llvm.aarch64.neon.facge.v4i16.v4f16(<4 x half> %a, <4 x half> %b)
-// CHECK:  ret <4 x i16> [[ABS]]
-uint16x4_t test_vcage_f16(float16x4_t a, float16x4_t b) {
-  return vcage_f16(a, b);
-}
-
-// CHECK-LABEL: test_vcageq_f16
-// CHECK:  [[ABS:%.*]] = call <8 x i16> @llvm.aarch64.neon.facge.v8i16.v8f16(<8 x half> %a, <8 x half> %b)
-// CHECK:  ret <8 x i16> [[ABS]]
-uint16x8_t test_vcageq_f16(float16x8_t a, float16x8_t b) {
-  return vcageq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vcagt_f16
-// CHECK:  [[ABS:%.*]] = call <4 x i16> @llvm.aarch64.neon.facgt.v4i16.v4f16(<4 x half> %a, <4 x half> %b)
-// CHECK:  ret <4 x i16> [[ABS]]
-uint16x4_t test_vcagt_f16(float16x4_t a, float16x4_t b) {
-  return vcagt_f16(a, b);
-}
-
-// CHECK-LABEL: test_vcagtq_f16
-// CHECK:  [[ABS:%.*]] = call <8 x i16> @llvm.aarch64.neon.facgt.v8i16.v8f16(<8 x half> %a, <8 x half> %b)
-// CHECK:  ret <8 x i16> [[ABS]]
-uint16x8_t test_vcagtq_f16(float16x8_t a, float16x8_t b) {
-  return vcagtq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vcale_f16
-// CHECK:  [[ABS:%.*]] = call <4 x i16> @llvm.aarch64.neon.facge.v4i16.v4f16(<4 x half> %b, <4 x half> %a)
-// CHECK:  ret <4 x i16> [[ABS]]
-uint16x4_t test_vcale_f16(float16x4_t a, float16x4_t b) {
-  return vcale_f16(a, b);
-}
-
-// CHECK-LABEL: test_vcaleq_f16
-// CHECK:  [[ABS:%.*]] = call <8 x i16> @llvm.aarch64.neon.facge.v8i16.v8f16(<8 x half> %b, <8 x half> %a)
-// CHECK:  ret <8 x i16> [[ABS]]
-uint16x8_t test_vcaleq_f16(float16x8_t a, float16x8_t b) {
-  return vcaleq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vcalt_f16
-// CHECK:  [[ABS:%.*]] = call <4 x i16> @llvm.aarch64.neon.facgt.v4i16.v4f16(<4 x half> %b, <4 x half> %a)
-// CHECK:  ret <4 x i16> [[ABS]]
-uint16x4_t test_vcalt_f16(float16x4_t a, float16x4_t b) {
-  return vcalt_f16(a, b);
-}
-
-// CHECK-LABEL: test_vcaltq_f16
-// CHECK:  [[ABS:%.*]] = call <8 x i16> @llvm.aarch64.neon.facgt.v8i16.v8f16(<8 x half> %b, <8 x half> %a)
-// CHECK:  ret <8 x i16> [[ABS]]
-uint16x8_t test_vcaltq_f16(float16x8_t a, float16x8_t b) {
-  return vcaltq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vceq_f16
-// CHECK:  [[TMP1:%.*]] = fcmp oeq <4 x half> %a, %b
-// CHECK:  [[TMP2:%.*]] = sext <4 x i1> [[TMP1]] to <4 x i16>
-// CHECK:  ret <4 x i16> [[TMP2]]
-uint16x4_t test_vceq_f16(float16x4_t a, float16x4_t b) {
-  return vceq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vceqq_f16
-// CHECK:  [[TMP1:%.*]] = fcmp oeq <8 x half> %a, %b
-// CHECK:  [[TMP2:%.*]] = sext <8 x i1> [[TMP1:%.*]] to <8 x i16>
-// CHECK:  ret <8 x i16> [[TMP2]]
-uint16x8_t test_vceqq_f16(float16x8_t a, float16x8_t b) {
-  return vceqq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vcge_f16
-// CHECK:  [[TMP1:%.*]] = fcmp oge <4 x half> %a, %b
-// CHECK:  [[TMP2:%.*]] = sext <4 x i1> [[TMP1]] to <4 x i16>
-// CHECK:  ret <4 x i16> [[TMP2]]
-uint16x4_t test_vcge_f16(float16x4_t a, float16x4_t b) {
-  return vcge_f16(a, b);
-}
-
-// CHECK-LABEL: test_vcgeq_f16
-// CHECK:  [[TMP1:%.*]] = fcmp oge <8 x half> %a, %b
-// CHECK:  [[TMP2:%.*]] = sext <8 x i1> [[TMP1:%.*]] to <8 x i16>
-// CHECK:  ret <8 x i16> [[TMP2]]
-uint16x8_t test_vcgeq_f16(float16x8_t a, float16x8_t b) {
-  return vcgeq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vcgt_f16
-// CHECK:  [[TMP1:%.*]] = fcmp ogt <4 x half> %a, %b
-// CHECK:  [[TMP2:%.*]] = sext <4 x i1> [[TMP1]] to <4 x i16>
-// CHECK:  ret <4 x i16> [[TMP2]]
-uint16x4_t test_vcgt_f16(float16x4_t a, float16x4_t b) {
-  return vcgt_f16(a, b);
-}
-
-// CHECK-LABEL: test_vcgtq_f16
-// CHECK:  [[TMP1:%.*]] = fcmp ogt <8 x half> %a, %b
-// CHECK:  [[TMP2:%.*]] = sext <8 x i1> [[TMP1:%.*]] to <8 x i16>
-// CHECK:  ret <8 x i16> [[TMP2]]
-uint16x8_t test_vcgtq_f16(float16x8_t a, float16x8_t b) {
-  return vcgtq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vcle_f16
-// CHECK:  [[TMP1:%.*]] = fcmp ole <4 x half> %a, %b
-// CHECK:  [[TMP2:%.*]] = sext <4 x i1> [[TMP1]] to <4 x i16>
-// CHECK:  ret <4 x i16> [[TMP2]]
-uint16x4_t test_vcle_f16(float16x4_t a, float16x4_t b) {
-  return vcle_f16(a, b);
-}
-
-// CHECK-LABEL: test_vcleq_f16
-// CHECK:  [[TMP1:%.*]] = fcmp ole <8 x half> %a, %b
-// CHECK:  [[TMP2:%.*]] = sext <8 x i1> [[TMP1:%.*]] to <8 x i16>
-// CHECK:  ret <8 x i16> [[TMP2]]
-uint16x8_t test_vcleq_f16(float16x8_t a, float16x8_t b) {
-  return vcleq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vclt_f16
-// CHECK:  [[TMP1:%.*]] = fcmp olt <4 x half> %a, %b
-// CHECK:  [[TMP2:%.*]] = sext <4 x i1> [[TMP1]] to <4 x i16>
-// CHECK:  ret <4 x i16> [[TMP2]]
-uint16x4_t test_vclt_f16(float16x4_t a, float16x4_t b) {
-  return vclt_f16(a, b);
-}
-
-// CHECK-LABEL: test_vcltq_f16
-// CHECK:  [[TMP1:%.*]] = fcmp olt <8 x half> %a, %b
-// CHECK:  [[TMP2:%.*]] = sext <8 x i1> [[TMP1:%.*]] to <8 x i16>
-// CHECK:  ret <8 x i16> [[TMP2]]
-uint16x8_t test_vcltq_f16(float16x8_t a, float16x8_t b) {
-  return vcltq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vcvt_n_f16_s16
-// CHECK:  [[CVT:%.*]] = call <4 x half> @llvm.aarch64.neon.vcvtfxs2fp.v4f16.v4i16(<4 x i16> %vcvt_n, i32 2)
-// CHECK:  ret <4 x half> [[CVT]]
-float16x4_t test_vcvt_n_f16_s16(int16x4_t a) {
-  return vcvt_n_f16_s16(a, 2);
-}
-
-// CHECK-LABEL: test_vcvtq_n_f16_s16
-// CHECK:  [[CVT:%.*]] = call <8 x half> @llvm.aarch64.neon.vcvtfxs2fp.v8f16.v8i16(<8 x i16> %vcvt_n, i32 2)
-// CHECK:  ret <8 x half> [[CVT]]
-float16x8_t test_vcvtq_n_f16_s16(int16x8_t a) {
-  return vcvtq_n_f16_s16(a, 2);
-}
-
-// CHECK-LABEL: test_vcvt_n_f16_u16
-// CHECK:  [[CVT:%.*]] = call <4 x half> @llvm.aarch64.neon.vcvtfxu2fp.v4f16.v4i16(<4 x i16> %vcvt_n, i32 2)
-// CHECK:  ret <4 x half> [[CVT]]
-float16x4_t test_vcvt_n_f16_u16(uint16x4_t a) {
-  return vcvt_n_f16_u16(a, 2);
-}
-
-// CHECK-LABEL: test_vcvtq_n_f16_u16
-// CHECK:  [[CVT:%.*]] = call <8 x half> @llvm.aarch64.neon.vcvtfxu2fp.v8f16.v8i16(<8 x i16> %vcvt_n, i32 2)
-// CHECK:  ret <8 x half> [[CVT]]
-float16x8_t test_vcvtq_n_f16_u16(uint16x8_t a) {
-  return vcvtq_n_f16_u16(a, 2);
-}
-
-// CHECK-LABEL: test_vcvt_n_s16_f16
-// CHECK:  [[CVT:%.*]] = call <4 x i16> @llvm.aarch64.neon.vcvtfp2fxs.v4i16.v4f16(<4 x half> %vcvt_n, i32 2)
-// CHECK:  ret <4 x i16> [[CVT]]
-int16x4_t test_vcvt_n_s16_f16(float16x4_t a) {
-  return vcvt_n_s16_f16(a, 2);
-}
-
-// CHECK-LABEL: test_vcvtq_n_s16_f16
-// CHECK:  [[CVT:%.*]] = call <8 x i16> @llvm.aarch64.neon.vcvtfp2fxs.v8i16.v8f16(<8 x half> %vcvt_n, i32 2)
-// CHECK:  ret <8 x i16> [[CVT]]
-int16x8_t test_vcvtq_n_s16_f16(float16x8_t a) {
-  return vcvtq_n_s16_f16(a, 2);
-}
-
-// CHECK-LABEL: test_vcvt_n_u16_f16
-// CHECK:  [[CVT:%.*]] = call <4 x i16> @llvm.aarch64.neon.vcvtfp2fxu.v4i16.v4f16(<4 x half> %vcvt_n, i32 2)
-// CHECK:  ret <4 x i16> [[CVT]]
-uint16x4_t test_vcvt_n_u16_f16(float16x4_t a) {
-  return vcvt_n_u16_f16(a, 2);
-}
-
-// CHECK-LABEL: test_vcvtq_n_u16_f16
-// CHECK:  [[CVT:%.*]] = call <8 x i16> @llvm.aarch64.neon.vcvtfp2fxu.v8i16.v8f16(<8 x half> %vcvt_n, i32 2)
-// CHECK:  ret <8 x i16> [[CVT]]
-uint16x8_t test_vcvtq_n_u16_f16(float16x8_t a) {
-  return vcvtq_n_u16_f16(a, 2);
-}
-
-// CHECK-LABEL: test_vdiv_f16
-// CHECK:  [[DIV:%.*]] = fdiv <4 x half> %a, %b
-// CHECK:  ret <4 x half> [[DIV]]
-float16x4_t test_vdiv_f16(float16x4_t a, float16x4_t b) {
-  return vdiv_f16(a, b);
-}
-
-// CHECK-LABEL: test_vdivq_f16
-// CHECK:  [[DIV:%.*]] = fdiv <8 x half> %a, %b
-// CHECK:  ret <8 x half> [[DIV]]
-float16x8_t test_vdivq_f16(float16x8_t a, float16x8_t b) {
-  return vdivq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vmax_f16
-// CHECK:  [[MAX:%.*]] = call <4 x half> @llvm.aarch64.neon.fmax.v4f16(<4 x half> %a, <4 x half> %b)
-// CHECK:  ret <4 x half> [[MAX]]
-float16x4_t test_vmax_f16(float16x4_t a, float16x4_t b) {
-  return vmax_f16(a, b);
-}
-
-// CHECK-LABEL: test_vmaxq_f16
-// CHECK:  [[MAX:%.*]] = call <8 x half> @llvm.aarch64.neon.fmax.v8f16(<8 x half> %a, <8 x half> %b)
-// CHECK:  ret <8 x half> [[MAX]]
-float16x8_t test_vmaxq_f16(float16x8_t a, float16x8_t b) {
-  return vmaxq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vmaxnm_f16
-// CHECK:  [[MAX:%.*]] = call <4 x half> @llvm.aarch64.neon.fmaxnm.v4f16(<4 x half> %a, <4 x half> %b)
-// CHECK:  ret <4 x half> [[MAX]]
-float16x4_t test_vmaxnm_f16(float16x4_t a, float16x4_t b) {
-  return vmaxnm_f16(a, b);
-}
-
-// CHECK-LABEL: test_vmaxnmq_f16
-// CHECK:  [[MAX:%.*]] = call <8 x half> @llvm.aarch64.neon.fmaxnm.v8f16(<8 x half> %a, <8 x half> %b)
-// CHECK:  ret <8 x half> [[MAX]]
-float16x8_t test_vmaxnmq_f16(float16x8_t a, float16x8_t b) {
-  return vmaxnmq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vmin_f16
-// CHECK:  [[MIN:%.*]] = call <4 x half> @llvm.aarch64.neon.fmin.v4f16(<4 x half> %a, <4 x half> %b)
-// CHECK:  ret <4 x half> [[MIN]]
-float16x4_t test_vmin_f16(float16x4_t a, float16x4_t b) {
-  return vmin_f16(a, b);
-}
-
-// CHECK-LABEL: test_vminq_f16
-// CHECK:  [[MIN:%.*]] = call <8 x half> @llvm.aarch64.neon.fmin.v8f16(<8 x half> %a, <8 x half> %b)
-// CHECK:  ret <8 x half> [[MIN]]
-float16x8_t test_vminq_f16(float16x8_t a, float16x8_t b) {
-  return vminq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vminnm_f16
-// CHECK:  [[MIN:%.*]] = call <4 x half> @llvm.aarch64.neon.fminnm.v4f16(<4 x half> %a, <4 x half> %b)
-// CHECK:  ret <4 x half> [[MIN]]
-float16x4_t test_vminnm_f16(float16x4_t a, float16x4_t b) {
-  return vminnm_f16(a, b);
-}
-
-// CHECK-LABEL: test_vminnmq_f16
-// CHECK:  [[MIN:%.*]] = call <8 x half> @llvm.aarch64.neon.fminnm.v8f16(<8 x half> %a, <8 x half> %b)
-// CHECK:  ret <8 x half> [[MIN]]
-float16x8_t test_vminnmq_f16(float16x8_t a, float16x8_t b) {
-  return vminnmq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vmul_f16
-// CHECK:  [[MUL:%.*]] = fmul <4 x half> %a, %b
-// CHECK:  ret <4 x half> [[MUL]]
-float16x4_t test_vmul_f16(float16x4_t a, float16x4_t b) {
-  return vmul_f16(a, b);
-}
-
-// CHECK-LABEL: test_vmulq_f16
-// CHECK:  [[MUL:%.*]] = fmul <8 x half> %a, %b
-// CHECK:  ret <8 x half> [[MUL]]
-float16x8_t test_vmulq_f16(float16x8_t a, float16x8_t b) {
-  return vmulq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vmulx_f16
-// CHECK:  [[MUL:%.*]] = call <4 x half> @llvm.aarch64.neon.fmulx.v4f16(<4 x half> %a, <4 x half> %b)
-// CHECK:  ret <4 x half> [[MUL]]
-float16x4_t test_vmulx_f16(float16x4_t a, float16x4_t b) {
-  return vmulx_f16(a, b);
-}
-
-// CHECK-LABEL: test_vmulxq_f16
-// CHECK:  [[MUL:%.*]] = call <8 x half> @llvm.aarch64.neon.fmulx.v8f16(<8 x half> %a, <8 x half> %b)
-// CHECK:  ret <8 x half> [[MUL]]
-float16x8_t test_vmulxq_f16(float16x8_t a, float16x8_t b) {
-  return vmulxq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vpadd_f16
-// CHECK:  [[ADD:%.*]] = call <4 x half> @llvm.aarch64.neon.addp.v4f16(<4 x half> %a, <4 x half> %b)
-// CHECK:  ret <4 x half> [[ADD]]
-float16x4_t test_vpadd_f16(float16x4_t a, float16x4_t b) {
-  return vpadd_f16(a, b);
-}
-
-// CHECK-LABEL: test_vpaddq_f16
-// CHECK:  [[ADD:%.*]] = call <8 x half> @llvm.aarch64.neon.addp.v8f16(<8 x half> %a, <8 x half> %b)
-// CHECK:  ret <8 x half> [[ADD]]
-float16x8_t test_vpaddq_f16(float16x8_t a, float16x8_t b) {
-  return vpaddq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vpmax_f16
-// CHECK:  [[MAX:%.*]] = call <4 x half> @llvm.aarch64.neon.fmaxp.v4f16(<4 x half> %a, <4 x half> %b)
-// CHECK:  ret <4 x half> [[MAX]]
-float16x4_t test_vpmax_f16(float16x4_t a, float16x4_t b) {
-  return vpmax_f16(a, b);
-}
-
-// CHECK-LABEL: test_vpmaxq_f16
-// CHECK:  [[MAX:%.*]] = call <8 x half> @llvm.aarch64.neon.fmaxp.v8f16(<8 x half> %a, <8 x half> %b)
-// CHECK:  ret <8 x half> [[MAX]]
-float16x8_t test_vpmaxq_f16(float16x8_t a, float16x8_t b) {
-  return vpmaxq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vpmaxnm_f16
-// CHECK:  [[MAX:%.*]] = call <4 x half> @llvm.aarch64.neon.fmaxnmp.v4f16(<4 x half> %a, <4 x half> %b)
-// CHECK:  ret <4 x half> [[MAX]]
-float16x4_t test_vpmaxnm_f16(float16x4_t a, float16x4_t b) {
-  return vpmaxnm_f16(a, b);
-}
-
-// CHECK-LABEL: test_vpmaxnmq_f16
-// CHECK:  [[MAX:%.*]] = call <8 x half> @llvm.aarch64.neon.fmaxnmp.v8f16(<8 x half> %a, <8 x half> %b)
-// CHECK:  ret <8 x half> [[MAX]]
-float16x8_t test_vpmaxnmq_f16(float16x8_t a, float16x8_t b) {
-  return vpmaxnmq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vpmin_f16
-// CHECK:  [[MIN:%.*]] = call <4 x half> @llvm.aarch64.neon.fminp.v4f16(<4 x half> %a, <4 x half> %b)
-// CHECK:  ret <4 x half> [[MIN]]
-float16x4_t test_vpmin_f16(float16x4_t a, float16x4_t b) {
-  return vpmin_f16(a, b);
-}
-
-// CHECK-LABEL: test_vpminq_f16
-// CHECK:  [[MIN:%.*]] = call <8 x half> @llvm.aarch64.neon.fminp.v8f16(<8 x half> %a, <8 x half> %b)
-// CHECK:  ret <8 x half> [[MIN]]
-float16x8_t test_vpminq_f16(float16x8_t a, float16x8_t b) {
-  return vpminq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vpminnm_f16
-// CHECK:  [[MIN:%.*]] = call <4 x half> @llvm.aarch64.neon.fminnmp.v4f16(<4 x half> %a, <4 x half> %b)
-// CHECK:  ret <4 x half> [[MIN]]
-float16x4_t test_vpminnm_f16(float16x4_t a, float16x4_t b) {
-  return vpminnm_f16(a, b);
-}
-
-// CHECK-LABEL: test_vpminnmq_f16
-// CHECK:  [[MIN:%.*]] = call <8 x half> @llvm.aarch64.neon.fminnmp.v8f16(<8 x half> %a, <8 x half> %b)
-// CHECK:  ret <8 x half> [[MIN]]
-float16x8_t test_vpminnmq_f16(float16x8_t a, float16x8_t b) {
-  return vpminnmq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vrecps_f16
-// CHECK:  [[MIN:%.*]] = call <4 x half> @llvm.aarch64.neon.frecps.v4f16(<4 x half> %a, <4 x half> %b)
-// CHECK:  ret <4 x half> [[MIN]]
-float16x4_t test_vrecps_f16(float16x4_t a, float16x4_t b) {
-  return vrecps_f16(a, b);
-}
-
-// CHECK-LABEL: test_vrecpsq_f16
-// CHECK:  [[MIN:%.*]] =  call <8 x half> @llvm.aarch64.neon.frecps.v8f16(<8 x half> %a, <8 x half> %b)
-// CHECK:  ret <8 x half> [[MIN]]
-float16x8_t test_vrecpsq_f16(float16x8_t a, float16x8_t b) {
-  return vrecpsq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vrsqrts_f16
-// CHECK:  [[MIN:%.*]] = call <4 x half> @llvm.aarch64.neon.frsqrts.v4f16(<4 x half> %a, <4 x half> %b)
-// CHECK:  ret <4 x half> [[MIN]]
-float16x4_t test_vrsqrts_f16(float16x4_t a, float16x4_t b) {
-  return vrsqrts_f16(a, b);
-}
-
-// CHECK-LABEL: test_vrsqrtsq_f16
-// CHECK:  [[MIN:%.*]] =  call <8 x half> @llvm.aarch64.neon.frsqrts.v8f16(<8 x half> %a, <8 x half> %b)
-// CHECK:  ret <8 x half> [[MIN]]
-float16x8_t test_vrsqrtsq_f16(float16x8_t a, float16x8_t b) {
-  return vrsqrtsq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vsub_f16
-// CHECK:  [[ADD:%.*]] = fsub <4 x half> %a, %b 
-// CHECK:  ret <4 x half> [[ADD]]
-float16x4_t test_vsub_f16(float16x4_t a, float16x4_t b) {
-  return vsub_f16(a, b);
-}
-
-// CHECK-LABEL: test_vsubq_f16
-// CHECK:  [[ADD:%.*]] = fsub <8 x half> %a, %b
-// CHECK:  ret <8 x half> [[ADD]]
-float16x8_t test_vsubq_f16(float16x8_t a, float16x8_t b) {
-  return vsubq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vfma_f16
-// CHECK:  [[ADD:%.*]] = call <4 x half> @llvm.fma.v4f16(<4 x half> %b, <4 x half> %c, <4 x half> %a)
-// CHECK:  ret <4 x half> [[ADD]]
-float16x4_t test_vfma_f16(float16x4_t a, float16x4_t b, float16x4_t c) {
-  return vfma_f16(a, b, c);
-}
-
-// CHECK-LABEL: test_vfmaq_f16
-// CHECK:  [[ADD:%.*]] = call <8 x half> @llvm.fma.v8f16(<8 x half> %b, <8 x half> %c, <8 x half> %a)
-// CHECK:  ret <8 x half> [[ADD]]
-float16x8_t test_vfmaq_f16(float16x8_t a, float16x8_t b, float16x8_t c) {
-  return vfmaq_f16(a, b, c);
-}
-
-// CHECK-LABEL: test_vfms_f16
-// CHECK:  [[SUB:%.*]] = fsub <4 x half> <half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000>, %b
-// CHECK:  [[ADD:%.*]] = call <4 x half> @llvm.fma.v4f16(<4 x half> [[SUB]], <4 x half> %c, <4 x half> %a)
-// CHECK:  ret <4 x half> [[ADD]]
-float16x4_t test_vfms_f16(float16x4_t a, float16x4_t b, float16x4_t c) {
-  return vfms_f16(a, b, c);
-}
-
-// CHECK-LABEL: test_vfmsq_f16
-// CHECK:  [[SUB:%.*]] = fsub <8 x half> <half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000>, %b
-// CHECK:  [[ADD:%.*]] = call <8 x half> @llvm.fma.v8f16(<8 x half> [[SUB]], <8 x half> %c, <8 x half> %a)
-// CHECK:  ret <8 x half> [[ADD]]
-float16x8_t test_vfmsq_f16(float16x8_t a, float16x8_t b, float16x8_t c) {
-  return vfmsq_f16(a, b, c);
-}
-
-// CHECK-LABEL: test_vfma_lane_f16
-// CHECK: [[TMP0:%.*]] = bitcast <4 x half> %a to <8 x i8>
-// CHECK: [[TMP1:%.*]] = bitcast <4 x half> %b to <8 x i8>
-// CHECK: [[TMP2:%.*]] = bitcast <4 x half> %c to <8 x i8>
-// CHECK: [[TMP3:%.*]] = bitcast <8 x i8> [[TMP2]] to <4 x half>
-// CHECK: [[LANE:%.*]] = shufflevector <4 x half> [[TMP3]], <4 x half> [[TMP3]], <4 x i32> <i32 3, i32 3, i32 3, i32 3>
-// CHECK: [[TMP4:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x half>
-// CHECK: [[TMP5:%.*]] = bitcast <8 x i8> [[TMP0]] to <4 x half>
-// CHECK: [[FMLA:%.*]] = call <4 x half> @llvm.fma.v4f16(<4 x half> [[TMP4]], <4 x half> [[LANE]], <4 x half> [[TMP5]])
-// CHECK: ret <4 x half> [[FMLA]]
-float16x4_t test_vfma_lane_f16(float16x4_t a, float16x4_t b, float16x4_t c) {
-  return vfma_lane_f16(a, b, c, 3);
-}
-
-// CHECK-LABEL: test_vfmaq_lane_f16
-// CHECK: [[TMP0:%.*]] = bitcast <8 x half> %a to <16 x i8>
-// CHECK: [[TMP1:%.*]] = bitcast <8 x half> %b to <16 x i8>
-// CHECK: [[TMP2:%.*]] = bitcast <4 x half> %c to <8 x i8>
-// CHECK: [[TMP3:%.*]] = bitcast <8 x i8> [[TMP2]] to <4 x half>
-// CHECK: [[LANE:%.*]] = shufflevector <4 x half> [[TMP3]], <4 x half> [[TMP3]], <8 x i32> <i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3>
-// CHECK: [[TMP4:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x half>
-// CHECK: [[TMP5:%.*]] = bitcast <16 x i8> [[TMP0]] to <8 x half>
-// CHECK: [[FMLA:%.*]] = call <8 x half> @llvm.fma.v8f16(<8 x half> [[TMP4]], <8 x half> [[LANE]], <8 x half> [[TMP5]])
-// CHECK: ret <8 x half> [[FMLA]]
-float16x8_t test_vfmaq_lane_f16(float16x8_t a, float16x8_t b, float16x4_t c) {
-  return vfmaq_lane_f16(a, b, c, 3);
-}
-
-// CHECK-LABEL: test_vfma_laneq_f16
-// CHECK: [[TMP0:%.*]] = bitcast <4 x half> %a to <8 x i8>
-// CHECK: [[TMP1:%.*]] = bitcast <4 x half> %b to <8 x i8>
-// CHECK: [[TMP2:%.*]] = bitcast <8 x half> %c to <16 x i8>
-// CHECK: [[TMP3:%.*]] = bitcast <8 x i8> [[TMP0]] to <4 x half>
-// CHECK: [[TMP4:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x half>
-// CHECK: [[TMP5:%.*]] = bitcast <16 x i8> [[TMP2]] to <8 x half>
-// CHECK: [[LANE:%.*]] = shufflevector <8 x half> [[TMP5]], <8 x half> [[TMP5]], <4 x i32> <i32 7, i32 7, i32 7, i32 7>
-// CHECK: [[FMLA:%.*]] = call <4 x half> @llvm.fma.v4f16(<4 x half> [[LANE]], <4 x half> [[TMP4]], <4 x half> [[TMP3]])
-// CHECK: ret <4 x half> [[FMLA]]
-float16x4_t test_vfma_laneq_f16(float16x4_t a, float16x4_t b, float16x8_t c) {
-  return vfma_laneq_f16(a, b, c, 7);
-}
-
-// CHECK-LABEL: test_vfmaq_laneq_f16
-// CHECK: [[TMP0:%.*]] = bitcast <8 x half> %a to <16 x i8>
-// CHECK: [[TMP1:%.*]] = bitcast <8 x half> %b to <16 x i8>
-// CHECK: [[TMP2:%.*]] = bitcast <8 x half> %c to <16 x i8>
-// CHECK: [[TMP3:%.*]] = bitcast <16 x i8> [[TMP0]] to <8 x half>
-// CHECK: [[TMP4:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x half>
-// CHECK: [[TMP5:%.*]] = bitcast <16 x i8> [[TMP2]] to <8 x half>
-// CHECK: [[LANE:%.*]] = shufflevector <8 x half> [[TMP5]], <8 x half> [[TMP5]], <8 x i32> <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-// CHECK: [[FMLA:%.*]] = call <8 x half> @llvm.fma.v8f16(<8 x half> [[LANE]], <8 x half> [[TMP4]], <8 x half> [[TMP3]])
-// CHECK: ret <8 x half> [[FMLA]]
-float16x8_t test_vfmaq_laneq_f16(float16x8_t a, float16x8_t b, float16x8_t c) {
-  return vfmaq_laneq_f16(a, b, c, 7);
-}
-
-// CHECK-LABEL: test_vfma_n_f16
-// CHECK: [[TMP0:%.*]] = insertelement <4 x half> undef, half %c, i32 0
-// CHECK: [[TMP1:%.*]] = insertelement <4 x half> [[TMP0]], half %c, i32 1
-// CHECK: [[TMP2:%.*]] = insertelement <4 x half> [[TMP1]], half %c, i32 2
-// CHECK: [[TMP3:%.*]] = insertelement <4 x half> [[TMP2]], half %c, i32 3
-// CHECK: [[FMA:%.*]]  = call <4 x half> @llvm.fma.v4f16(<4 x half> %b, <4 x half> [[TMP3]], <4 x half> %a)
-// CHECK: ret <4 x half> [[FMA]]
-float16x4_t test_vfma_n_f16(float16x4_t a, float16x4_t b, float16_t c) {
-  return vfma_n_f16(a, b, c);
-}
-
-// CHECK-LABEL: test_vfmaq_n_f16
-// CHECK: [[TMP0:%.*]] = insertelement <8 x half> undef, half %c, i32 0
-// CHECK: [[TMP1:%.*]] = insertelement <8 x half> [[TMP0]], half %c, i32 1
-// CHECK: [[TMP2:%.*]] = insertelement <8 x half> [[TMP1]], half %c, i32 2
-// CHECK: [[TMP3:%.*]] = insertelement <8 x half> [[TMP2]], half %c, i32 3
-// CHECK: [[TMP4:%.*]] = insertelement <8 x half> [[TMP3]], half %c, i32 4
-// CHECK: [[TMP5:%.*]] = insertelement <8 x half> [[TMP4]], half %c, i32 5
-// CHECK: [[TMP6:%.*]] = insertelement <8 x half> [[TMP5]], half %c, i32 6
-// CHECK: [[TMP7:%.*]] = insertelement <8 x half> [[TMP6]], half %c, i32 7
-// CHECK: [[FMA:%.*]]  = call <8 x half> @llvm.fma.v8f16(<8 x half> %b, <8 x half> [[TMP7]], <8 x half> %a)
-// CHECK: ret <8 x half> [[FMA]]
-float16x8_t test_vfmaq_n_f16(float16x8_t a, float16x8_t b, float16_t c) {
-  return vfmaq_n_f16(a, b, c);
-}
-
-// CHECK-LABEL: test_vfmah_lane_f16
-// CHECK: [[TMP0:%.*]] = bitcast <4 x half> %c to <8 x i8>
-// CHECK: [[TMP1:%.*]] = bitcast <8 x i8> [[TMP0]] to <4 x half>
-// CHECK: [[EXTR:%.*]] = extractelement <4 x half> [[TMP1]], i32 3
-// CHECK: [[FMA:%.*]]  = call half @llvm.fma.f16(half %b, half [[EXTR]], half %a)
-// CHECK: ret half [[FMA]]
-float16_t test_vfmah_lane_f16(float16_t a, float16_t b, float16x4_t c) {
-  return vfmah_lane_f16(a, b, c, 3);
-}
-
-// CHECK-LABEL: test_vfmah_laneq_f16
-// CHECK: [[TMP0:%.*]] = bitcast <8 x half> %c to <16 x i8>
-// CHECK: [[TMP1:%.*]] = bitcast <16 x i8> [[TMP0]] to <8 x half>
-// CHECK: [[EXTR:%.*]] = extractelement <8 x half> [[TMP1]], i32 7
-// CHECK: [[FMA:%.*]]  = call half @llvm.fma.f16(half %b, half [[EXTR]], half %a)
-// CHECK: ret half [[FMA]]
-float16_t test_vfmah_laneq_f16(float16_t a, float16_t b, float16x8_t c) {
-  return vfmah_laneq_f16(a, b, c, 7);
-}
-
-// CHECK-LABEL: test_vfms_lane_f16
-// CHECK: [[SUB:%.*]]  = fsub <4 x half> <half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000>, %b
-// CHECK: [[TMP0:%.*]] = bitcast <4 x half> %a to <8 x i8>
-// CHECK: [[TMP1:%.*]] = bitcast <4 x half> [[SUB]] to <8 x i8>
-// CHECK: [[TMP2:%.*]] = bitcast <4 x half> %c to <8 x i8>
-// CHECK: [[TMP3:%.*]] = bitcast <8 x i8> [[TMP2]] to <4 x half>
-// CHECK: [[LANE:%.*]] = shufflevector <4 x half> [[TMP3]], <4 x half> [[TMP3]], <4 x i32> <i32 3, i32 3, i32 3, i32 3>
-// CHECK: [[TMP4:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x half>
-// CHECK: [[TMP5:%.*]] = bitcast <8 x i8> [[TMP0]] to <4 x half>
-// CHECK: [[FMA:%.*]] = call <4 x half> @llvm.fma.v4f16(<4 x half> [[TMP4]], <4 x half> [[LANE]], <4 x half> [[TMP5]])
-// CHECK: ret <4 x half> [[FMA]]
-float16x4_t test_vfms_lane_f16(float16x4_t a, float16x4_t b, float16x4_t c) {
-  return vfms_lane_f16(a, b, c, 3);
-}
-
-// CHECK-LABEL: test_vfmsq_lane_f16
-// CHECK: [[SUB:%.*]]  = fsub <8 x half> <half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000>, %b
-// CHECK: [[TMP0:%.*]] = bitcast <8 x half> %a to <16 x i8>
-// CHECK: [[TMP1:%.*]] = bitcast <8 x half> [[SUB]] to <16 x i8>
-// CHECK: [[TMP2:%.*]] = bitcast <4 x half> %c to <8 x i8>
-// CHECK: [[TMP3:%.*]] = bitcast <8 x i8> [[TMP2]] to <4 x half>
-// CHECK: [[LANE:%.*]] = shufflevector <4 x half> [[TMP3]], <4 x half> [[TMP3]], <8 x i32> <i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3, i32 3>
-// CHECK: [[TMP4:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x half>
-// CHECK: [[TMP5:%.*]] = bitcast <16 x i8> [[TMP0]] to <8 x half>
-// CHECK: [[FMLA:%.*]] = call <8 x half> @llvm.fma.v8f16(<8 x half> [[TMP4]], <8 x half> [[LANE]], <8 x half> [[TMP5]])
-// CHECK: ret <8 x half> [[FMLA]]
-float16x8_t test_vfmsq_lane_f16(float16x8_t a, float16x8_t b, float16x4_t c) {
-  return vfmsq_lane_f16(a, b, c, 3);
-}
-
-// CHECK-LABEL: test_vfms_laneq_f16
-// CHECK: [[SUB:%.*]]  = fsub <4 x half> <half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000>, %b
-// CHECK: [[TMP0:%.*]] = bitcast <4 x half> %a to <8 x i8>
-// CHECK: [[TMP1:%.*]] = bitcast <4 x half> [[SUB]] to <8 x i8>
-// CHECK: [[TMP2:%.*]] = bitcast <8 x half> %c to <16 x i8>
-// CHECK: [[TMP3:%.*]] = bitcast <8 x i8> [[TMP0]] to <4 x half>
-// CHECK: [[TMP4:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x half>
-// CHECK: [[TMP5:%.*]] = bitcast <16 x i8> [[TMP2]] to <8 x half>
-// CHECK: [[LANE:%.*]] = shufflevector <8 x half> [[TMP5]], <8 x half> [[TMP5]], <4 x i32> <i32 7, i32 7, i32 7, i32 7>
-// CHECK: [[FMLA:%.*]] = call <4 x half> @llvm.fma.v4f16(<4 x half> [[LANE]], <4 x half> [[TMP4]], <4 x half> [[TMP3]])
-// CHECK: ret <4 x half> [[FMLA]]
-float16x4_t test_vfms_laneq_f16(float16x4_t a, float16x4_t b, float16x8_t c) {
-  return vfms_laneq_f16(a, b, c, 7);
-}
-
-// CHECK-LABEL: test_vfmsq_laneq_f16
-// CHECK: [[SUB:%.*]]  = fsub <8 x half> <half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000>, %b
-// CHECK: [[TMP0:%.*]] = bitcast <8 x half> %a to <16 x i8>
-// CHECK: [[TMP1:%.*]] = bitcast <8 x half> [[SUB]] to <16 x i8>
-// CHECK: [[TMP2:%.*]] = bitcast <8 x half> %c to <16 x i8>
-// CHECK: [[TMP3:%.*]] = bitcast <16 x i8> [[TMP0]] to <8 x half>
-// CHECK: [[TMP4:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x half>
-// CHECK: [[TMP5:%.*]] = bitcast <16 x i8> [[TMP2]] to <8 x half>
-// CHECK: [[LANE:%.*]] = shufflevector <8 x half> [[TMP5]], <8 x half> [[TMP5]], <8 x i32> <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-// CHECK: [[FMLA:%.*]] = call <8 x half> @llvm.fma.v8f16(<8 x half> [[LANE]], <8 x half> [[TMP4]], <8 x half> [[TMP3]])
-// CHECK: ret <8 x half> [[FMLA]]
-float16x8_t test_vfmsq_laneq_f16(float16x8_t a, float16x8_t b, float16x8_t c) {
-  return vfmsq_laneq_f16(a, b, c, 7);
-}
-
-// CHECK-LABEL: test_vfms_n_f16
-// CHECK: [[SUB:%.*]]  = fsub <4 x half> <half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000>, %b
-// CHECK: [[TMP0:%.*]] = insertelement <4 x half> undef, half %c, i32 0
-// CHECK: [[TMP1:%.*]] = insertelement <4 x half> [[TMP0]], half %c, i32 1
-// CHECK: [[TMP2:%.*]] = insertelement <4 x half> [[TMP1]], half %c, i32 2
-// CHECK: [[TMP3:%.*]] = insertelement <4 x half> [[TMP2]], half %c, i32 3
-// CHECK: [[FMA:%.*]]  = call <4 x half> @llvm.fma.v4f16(<4 x half> [[SUB]], <4 x half> [[TMP3]], <4 x half> %a)
-// CHECK: ret <4 x half> [[FMA]]
-float16x4_t test_vfms_n_f16(float16x4_t a, float16x4_t b, float16_t c) {
-  return vfms_n_f16(a, b, c);
-}
-
-// CHECK-LABEL: test_vfmsq_n_f16
-// CHECK: [[SUB:%.*]]  = fsub <8 x half> <half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000, half 0xH8000>, %b
-// CHECK: [[TMP0:%.*]] = insertelement <8 x half> undef, half %c, i32 0
-// CHECK: [[TMP1:%.*]] = insertelement <8 x half> [[TMP0]], half %c, i32 1
-// CHECK: [[TMP2:%.*]] = insertelement <8 x half> [[TMP1]], half %c, i32 2
-// CHECK: [[TMP3:%.*]] = insertelement <8 x half> [[TMP2]], half %c, i32 3
-// CHECK: [[TMP4:%.*]] = insertelement <8 x half> [[TMP3]], half %c, i32 4
-// CHECK: [[TMP5:%.*]] = insertelement <8 x half> [[TMP4]], half %c, i32 5
-// CHECK: [[TMP6:%.*]] = insertelement <8 x half> [[TMP5]], half %c, i32 6
-// CHECK: [[TMP7:%.*]] = insertelement <8 x half> [[TMP6]], half %c, i32 7
-// CHECK: [[FMA:%.*]]  = call <8 x half> @llvm.fma.v8f16(<8 x half> [[SUB]], <8 x half> [[TMP7]], <8 x half> %a)
-// CHECK: ret <8 x half> [[FMA]]
-float16x8_t test_vfmsq_n_f16(float16x8_t a, float16x8_t b, float16_t c) {
-  return vfmsq_n_f16(a, b, c);
-}
-
-// CHECK-LABEL: test_vfmsh_lane_f16
-// CHECK: [[TMP0:%.*]] = fpext half %b to float
-// CHECK: [[TMP1:%.*]] = fsub float -0.000000e+00, [[TMP0]]
-// CHECK: [[SUB:%.*]]  = fptrunc float [[TMP1]] to half
-// CHECK: [[TMP2:%.*]] = bitcast <4 x half> %c to <8 x i8>
-// CHECK: [[TMP3:%.*]] = bitcast <8 x i8> [[TMP2]] to <4 x half>
-// CHECK: [[EXTR:%.*]] = extractelement <4 x half> [[TMP3]], i32 3
-// CHECK: [[FMA:%.*]]  = call half @llvm.fma.f16(half [[SUB]], half [[EXTR]], half %a)
-// CHECK: ret half [[FMA]]
-float16_t test_vfmsh_lane_f16(float16_t a, float16_t b, float16x4_t c) {
-  return vfmsh_lane_f16(a, b, c, 3);
-}
-
-// CHECK-LABEL: test_vfmsh_laneq_f16
-// CHECK: [[TMP0:%.*]] = fpext half %b to float
-// CHECK: [[TMP1:%.*]] = fsub float -0.000000e+00, [[TMP0]]
-// CHECK: [[SUB:%.*]]  = fptrunc float [[TMP1]] to half
-// CHECK: [[TMP2:%.*]] = bitcast <8 x half> %c to <16 x i8>
-// CHECK: [[TMP3:%.*]] = bitcast <16 x i8> [[TMP2]] to <8 x half>
-// CHECK: [[EXTR:%.*]] = extractelement <8 x half> [[TMP3]], i32 7
-// CHECK: [[FMA:%.*]]  = call half @llvm.fma.f16(half [[SUB]], half [[EXTR]], half %a)
-// CHECK: ret half [[FMA]]
-float16_t test_vfmsh_laneq_f16(float16_t a, float16_t b, float16x8_t c) {
-  return vfmsh_laneq_f16(a, b, c, 7);
-}
-
-// CHECK-LABEL: test_vmul_lane_f16
-// CHECK: [[TMP0:%.*]] = shufflevector <4 x half> %b, <4 x half> %b, <4 x i32> <i32 3, i32 3, i32 3, i32 3>
-// CHECK: [[MUL:%.*]]  = fmul <4 x half> %a, [[TMP0]]
-// CHECK: ret <4 x half> [[MUL]]
-float16x4_t test_vmul_lane_f16(float16x4_t a, float16x4_t b) {
-  return vmul_lane_f16(a, b, 3);
-}
-
-// CHECK-LABEL: test_vmulq_lane_f16
-// CHECK: [[TMP0:%.*]] = shufflevector <4 x half> %b, <4 x half> %b, <8 x i32> <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-// CHECK: [[MUL:%.*]]  = fmul <8 x half> %a, [[TMP0]]
-// CHECK: ret <8 x half> [[MUL]]
-float16x8_t test_vmulq_lane_f16(float16x8_t a, float16x4_t b) {
-  return vmulq_lane_f16(a, b, 7);
-}
-
-// CHECK-LABEL: test_vmul_laneq_f16
-// CHECK: [[TMP0:%.*]] = shufflevector <8 x half> %b, <8 x half> %b, <4 x i32> <i32 7, i32 7, i32 7, i32 7>
-// CHECK: [[MUL:%.*]]  = fmul <4 x half> %a, [[TMP0]]
-// CHECK: ret <4 x half> [[MUL]]
-float16x4_t test_vmul_laneq_f16(float16x4_t a, float16x8_t b) {
-  return vmul_laneq_f16(a, b, 7);
-}
-
-// CHECK-LABEL: test_vmulq_laneq_f16
-// CHECK: [[TMP0:%.*]] = shufflevector <8 x half> %b, <8 x half> %b, <8 x i32> <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-// CHECK: [[MUL:%.*]]  = fmul <8 x half> %a, [[TMP0]]
-// CHECK: ret <8 x half> [[MUL]]
-float16x8_t test_vmulq_laneq_f16(float16x8_t a, float16x8_t b) {
-  return vmulq_laneq_f16(a, b, 7);
-}
-
-// CHECK-LABEL: test_vmul_n_f16
-// CHECK: [[TMP0:%.*]] = insertelement <4 x half> undef, half %b, i32 0
-// CHECK: [[TMP1:%.*]] = insertelement <4 x half> [[TMP0]], half %b, i32 1
-// CHECK: [[TMP2:%.*]] = insertelement <4 x half> [[TMP1]], half %b, i32 2
-// CHECK: [[TMP3:%.*]] = insertelement <4 x half> [[TMP2]], half %b, i32 3
-// CHECK: [[MUL:%.*]]  = fmul <4 x half> %a, [[TMP3]]
-// CHECK: ret <4 x half> [[MUL]]
-float16x4_t test_vmul_n_f16(float16x4_t a, float16_t b) {
-  return vmul_n_f16(a, b);
-}
-
-// CHECK-LABEL: test_vmulq_n_f16
-// CHECK: [[TMP0:%.*]] = insertelement <8 x half> undef, half %b, i32 0
-// CHECK: [[TMP1:%.*]] = insertelement <8 x half> [[TMP0]], half %b, i32 1
-// CHECK: [[TMP2:%.*]] = insertelement <8 x half> [[TMP1]], half %b, i32 2
-// CHECK: [[TMP3:%.*]] = insertelement <8 x half> [[TMP2]], half %b, i32 3
-// CHECK: [[TMP4:%.*]] = insertelement <8 x half> [[TMP3]], half %b, i32 4
-// CHECK: [[TMP5:%.*]] = insertelement <8 x half> [[TMP4]], half %b, i32 5
-// CHECK: [[TMP6:%.*]] = insertelement <8 x half> [[TMP5]], half %b, i32 6
-// CHECK: [[TMP7:%.*]] = insertelement <8 x half> [[TMP6]], half %b, i32 7
-// CHECK: [[MUL:%.*]]  = fmul <8 x half> %a, [[TMP7]]
-// CHECK: ret <8 x half> [[MUL]]
-float16x8_t test_vmulq_n_f16(float16x8_t a, float16_t b) {
-  return vmulq_n_f16(a, b);
-}
-
-// FIXME: Fix it when fp16 non-storage-only type becomes available.
-// CHECK-LABEL: test_vmulh_lane_f16
-// CHECK: [[CONV0:%.*]] = fpext half %a to float
-// CHECK: [[CONV1:%.*]] = fpext half %{{.*}} to float
-// CHECK: [[MUL:%.*]]   = fmul float [[CONV0:%.*]], [[CONV0:%.*]]
-// CHECK: [[CONV3:%.*]] = fptrunc float %mul to half
-// CHECK: ret half [[CONV3:%.*]]
-float16_t test_vmulh_lane_f16(float16_t a, float16x4_t b) {
-  return vmulh_lane_f16(a, b, 3);
-}
-
-// CHECK-LABEL: test_vmulh_laneq_f16
-// CHECK: [[CONV0:%.*]] = fpext half %a to float
-// CHECK: [[CONV1:%.*]] = fpext half %{{.*}} to float
-// CHECK: [[MUL:%.*]]   = fmul float [[CONV0:%.*]], [[CONV0:%.*]]
-// CHECK: [[CONV3:%.*]] = fptrunc float %mul to half
-// CHECK: ret half [[CONV3:%.*]]
-float16_t test_vmulh_laneq_f16(float16_t a, float16x8_t b) {
-  return vmulh_laneq_f16(a, b, 7);
-}
-
-// CHECK-LABEL: test_vmulx_lane_f16
-// CHECK: [[TMP0:%.*]] = shufflevector <4 x half> %b, <4 x half> %b, <4 x i32> <i32 3, i32 3, i32 3, i32 3>
-// CHECK: [[MUL:%.*]] = call <4 x half> @llvm.aarch64.neon.fmulx.v4f16(<4 x half> %a, <4 x half> [[TMP0]])
-// CHECK: ret <4 x half> [[MUL]]
-float16x4_t test_vmulx_lane_f16(float16x4_t a, float16x4_t b) {
-  return vmulx_lane_f16(a, b, 3);
-}
-
-// CHECK-LABEL: test_vmulxq_lane_f16
-// CHECK: [[TMP0:%.*]] = shufflevector <4 x half> %b, <4 x half> %b, <8 x i32> <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-// CHECK: [[MUL:%.*]] = call <8 x half> @llvm.aarch64.neon.fmulx.v8f16(<8 x half> %a, <8 x half> [[TMP0]])
-// CHECK: ret <8 x half> [[MUL]]
-float16x8_t test_vmulxq_lane_f16(float16x8_t a, float16x4_t b) {
-  return vmulxq_lane_f16(a, b, 7);
-}
-
-// CHECK-LABEL: test_vmulx_laneq_f16
-// CHECK: [[TMP0:%.*]] = shufflevector <8 x half> %b, <8 x half> %b, <4 x i32> <i32 7, i32 7, i32 7, i32 7>
-// CHECK: [[MUL:%.*]]  = call <4 x half> @llvm.aarch64.neon.fmulx.v4f16(<4 x half> %a, <4 x half> [[TMP0]])
-// CHECK: ret <4 x half> [[MUL]]
-float16x4_t test_vmulx_laneq_f16(float16x4_t a, float16x8_t b) {
-  return vmulx_laneq_f16(a, b, 7);
-}
-
-// CHECK-LABEL: test_vmulxq_laneq_f16
-// CHECK: [[TMP0:%.*]] = shufflevector <8 x half> %b, <8 x half> %b, <8 x i32> <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-// CHECK: [[MUL:%.*]]  = call <8 x half> @llvm.aarch64.neon.fmulx.v8f16(<8 x half> %a, <8 x half> [[TMP0]])
-// CHECK: ret <8 x half> [[MUL]]
-float16x8_t test_vmulxq_laneq_f16(float16x8_t a, float16x8_t b) {
-  return vmulxq_laneq_f16(a, b, 7);
-}
-
-// CHECK-LABEL: test_vmulx_n_f16
-// CHECK: [[TMP0:%.*]] = insertelement <4 x half> undef, half %b, i32 0
-// CHECK: [[TMP1:%.*]] = insertelement <4 x half> [[TMP0]], half %b, i32 1
-// CHECK: [[TMP2:%.*]] = insertelement <4 x half> [[TMP1]], half %b, i32 2
-// CHECK: [[TMP3:%.*]] = insertelement <4 x half> [[TMP2]], half %b, i32 3
-// CHECK: [[MUL:%.*]]  = call <4 x half> @llvm.aarch64.neon.fmulx.v4f16(<4 x half> %a, <4 x half> [[TMP3]])
-// CHECK: ret <4 x half> [[MUL]]
-float16x4_t test_vmulx_n_f16(float16x4_t a, float16_t b) {
-  return vmulx_n_f16(a, b);
-}
-
-// CHECK-LABEL: test_vmulxq_n_f16
-// CHECK: [[TMP0:%.*]] = insertelement <8 x half> undef, half %b, i32 0
-// CHECK: [[TMP1:%.*]] = insertelement <8 x half> [[TMP0]], half %b, i32 1
-// CHECK: [[TMP2:%.*]] = insertelement <8 x half> [[TMP1]], half %b, i32 2
-// CHECK: [[TMP3:%.*]] = insertelement <8 x half> [[TMP2]], half %b, i32 3
-// CHECK: [[TMP4:%.*]] = insertelement <8 x half> [[TMP3]], half %b, i32 4
-// CHECK: [[TMP5:%.*]] = insertelement <8 x half> [[TMP4]], half %b, i32 5
-// CHECK: [[TMP6:%.*]] = insertelement <8 x half> [[TMP5]], half %b, i32 6
-// CHECK: [[TMP7:%.*]] = insertelement <8 x half> [[TMP6]], half %b, i32 7
-// CHECK: [[MUL:%.*]]  = call <8 x half> @llvm.aarch64.neon.fmulx.v8f16(<8 x half> %a, <8 x half> [[TMP7]])
-// CHECK: ret <8 x half> [[MUL]]
-float16x8_t test_vmulxq_n_f16(float16x8_t a, float16_t b) {
-  return vmulxq_n_f16(a, b);
-}
-
-/* TODO: Not implemented yet (needs scalar intrinsic from arm_fp16.h)
-// CCHECK-LABEL: test_vmulxh_lane_f16
-// CCHECK: [[CONV0:%.*]] = fpext half %a to float
-// CCHECK: [[CONV1:%.*]] = fpext half %{{.*}} to float
-// CCHECK: [[MUL:%.*]]   = fmul float [[CONV0:%.*]], [[CONV0:%.*]]
-// CCHECK: [[CONV3:%.*]] = fptrunc float %mul to half
-// CCHECK: ret half [[CONV3:%.*]]
-float16_t test_vmulxh_lane_f16(float16_t a, float16x4_t b) {
-  return vmulxh_lane_f16(a, b, 3);
-}
-
-// CCHECK-LABEL: test_vmulxh_laneq_f16
-// CCHECK: [[CONV0:%.*]] = fpext half %a to float
-// CCHECK: [[CONV1:%.*]] = fpext half %{{.*}} to float
-// CCHECK: [[MUL:%.*]]   = fmul float [[CONV0:%.*]], [[CONV0:%.*]]
-// CCHECK: [[CONV3:%.*]] = fptrunc float %mul to half
-// CCHECK: ret half [[CONV3:%.*]]
-float16_t test_vmulxh_laneq_f16(float16_t a, float16x8_t b) {
-  return vmulxh_laneq_f16(a, b, 7);
-}
-*/
-
-// CHECK-LABEL: test_vmaxv_f16
-// CHECK: [[TMP0:%.*]] = bitcast <4 x half> %a to <8 x i8>
-// CHECK: [[TMP1:%.*]] = bitcast <8 x i8> [[TMP0]] to <4 x half>
-// CHECK: [[MAX:%.*]]  = call half @llvm.aarch64.neon.fmaxv.f16.v4f16(<4 x half> [[TMP1]])
-// CHECK: ret half [[MAX]]
-float16_t test_vmaxv_f16(float16x4_t a) {
-  return vmaxv_f16(a);
-}
-
-// CHECK-LABEL: test_vmaxvq_f16
-// CHECK: [[TMP0:%.*]] = bitcast <8 x half> %a to <16 x i8>
-// CHECK: [[TMP1:%.*]] = bitcast <16 x i8> [[TMP0]] to <8 x half>
-// CHECK: [[MAX:%.*]]  = call half @llvm.aarch64.neon.fmaxv.f16.v8f16(<8 x half> [[TMP1]])
-// CHECK: ret half [[MAX]]
-float16_t test_vmaxvq_f16(float16x8_t a) {
-  return vmaxvq_f16(a);
-}
-
-// CHECK-LABEL: test_vminv_f16
-// CHECK: [[TMP0:%.*]] = bitcast <4 x half> %a to <8 x i8>
-// CHECK: [[TMP1:%.*]] = bitcast <8 x i8> [[TMP0]] to <4 x half>
-// CHECK: [[MAX:%.*]]  = call half @llvm.aarch64.neon.fminv.f16.v4f16(<4 x half> [[TMP1]])
-// CHECK: ret half [[MAX]]
-float16_t test_vminv_f16(float16x4_t a) {
-  return vminv_f16(a);
-}
-
-// CHECK-LABEL: test_vminvq_f16
-// CHECK: [[TMP0:%.*]] = bitcast <8 x half> %a to <16 x i8>
-// CHECK: [[TMP1:%.*]] = bitcast <16 x i8> [[TMP0]] to <8 x half>
-// CHECK: [[MAX:%.*]]  = call half @llvm.aarch64.neon.fminv.f16.v8f16(<8 x half> [[TMP1]])
-// CHECK: ret half [[MAX]]
-float16_t test_vminvq_f16(float16x8_t a) {
-  return vminvq_f16(a);
-}
-
-// CHECK-LABEL: test_vmaxnmv_f16
-// CHECK: [[TMP0:%.*]] = bitcast <4 x half> %a to <8 x i8>
-// CHECK: [[TMP1:%.*]] = bitcast <8 x i8> [[TMP0]] to <4 x half>
-// CHECK: [[MAX:%.*]]  = call half @llvm.aarch64.neon.fmaxnmv.f16.v4f16(<4 x half> [[TMP1]])
-// CHECK: ret half [[MAX]]
-float16_t test_vmaxnmv_f16(float16x4_t a) {
-  return vmaxnmv_f16(a);
-}
-
-// CHECK-LABEL: test_vmaxnmvq_f16
-// CHECK: [[TMP0:%.*]] = bitcast <8 x half> %a to <16 x i8>
-// CHECK: [[TMP1:%.*]] = bitcast <16 x i8> [[TMP0]] to <8 x half>
-// CHECK: [[MAX:%.*]]  = call half @llvm.aarch64.neon.fmaxnmv.f16.v8f16(<8 x half> [[TMP1]])
-// CHECK: ret half [[MAX]]
-float16_t test_vmaxnmvq_f16(float16x8_t a) {
-  return vmaxnmvq_f16(a);
-}
-
-// CHECK-LABEL: test_vminnmv_f16
-// CHECK: [[TMP0:%.*]] = bitcast <4 x half> %a to <8 x i8>
-// CHECK: [[TMP1:%.*]] = bitcast <8 x i8> [[TMP0]] to <4 x half>
-// CHECK: [[MAX:%.*]]  = call half @llvm.aarch64.neon.fminnmv.f16.v4f16(<4 x half> [[TMP1]])
-// CHECK: ret half [[MAX]]
-float16_t test_vminnmv_f16(float16x4_t a) {
-  return vminnmv_f16(a);
-}
-
-// CHECK-LABEL: test_vminnmvq_f16
-// CHECK: [[TMP0:%.*]] = bitcast <8 x half> %a to <16 x i8>
-// CHECK: [[TMP1:%.*]] = bitcast <16 x i8> [[TMP0]] to <8 x half>
-// CHECK: [[MAX:%.*]]  = call half @llvm.aarch64.neon.fminnmv.f16.v8f16(<8 x half> [[TMP1]])
-// CHECK: ret half [[MAX]]
-float16_t test_vminnmvq_f16(float16x8_t a) {
-  return vminnmvq_f16(a);
-}
-
-// CHECK-LABEL: test_vbsl_f16
-// CHECK:  [[TMP0:%.*]] = bitcast <4 x half> %b to <8 x i8>
-// CHECK:  [[TMP1:%.*]] = bitcast <4 x half> %c to <8 x i8>
-// CHECK:  [[TMP2:%.*]] = bitcast <8 x i8> [[TMP0]] to <4 x i16>
-// CHECK:  [[TMP3:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x i16>
-// CHECK:  [[TMP4:%.*]] = and <4 x i16> %a, [[TMP2]]
-// CHECK:  [[TMP5:%.*]] = xor <4 x i16> %a, <i16 -1, i16 -1, i16 -1, i16 -1>
-// CHECK:  [[TMP6:%.*]] = and <4 x i16> [[TMP5]], [[TMP3]]
-// CHECK:  [[TMP7:%.*]] = or <4 x i16> [[TMP4]], [[TMP6]]
-// CHECK:  [[TMP8:%.*]] = bitcast <4 x i16> [[TMP7]] to <4 x half>
-// CHECK:  ret <4 x half> [[TMP8]]
-float16x4_t test_vbsl_f16(uint16x4_t a, float16x4_t b, float16x4_t c) {
-  return vbsl_f16(a, b, c);
-}
-
-// CHECK-LABEL: test_vbslq_f16
-// CHECK:  [[TMP0:%.*]] = bitcast <8 x half> %b to <16 x i8>
-// CHECK:  [[TMP1:%.*]] = bitcast <8 x half> %c to <16 x i8>
-// CHECK:  [[TMP2:%.*]] = bitcast <16 x i8> [[TMP0]] to <8 x i16>
-// CHECK:  [[TMP3:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x i16>
-// CHECK:  [[TMP4:%.*]] = and <8 x i16> %a, [[TMP2]]
-// CHECK:  [[TMP5:%.*]] = xor <8 x i16> %a, <i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1, i16 -1>
-// CHECK:  [[TMP6:%.*]] = and <8 x i16> [[TMP5]], [[TMP3]]
-// CHECK:  [[TMP7:%.*]] = or <8 x i16> [[TMP4]], [[TMP6]]
-// CHECK:  [[TMP8:%.*]] = bitcast <8 x i16> [[TMP7]] to <8 x half>
-// CHECK:  ret <8 x half> [[TMP8]]
-float16x8_t test_vbslq_f16(uint16x8_t a, float16x8_t b, float16x8_t c) {
-  return vbslq_f16(a, b, c);
-}
-
-// CHECK-LABEL: test_vzip_f16
-// CHECK:   [[RETVAL:%.*]]  = alloca %struct.float16x4x2_t, align 8
-// CHECK:   [[__RET_I:%.*]] = alloca %struct.float16x4x2_t, align 8
-// CHECK:   [[TMP0:%.*]]  = bitcast %struct.float16x4x2_t* [[__RET_I]] to i8*
-// CHECK:   [[TMP1:%.*]]  = bitcast i8* [[TMP0]] to <4 x half>*
-// CHECK:   [[VZIP0_I:%.*]] = shufflevector <4 x half> %a, <4 x half> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
-// CHECK:   store <4 x half> [[VZIP0_I]], <4 x half>* [[TMP1]]
-// CHECK:   [[TMP2:%.*]] = getelementptr inbounds <4 x half>, <4 x half>* [[TMP1]], i32 1
-// CHECK:   [[VZIP1_I:%.*]] = shufflevector <4 x half> %a, <4 x half> %b, <4 x i32> <i32 2, i32 6, i32 3, i32 7>
-// CHECK:   store <4 x half> [[VZIP1_I]], <4 x half>* [[TMP2]]
-// CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x4x2_t* [[RETVAL]] to i8*
-// CHECK:   [[TMP6:%.*]] = bitcast %struct.float16x4x2_t* [[__RET_I]] to i8*
-// CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP5]], i8* [[TMP6]], i64 16, i32 8, i1 false)
-float16x4x2_t test_vzip_f16(float16x4_t a, float16x4_t b) {
-  return vzip_f16(a, b);
-}
-
-// CHECK-LABEL: test_vzipq_f16
-// CHECK:   [[RETVAL:%.*]]  = alloca %struct.float16x8x2_t, align 16
-// CHECK:   [[__RET_I:%.*]] = alloca %struct.float16x8x2_t, align 16
-// CHECK:   [[TMP0:%.*]]  = bitcast %struct.float16x8x2_t* [[__RET_I]] to i8*
-// CHECK:   [[TMP1:%.*]]  = bitcast i8* [[TMP0]] to <8 x half>*
-// CHECK:   [[VZIP0_I:%.*]] = shufflevector <8 x half> %a, <8 x half> %b, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 2, i32 10, i32 3, i32 11>
-// CHECK:   store <8 x half> [[VZIP0_I]], <8 x half>* [[TMP1]]
-// CHECK:   [[TMP2:%.*]] = getelementptr inbounds <8 x half>, <8 x half>* [[TMP1]], i32 1
-// CHECK:   [[VZIP1_I:%.*]] = shufflevector <8 x half> %a, <8 x half> %b, <8 x i32> <i32 4, i32 12, i32 5, i32 13, i32 6, i32 14, i32 7, i32 15>
-// CHECK:   store <8 x half> [[VZIP1_I]], <8 x half>* [[TMP2]]
-// CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x8x2_t* [[RETVAL]] to i8*
-// CHECK:   [[TMP6:%.*]] = bitcast %struct.float16x8x2_t* [[__RET_I]] to i8*
-// CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP5]], i8* [[TMP6]], i64 32, i32 16, i1 false)
-float16x8x2_t test_vzipq_f16(float16x8_t a, float16x8_t b) {
-  return vzipq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vuzp_f16
-// CHECK:   [[RETVAL:%.*]]  = alloca %struct.float16x4x2_t, align 8
-// CHECK:   [[__RET_I:%.*]] = alloca %struct.float16x4x2_t, align 8
-// CHECK:   [[TMP0:%.*]]  = bitcast %struct.float16x4x2_t* [[__RET_I]] to i8*
-// CHECK:   [[TMP1:%.*]]  = bitcast i8* [[TMP0]] to <4 x half>*
-// CHECK:   [[VZIP0_I:%.*]] = shufflevector <4 x half> %a, <4 x half> %b, <4 x i32> <i32 0, i32 2, i32 4, i32 6>
-// CHECK:   store <4 x half> [[VZIP0_I]], <4 x half>* [[TMP1]]
-// CHECK:   [[TMP2:%.*]] = getelementptr inbounds <4 x half>, <4 x half>* [[TMP1]], i32 1
-// CHECK:   [[VZIP1_I:%.*]] = shufflevector <4 x half> %a, <4 x half> %b, <4 x i32> <i32 1, i32 3, i32 5, i32 7>
-// CHECK:   store <4 x half> [[VZIP1_I]], <4 x half>* [[TMP2]]
-// CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x4x2_t* [[RETVAL]] to i8*
-// CHECK:   [[TMP6:%.*]] = bitcast %struct.float16x4x2_t* [[__RET_I]] to i8*
-// CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP5]], i8* [[TMP6]], i64 16, i32 8, i1 false)
-float16x4x2_t test_vuzp_f16(float16x4_t a, float16x4_t b) {
-  return vuzp_f16(a, b);
-}
-
-// CHECK-LABEL: test_vuzpq_f16
-// CHECK:   [[RETVAL:%.*]]  = alloca %struct.float16x8x2_t, align 16
-// CHECK:   [[__RET_I:%.*]] = alloca %struct.float16x8x2_t, align 16
-// CHECK:   [[TMP0:%.*]]  = bitcast %struct.float16x8x2_t* [[__RET_I]] to i8*
-// CHECK:   [[TMP1:%.*]]  = bitcast i8* [[TMP0]] to <8 x half>*
-// CHECK:   [[VZIP0_I:%.*]] = shufflevector <8 x half> %a, <8 x half> %b, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14>
-// CHECK:   store <8 x half> [[VZIP0_I]], <8 x half>* [[TMP1]]
-// CHECK:   [[TMP2:%.*]] = getelementptr inbounds <8 x half>, <8 x half>* [[TMP1]], i32 1
-// CHECK:   [[VZIP1_I:%.*]] = shufflevector <8 x half> %a, <8 x half> %b, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15>
-// CHECK:   store <8 x half> [[VZIP1_I]], <8 x half>* [[TMP2]]
-// CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x8x2_t* [[RETVAL]] to i8*
-// CHECK:   [[TMP6:%.*]] = bitcast %struct.float16x8x2_t* [[__RET_I]] to i8*
-// CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP5]], i8* [[TMP6]], i64 32, i32 16, i1 false)
-float16x8x2_t test_vuzpq_f16(float16x8_t a, float16x8_t b) {
-  return vuzpq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vtrn_f16
-// CHECK:   [[RETVAL:%.*]]  = alloca %struct.float16x4x2_t, align 8
-// CHECK:   [[__RET_I:%.*]] = alloca %struct.float16x4x2_t, align 8
-// CHECK:   [[TMP0:%.*]]  = bitcast %struct.float16x4x2_t* [[__RET_I]] to i8*
-// CHECK:   [[TMP1:%.*]]  = bitcast i8* [[TMP0]] to <4 x half>*
-// CHECK:   [[VZIP0_I:%.*]] = shufflevector <4 x half> %a, <4 x half> %b, <4 x i32> <i32 0, i32 4, i32 2, i32 6>
-// CHECK:   store <4 x half> [[VZIP0_I]], <4 x half>* [[TMP1]]
-// CHECK:   [[TMP2:%.*]] = getelementptr inbounds <4 x half>, <4 x half>* [[TMP1]], i32 1
-// CHECK:   [[VZIP1_I:%.*]] = shufflevector <4 x half> %a, <4 x half> %b, <4 x i32> <i32 1, i32 5, i32 3, i32 7>
-// CHECK:   store <4 x half> [[VZIP1_I]], <4 x half>* [[TMP2]]
-// CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x4x2_t* [[RETVAL]] to i8*
-// CHECK:   [[TMP6:%.*]] = bitcast %struct.float16x4x2_t* [[__RET_I]] to i8*
-// CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP5]], i8* [[TMP6]], i64 16, i32 8, i1 false)
-float16x4x2_t test_vtrn_f16(float16x4_t a, float16x4_t b) {
-  return vtrn_f16(a, b);
-}
-
-// CHECK-LABEL: test_vtrnq_f16
-// CHECK:   [[RETVAL:%.*]]  = alloca %struct.float16x8x2_t, align 16
-// CHECK:   [[__RET_I:%.*]] = alloca %struct.float16x8x2_t, align 16
-// CHECK:   [[TMP0:%.*]]  = bitcast %struct.float16x8x2_t* [[__RET_I]] to i8*
-// CHECK:   [[TMP1:%.*]]  = bitcast i8* [[TMP0]] to <8 x half>*
-// CHECK:   [[VZIP0_I:%.*]] = shufflevector <8 x half> %a, <8 x half> %b, <8 x i32> <i32 0, i32 8, i32 2, i32 10, i32 4, i32 12, i32 6, i32 14>
-// CHECK:   store <8 x half> [[VZIP0_I]], <8 x half>* [[TMP1]] 
-// CHECK:   [[TMP2:%.*]] = getelementptr inbounds <8 x half>, <8 x half>* [[TMP1]], i32 1
-// CHECK:   [[VZIP1_I:%.*]] = shufflevector <8 x half> %a, <8 x half> %b, <8 x i32>  <i32 1, i32 9, i32 3, i32 11, i32 5, i32 13, i32 7, i32 15>
-// CHECK:   store <8 x half> [[VZIP1_I]], <8 x half>* [[TMP2]]
-// CHECK:   [[TMP5:%.*]] = bitcast %struct.float16x8x2_t* [[RETVAL]] to i8*
-// CHECK:   [[TMP6:%.*]] = bitcast %struct.float16x8x2_t* [[__RET_I]] to i8*
-// CHECK:   call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP5]], i8* [[TMP6]], i64 32, i32 16, i1 false)
-float16x8x2_t test_vtrnq_f16(float16x8_t a, float16x8_t b) {
-  return vtrnq_f16(a, b);
-}
-
-// CHECK-LABEL: test_vmov_n_f16
-// CHECK:   [[TMP0:%.*]] = insertelement <4 x half> undef, half %a, i32 0
-// CHECK:   [[TMP1:%.*]] = insertelement <4 x half> [[TMP0]], half %a, i32 1
-// CHECK:   [[TMP2:%.*]] = insertelement <4 x half> [[TMP1]], half %a, i32 2
-// CHECK:   [[TMP3:%.*]] = insertelement <4 x half> [[TMP2]], half %a, i32 3
-// CHECK:   ret <4 x half> [[TMP3]]
-float16x4_t test_vmov_n_f16(float16_t a) {
-  return vmov_n_f16(a);
-}
-
-// CHECK-LABEL: test_vmovq_n_f16
-// CHECK:   [[TMP0:%.*]] = insertelement <8 x half> undef, half %a, i32 0
-// CHECK:   [[TMP1:%.*]] = insertelement <8 x half> [[TMP0]], half %a, i32 1
-// CHECK:   [[TMP2:%.*]] = insertelement <8 x half> [[TMP1]], half %a, i32 2
-// CHECK:   [[TMP3:%.*]] = insertelement <8 x half> [[TMP2]], half %a, i32 3
-// CHECK:   [[TMP4:%.*]] = insertelement <8 x half> [[TMP3]], half %a, i32 4
-// CHECK:   [[TMP5:%.*]] = insertelement <8 x half> [[TMP4]], half %a, i32 5
-// CHECK:   [[TMP6:%.*]] = insertelement <8 x half> [[TMP5]], half %a, i32 6
-// CHECK:   [[TMP7:%.*]] = insertelement <8 x half> [[TMP6]], half %a, i32 7
-// CHECK:   ret <8 x half> [[TMP7]]
-float16x8_t test_vmovq_n_f16(float16_t a) {
-  return vmovq_n_f16(a);
-}
-
-// CHECK-LABEL: test_vdup_n_f16
-// CHECK:   [[TMP0:%.*]] = insertelement <4 x half> undef, half %a, i32 0
-// CHECK:   [[TMP1:%.*]] = insertelement <4 x half> [[TMP0]], half %a, i32 1
-// CHECK:   [[TMP2:%.*]] = insertelement <4 x half> [[TMP1]], half %a, i32 2
-// CHECK:   [[TMP3:%.*]] = insertelement <4 x half> [[TMP2]], half %a, i32 3
-// CHECK:   ret <4 x half> [[TMP3]]
-float16x4_t test_vdup_n_f16(float16_t a) {
-  return vdup_n_f16(a);
-}
-
-// CHECK-LABEL: test_vdupq_n_f16
-// CHECK:   [[TMP0:%.*]] = insertelement <8 x half> undef, half %a, i32 0
-// CHECK:   [[TMP1:%.*]] = insertelement <8 x half> [[TMP0]], half %a, i32 1
-// CHECK:   [[TMP2:%.*]] = insertelement <8 x half> [[TMP1]], half %a, i32 2
-// CHECK:   [[TMP3:%.*]] = insertelement <8 x half> [[TMP2]], half %a, i32 3
-// CHECK:   [[TMP4:%.*]] = insertelement <8 x half> [[TMP3]], half %a, i32 4
-// CHECK:   [[TMP5:%.*]] = insertelement <8 x half> [[TMP4]], half %a, i32 5
-// CHECK:   [[TMP6:%.*]] = insertelement <8 x half> [[TMP5]], half %a, i32 6
-// CHECK:   [[TMP7:%.*]] = insertelement <8 x half> [[TMP6]], half %a, i32 7
-// CHECK:   ret <8 x half> [[TMP7]]
-float16x8_t test_vdupq_n_f16(float16_t a) {
-  return vdupq_n_f16(a);
-}
-
-// CHECK-LABEL: test_vdup_lane_f16
-// CHECK:   [[SHFL:%.*]] = shufflevector <4 x half> %a, <4 x half> %a, <4 x i32> <i32 3, i32 3, i32 3, i32 3>
-// CHECK:   ret <4 x half> [[SHFL]]
-float16x4_t test_vdup_lane_f16(float16x4_t a) {
-  return vdup_lane_f16(a, 3);
-}
-
-// CHECK-LABEL: test_vdupq_lane_f16
-// CHECK:   [[SHFL:%.*]] = shufflevector <4 x half> %a, <4 x half> %a, <8 x i32> <i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7, i32 7>
-// CHECK:   ret <8 x half> [[SHFL]]
-float16x8_t test_vdupq_lane_f16(float16x4_t a) {
-  return vdupq_lane_f16(a, 7);
-}
-
-// CHECK-LABEL: @test_vext_f16(
-// CHECK:   [[TMP0:%.*]] = bitcast <4 x half> %a to <8 x i8>
-// CHECK:   [[TMP1:%.*]] = bitcast <4 x half> %b to <8 x i8>
-// CHECK:   [[TMP2:%.*]] = bitcast <8 x i8> [[TMP0]] to <4 x half>
-// CHECK:   [[TMP3:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x half>
-// CHECK:   [[VEXT:%.*]] = shufflevector <4 x half> [[TMP2]], <4 x half> [[TMP3]], <4 x i32> <i32 2, i32 3, i32 4, i32 5>
-// CHECK:   ret <4 x half> [[VEXT]]
-float16x4_t test_vext_f16(float16x4_t a, float16x4_t b) {
-  return vext_f16(a, b, 2);
-}
-
-// CHECK-LABEL: @test_vextq_f16(
-// CHECK:   [[TMP0:%.*]] = bitcast <8 x half> %a to <16 x i8>
-// CHECK:   [[TMP1:%.*]] = bitcast <8 x half> %b to <16 x i8>
-// CHECK:   [[TMP2:%.*]] = bitcast <16 x i8> [[TMP0]] to <8 x half>
-// CHECK:   [[TMP3:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x half>
-// CHECK:   [[VEXT:%.*]] = shufflevector <8 x half> [[TMP2]], <8 x half> [[TMP3]], <8 x i32> <i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12>
-// CHECK:   ret <8 x half> [[VEXT]]
-float16x8_t test_vextq_f16(float16x8_t a, float16x8_t b) {
-  return vextq_f16(a, b, 5);
-}
-
-// CHECK-LABEL: @test_vrev64_f16(
-// CHECK:   [[SHFL:%.*]] = shufflevector <4 x half> %a, <4 x half> %a, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
-// CHECK:   ret <4 x half> [[SHFL]]
-float16x4_t test_vrev64_f16(float16x4_t a) {
-  return vrev64_f16(a);
-}
-
-// CHECK-LABEL: @test_vrev64q_f16(
-// CHECK:   [[SHFL:%.*]] = shufflevector <8 x half> %a, <8 x half> %a, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4>
-// CHECK:   ret <8 x half> [[SHFL]]
-float16x8_t test_vrev64q_f16(float16x8_t a) {
-  return vrev64q_f16(a);
-}
-
-// CHECK-LABEL: @test_vzip1_f16(
-// CHECK:   [[SHFL:%.*]] = shufflevector <4 x half> %a, <4 x half> %b, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
-// CHECK:   ret <4 x half> [[SHFL]]
-float16x4_t test_vzip1_f16(float16x4_t a, float16x4_t b) {
-  return vzip1_f16(a, b);
-}
-
-// CHECK-LABEL: @test_vzip1q_f16(
-// CHECK:   [[SHFL:%.*]] = shufflevector <8 x half> %a, <8 x half> %b, <8 x i32> <i32 0, i32 8, i32 1, i32 9, i32 2, i32 10, i32 3, i32 11>
-// CHECK:   ret <8 x half> [[SHFL]]
-float16x8_t test_vzip1q_f16(float16x8_t a, float16x8_t b) {
-  return vzip1q_f16(a, b);
-}
-
-// CHECK-LABEL: @test_vzip2_f16(
-// CHECK:   [[SHFL:%.*]] = shufflevector <4 x half> %a, <4 x half> %b, <4 x i32> <i32 2, i32 6, i32 3, i32 7>
-// CHECK:   ret <4 x half> [[SHFL]]
-float16x4_t test_vzip2_f16(float16x4_t a, float16x4_t b) {
-  return vzip2_f16(a, b);
-}
-
-// CHECK-LABEL: @test_vzip2q_f16(
-// CHECK:   [[SHFL:%.*]] = shufflevector <8 x half> %a, <8 x half> %b, <8 x i32> <i32 4, i32 12, i32 5, i32 13, i32 6, i32 14, i32 7, i32 15>
-// CHECK:   ret <8 x half> [[SHFL]]
-float16x8_t test_vzip2q_f16(float16x8_t a, float16x8_t b) {
-  return vzip2q_f16(a, b);
-}
-
-// CHECK-LABEL: @test_vuzp1_f16(
-// CHECK:   [[SHUFFLE_I:%.*]] = shufflevector <4 x half> %a, <4 x half> %b, <4 x i32> <i32 0, i32 2, i32 4, i32 6>
-// CHECK:   ret <4 x half> [[SHUFFLE_I]]
-float16x4_t test_vuzp1_f16(float16x4_t a, float16x4_t b) {
-  return vuzp1_f16(a, b);
-}
-
-// CHECK-LABEL: @test_vuzp1q_f16(
-// CHECK:   [[SHUFFLE_I:%.*]] = shufflevector <8 x half> %a, <8 x half> %b, <8 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 10, i32 12, i32 14>
-// CHECK:   ret <8 x half> [[SHUFFLE_I]]
-float16x8_t test_vuzp1q_f16(float16x8_t a, float16x8_t b) {
-  return vuzp1q_f16(a, b);
-}
-
-// CHECK-LABEL: @test_vuzp2_f16(
-// CHECK:   [[SHUFFLE_I:%.*]] = shufflevector <4 x half> %a, <4 x half> %b, <4 x i32> <i32 1, i32 3, i32 5, i32 7>
-// CHECK:   ret <4 x half> [[SHUFFLE_I]]
-float16x4_t test_vuzp2_f16(float16x4_t a, float16x4_t b) {
-  return vuzp2_f16(a, b);
-}
-
-// CHECK-LABEL: @test_vuzp2q_f16(
-// CHECK:   [[SHUFFLE_I:%.*]] = shufflevector <8 x half> %a, <8 x half> %b, <8 x i32> <i32 1, i32 3, i32 5, i32 7, i32 9, i32 11, i32 13, i32 15>
-// CHECK:   ret <8 x half> [[SHUFFLE_I]]
-float16x8_t test_vuzp2q_f16(float16x8_t a, float16x8_t b) {
-  return vuzp2q_f16(a, b);
-}
-
-// CHECK-LABEL: @test_vtrn1_f16(
-// CHECK:   [[SHUFFLE_I:%.*]] = shufflevector <4 x half> %a, <4 x half> %b, <4 x i32> <i32 0, i32 4, i32 2, i32 6>
-// CHECK:   ret <4 x half> [[SHUFFLE_I]]
-float16x4_t test_vtrn1_f16(float16x4_t a, float16x4_t b) {
-  return vtrn1_f16(a, b);
-}
-
-// CHECK-LABEL: @test_vtrn1q_f16(
-// CHECK:   [[SHUFFLE_I:%.*]] = shufflevector <8 x half> %a, <8 x half> %b, <8 x i32>  <i32 0, i32 8, i32 2, i32 10, i32 4, i32 12, i32 6, i32 14>
-// CHECK:   ret <8 x half> [[SHUFFLE_I]]
-float16x8_t test_vtrn1q_f16(float16x8_t a, float16x8_t b) {
-  return vtrn1q_f16(a, b);
-}
-
-// CHECK-LABEL: @test_vtrn2_f16(
-// CHECK:   [[SHUFFLE_I:%.*]] = shufflevector <4 x half> %a, <4 x half> %b, <4 x i32> <i32 1, i32 5, i32 3, i32 7>
-// CHECK:   ret <4 x half> [[SHUFFLE_I]]
-float16x4_t test_vtrn2_f16(float16x4_t a, float16x4_t b) {
-  return vtrn2_f16(a, b);
-}
-
-// CHECK-LABEL: @test_vtrn2q_f16(
-// CHECK:   [[SHUFFLE_I:%.*]] = shufflevector <8 x half> %a, <8 x half> %b, <8 x i32> <i32 1, i32 9, i32 3, i32 11, i32 5, i32 13, i32 7, i32 15>
-// CHECK:   ret <8 x half> [[SHUFFLE_I]]
-float16x8_t test_vtrn2q_f16(float16x8_t a, float16x8_t b) {
-  return vtrn2q_f16(a, b);
-}
-
diff --git a/test/CodeGen/address-space.c b/test/CodeGen/address-space.c
index 54e059385772..28b3954ab7d2 100644
--- a/test/CodeGen/address-space.c
+++ b/test/CodeGen/address-space.c
@@ -24,11 +24,13 @@ int test2(int i) { return ban[i]; }
 __attribute__((address_space(2))) int *A, *B;
 
 // CHECK-LABEL: define void @test3()
-// GIZ: load i32 addrspace(2)*, i32 addrspace(2)** @B
-// PIZ: load i32 addrspace(2)*, i32 addrspace(2)* addrspace(4)* @B
+// X86: load i32 addrspace(2)*, i32 addrspace(2)** @B
+// AMDGIZ: load i32 addrspace(2)*, i32 addrspace(2)** addrspacecast (i32 addrspace(2)* addrspace(1)* @B to i32 addrspace(2)**)
+// PIZ: load i32 addrspace(2)*, i32 addrspace(2)* addrspace(4)* addrspacecast (i32 addrspace(2)* addrspace(1)* @B to i32 addrspace(2)* addrspace(4)*)
 // CHECK: load i32, i32 addrspace(2)*
-// GIZ: load i32 addrspace(2)*, i32 addrspace(2)** @A
-// PIZ: load i32 addrspace(2)*, i32 addrspace(2)* addrspace(4)* @A
+// X86: load i32 addrspace(2)*, i32 addrspace(2)** @A
+// AMDGIZ: load i32 addrspace(2)*, i32 addrspace(2)** addrspacecast (i32 addrspace(2)* addrspace(1)* @A to i32 addrspace(2)**)
+// PIZ: load i32 addrspace(2)*, i32 addrspace(2)* addrspace(4)* addrspacecast (i32 addrspace(2)* addrspace(1)* @A to i32 addrspace(2)* addrspace(4)*)
 // CHECK: store i32 {{.*}}, i32 addrspace(2)*
 void test3() {
   *A = *B;
diff --git a/test/CodeGen/arm_neon_intrinsics.c b/test/CodeGen/arm_neon_intrinsics.c
index b01c90c03a96..62888dd73339 100644
--- a/test/CodeGen/arm_neon_intrinsics.c
+++ b/test/CodeGen/arm_neon_intrinsics.c
@@ -3896,8 +3896,9 @@ int64x2_t test_vld1q_s64(int64_t const * a) {
 
 // CHECK-LABEL: @test_vld1q_f16(
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[VLD1:%.*]] = call <8 x half> @llvm.arm.neon.vld1.v8f16.p0i8(i8* [[TMP0]], i32 2)
-// CHECK:   ret <8 x half> [[VLD1]]
+// CHECK:   [[VLD1:%.*]] = call <8 x i16> @llvm.arm.neon.vld1.v8i16.p0i8(i8* [[TMP0]], i32 2)
+// CHECK:   [[TMP1:%.*]] = bitcast <8 x i16> [[VLD1]] to <8 x half>
+// CHECK:   ret <8 x half> [[TMP1]]
 float16x8_t test_vld1q_f16(float16_t const * a) {
   return vld1q_f16(a);
 }
@@ -3989,8 +3990,9 @@ int64x1_t test_vld1_s64(int64_t const * a) {
 
 // CHECK-LABEL: @test_vld1_f16(
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[VLD1:%.*]] = call <4 x half> @llvm.arm.neon.vld1.v4f16.p0i8(i8* [[TMP0]], i32 2)
-// CHECK:   ret <4 x half> [[VLD1]]
+// CHECK:   [[VLD1:%.*]] = call <4 x i16> @llvm.arm.neon.vld1.v4i16.p0i8(i8* [[TMP0]], i32 2)
+// CHECK:   [[TMP1:%.*]] = bitcast <4 x i16> [[VLD1]] to <4 x half>
+// CHECK:   ret <4 x half> [[TMP1]]
 float16x4_t test_vld1_f16(float16_t const * a) {
   return vld1_f16(a);
 }
@@ -4104,11 +4106,12 @@ int64x2_t test_vld1q_dup_s64(int64_t const * a) {
 
 // CHECK-LABEL: @test_vld1q_dup_f16(
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP1:%.*]] = bitcast i8* [[TMP0]] to half*
-// CHECK:   [[TMP2:%.*]] = load half, half* [[TMP1]], align 2
-// CHECK:   [[TMP3:%.*]] = insertelement <8 x half> undef, half [[TMP2]], i32 0
-// CHECK:   [[LANE:%.*]] = shufflevector <8 x half> [[TMP3]], <8 x half> [[TMP3]], <8 x i32> zeroinitializer
-// CHECK:   ret <8 x half> [[LANE]]
+// CHECK:   [[TMP1:%.*]] = bitcast i8* [[TMP0]] to i16*
+// CHECK:   [[TMP2:%.*]] = load i16, i16* [[TMP1]], align 2
+// CHECK:   [[TMP3:%.*]] = insertelement <8 x i16> undef, i16 [[TMP2]], i32 0
+// CHECK:   [[LANE:%.*]] = shufflevector <8 x i16> [[TMP3]], <8 x i16> [[TMP3]], <8 x i32> zeroinitializer
+// CHECK:   [[TMP4:%.*]] = bitcast <8 x i16> [[LANE]] to <8 x half>
+// CHECK:   ret <8 x half> [[TMP4]]
 float16x8_t test_vld1q_dup_f16(float16_t const * a) {
   return vld1q_dup_f16(a);
 }
@@ -4230,11 +4233,12 @@ int64x1_t test_vld1_dup_s64(int64_t const * a) {
 
 // CHECK-LABEL: @test_vld1_dup_f16(
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[TMP1:%.*]] = bitcast i8* [[TMP0]] to half*
-// CHECK:   [[TMP2:%.*]] = load half, half* [[TMP1]], align 2
-// CHECK:   [[TMP3:%.*]] = insertelement <4 x half> undef, half [[TMP2]], i32 0
-// CHECK:   [[LANE:%.*]] = shufflevector <4 x half> [[TMP3]], <4 x half> [[TMP3]], <4 x i32> zeroinitializer
-// CHECK:   ret <4 x half> [[LANE]]
+// CHECK:   [[TMP1:%.*]] = bitcast i8* [[TMP0]] to i16*
+// CHECK:   [[TMP2:%.*]] = load i16, i16* [[TMP1]], align 2
+// CHECK:   [[TMP3:%.*]] = insertelement <4 x i16> undef, i16 [[TMP2]], i32 0
+// CHECK:   [[LANE:%.*]] = shufflevector <4 x i16> [[TMP3]], <4 x i16> [[TMP3]], <4 x i32> zeroinitializer
+// CHECK:   [[TMP4:%.*]] = bitcast <4 x i16> [[LANE]] to <4 x half>
+// CHECK:   ret <4 x half> [[TMP4]]
 float16x4_t test_vld1_dup_f16(float16_t const * a) {
   return vld1_dup_f16(a);
 }
@@ -4361,11 +4365,12 @@ int64x2_t test_vld1q_lane_s64(int64_t const * a, int64x2_t b) {
 // CHECK-LABEL: @test_vld1q_lane_f16(
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast <8 x half> %b to <16 x i8>
-// CHECK:   [[TMP2:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x half>
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to half*
-// CHECK:   [[TMP4:%.*]] = load half, half* [[TMP3]], align 2
-// CHECK:   [[VLD1_LANE:%.*]] = insertelement <8 x half> [[TMP2]], half [[TMP4]], i32 7
-// CHECK:   ret <8 x half> [[VLD1_LANE]]
+// CHECK:   [[TMP2:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x i16>
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to i16*
+// CHECK:   [[TMP4:%.*]] = load i16, i16* [[TMP3]], align 2
+// CHECK:   [[VLD1_LANE:%.*]] = insertelement <8 x i16> [[TMP2]], i16 [[TMP4]], i32 7
+// CHECK:   [[TMP5:%.*]] = bitcast <8 x i16> [[VLD1_LANE]] to <8 x half>
+// CHECK:   ret <8 x half> [[TMP5]]
 float16x8_t test_vld1q_lane_f16(float16_t const * a, float16x8_t b) {
   return vld1q_lane_f16(a, b, 7);
 }
@@ -4493,11 +4498,12 @@ int64x1_t test_vld1_lane_s64(int64_t const * a, int64x1_t b) {
 // CHECK-LABEL: @test_vld1_lane_f16(
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast <4 x half> %b to <8 x i8>
-// CHECK:   [[TMP2:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x half>
-// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to half*
-// CHECK:   [[TMP4:%.*]] = load half, half* [[TMP3]], align 2
-// CHECK:   [[VLD1_LANE:%.*]] = insertelement <4 x half> [[TMP2]], half [[TMP4]], i32 3
-// CHECK:   ret <4 x half> [[VLD1_LANE]]
+// CHECK:   [[TMP2:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x i16>
+// CHECK:   [[TMP3:%.*]] = bitcast i8* [[TMP0]] to i16*
+// CHECK:   [[TMP4:%.*]] = load i16, i16* [[TMP3]], align 2
+// CHECK:   [[VLD1_LANE:%.*]] = insertelement <4 x i16> [[TMP2]], i16 [[TMP4]], i32 3
+// CHECK:   [[TMP5:%.*]] = bitcast <4 x i16> [[VLD1_LANE]] to <4 x half>
+// CHECK:   ret <4 x half> [[TMP5]]
 float16x4_t test_vld1_lane_f16(float16_t const * a, float16x4_t b) {
   return vld1_lane_f16(a, b, 3);
 }
@@ -4590,7 +4596,7 @@ int32x4x2_t test_vld2q_s32(int32_t const * a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x8x2_t, align 16
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x8x2_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[VLD2Q_V:%.*]] = call { <8 x half>, <8 x half>
+// CHECK:   [[VLD2Q_V:%.*]] = call { <8 x i16>, <8 x i16>
 float16x8x2_t test_vld2q_f16(float16_t const * a) {
   return vld2q_f16(a);
 }
@@ -4695,7 +4701,7 @@ int64x1x2_t test_vld2_s64(int64_t const * a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x4x2_t, align 8
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x4x2_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[VLD2_V:%.*]] = call { <4 x half>, <4 x half>
+// CHECK:   [[VLD2_V:%.*]] = call { <4 x i16>, <4 x i16>
 float16x4x2_t test_vld2_f16(float16_t const * a) {
   return vld2_f16(a);
 }
@@ -4800,7 +4806,7 @@ int64x1x2_t test_vld2_dup_s64(int64_t const * a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x4x2_t, align 8
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x4x2_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[VLD_DUP:%.*]] = call { <4 x half>, <4 x half>
+// CHECK:   [[VLD_DUP:%.*]] = call { <4 x i16>, <4 x i16>
 float16x4x2_t test_vld2_dup_f16(float16_t const * a) {
   return vld2_dup_f16(a);
 }
@@ -4959,9 +4965,9 @@ int32x4x2_t test_vld2q_lane_s32(int32_t const * a, int32x4x2_t b) {
 // CHECK:   [[ARRAYIDX2:%.*]] = getelementptr inbounds [2 x <8 x half>], [2 x <8 x half>]* [[VAL1]], i32 0, i32 1
 // CHECK:   [[TMP7:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX2]], align 16
 // CHECK:   [[TMP8:%.*]] = bitcast <8 x half> [[TMP7]] to <16 x i8>
-// CHECK:   [[TMP9:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x half>
-// CHECK:   [[TMP10:%.*]] = bitcast <16 x i8> [[TMP8]] to <8 x half>
-// CHECK:   [[VLD2Q_LANE_V:%.*]] = call { <8 x half>, <8 x half>
+// CHECK:   [[TMP9:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x i16>
+// CHECK:   [[TMP10:%.*]] = bitcast <16 x i8> [[TMP8]] to <8 x i16>
+// CHECK:   [[VLD2Q_LANE_V:%.*]] = call { <8 x i16>, <8 x i16>
 float16x8x2_t test_vld2q_lane_f16(float16_t const * a, float16x8x2_t b) {
   return vld2q_lane_f16(a, b, 7);
 }
@@ -5192,9 +5198,9 @@ int32x2x2_t test_vld2_lane_s32(int32_t const * a, int32x2x2_t b) {
 // CHECK:   [[ARRAYIDX2:%.*]] = getelementptr inbounds [2 x <4 x half>], [2 x <4 x half>]* [[VAL1]], i32 0, i32 1
 // CHECK:   [[TMP7:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX2]], align 8
 // CHECK:   [[TMP8:%.*]] = bitcast <4 x half> [[TMP7]] to <8 x i8>
-// CHECK:   [[TMP9:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x half>
-// CHECK:   [[TMP10:%.*]] = bitcast <8 x i8> [[TMP8]] to <4 x half>
-// CHECK:   [[VLD2_LANE_V:%.*]] = call { <4 x half>, <4 x half>
+// CHECK:   [[TMP9:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x i16>
+// CHECK:   [[TMP10:%.*]] = bitcast <8 x i8> [[TMP8]] to <4 x i16>
+// CHECK:   [[VLD2_LANE_V:%.*]] = call { <4 x i16>, <4 x i16>
 float16x4x2_t test_vld2_lane_f16(float16_t const * a, float16x4x2_t b) {
   return vld2_lane_f16(a, b, 3);
 }
@@ -5331,7 +5337,7 @@ int32x4x3_t test_vld3q_s32(int32_t const * a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x8x3_t, align 16
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x8x3_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[VLD3Q_V:%.*]] = call { <8 x half>, <8 x half>, <8 x half>
+// CHECK:   [[VLD3Q_V:%.*]] = call { <8 x i16>, <8 x i16>, <8 x i16>
 float16x8x3_t test_vld3q_f16(float16_t const * a) {
   return vld3q_f16(a);
 }
@@ -5436,7 +5442,7 @@ int64x1x3_t test_vld3_s64(int64_t const * a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x4x3_t, align 8
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x4x3_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[VLD3_V:%.*]] = call { <4 x half>, <4 x half>, <4 x half>
+// CHECK:   [[VLD3_V:%.*]] = call { <4 x i16>, <4 x i16>, <4 x i16>
 float16x4x3_t test_vld3_f16(float16_t const * a) {
   return vld3_f16(a);
 }
@@ -5541,7 +5547,7 @@ int64x1x3_t test_vld3_dup_s64(int64_t const * a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x4x3_t, align 8
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x4x3_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[VLD_DUP:%.*]] = call { <4 x half>, <4 x half>, <4 x half>
+// CHECK:   [[VLD_DUP:%.*]] = call { <4 x i16>, <4 x i16>, <4 x i16>
 float16x4x3_t test_vld3_dup_f16(float16_t const * a) {
   return vld3_dup_f16(a);
 }
@@ -5724,10 +5730,10 @@ int32x4x3_t test_vld3q_lane_s32(int32_t const * a, int32x4x3_t b) {
 // CHECK:   [[ARRAYIDX4:%.*]] = getelementptr inbounds [3 x <8 x half>], [3 x <8 x half>]* [[VAL3]], i32 0, i32 2
 // CHECK:   [[TMP9:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX4]], align 16
 // CHECK:   [[TMP10:%.*]] = bitcast <8 x half> [[TMP9]] to <16 x i8>
-// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x half>
-// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP8]] to <8 x half>
-// CHECK:   [[TMP13:%.*]] = bitcast <16 x i8> [[TMP10]] to <8 x half>
-// CHECK:   [[VLD3Q_LANE_V:%.*]] = call { <8 x half>, <8 x half>, <8 x half>
+// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x i16>
+// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP8]] to <8 x i16>
+// CHECK:   [[TMP13:%.*]] = bitcast <16 x i8> [[TMP10]] to <8 x i16>
+// CHECK:   [[VLD3Q_LANE_V:%.*]] = call { <8 x i16>, <8 x i16>, <8 x i16>
 float16x8x3_t test_vld3q_lane_f16(float16_t const * a, float16x8x3_t b) {
   return vld3q_lane_f16(a, b, 7);
 }
@@ -5998,10 +6004,10 @@ int32x2x3_t test_vld3_lane_s32(int32_t const * a, int32x2x3_t b) {
 // CHECK:   [[ARRAYIDX4:%.*]] = getelementptr inbounds [3 x <4 x half>], [3 x <4 x half>]* [[VAL3]], i32 0, i32 2
 // CHECK:   [[TMP9:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX4]], align 8
 // CHECK:   [[TMP10:%.*]] = bitcast <4 x half> [[TMP9]] to <8 x i8>
-// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x half>
-// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP8]] to <4 x half>
-// CHECK:   [[TMP13:%.*]] = bitcast <8 x i8> [[TMP10]] to <4 x half>
-// CHECK:   [[VLD3_LANE_V:%.*]] = call { <4 x half>, <4 x half>, <4 x half>
+// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x i16>
+// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP8]] to <4 x i16>
+// CHECK:   [[TMP13:%.*]] = bitcast <8 x i8> [[TMP10]] to <4 x i16>
+// CHECK:   [[VLD3_LANE_V:%.*]] = call { <4 x i16>, <4 x i16>, <4 x i16>
 float16x4x3_t test_vld3_lane_f16(float16_t const * a, float16x4x3_t b) {
   return vld3_lane_f16(a, b, 3);
 }
@@ -6151,7 +6157,7 @@ int32x4x4_t test_vld4q_s32(int32_t const * a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x8x4_t, align 16
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x8x4_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[VLD4Q_V:%.*]] = call { <8 x half>, <8 x half>, <8 x half>, <8 x half>
+// CHECK:   [[VLD4Q_V:%.*]] = call { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16>
 float16x8x4_t test_vld4q_f16(float16_t const * a) {
   return vld4q_f16(a);
 }
@@ -6256,7 +6262,7 @@ int64x1x4_t test_vld4_s64(int64_t const * a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x4x4_t, align 8
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x4x4_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[VLD4_V:%.*]] = call { <4 x half>, <4 x half>, <4 x half>, <4 x half>
+// CHECK:   [[VLD4_V:%.*]] = call { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16>
 float16x4x4_t test_vld4_f16(float16_t const * a) {
   return vld4_f16(a);
 }
@@ -6361,7 +6367,7 @@ int64x1x4_t test_vld4_dup_s64(int64_t const * a) {
 // CHECK:   [[__RET:%.*]] = alloca %struct.float16x4x4_t, align 8
 // CHECK:   [[TMP0:%.*]] = bitcast %struct.float16x4x4_t* [[__RET]] to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast half* %a to i8*
-// CHECK:   [[VLD_DUP:%.*]] = call { <4 x half>, <4 x half>, <4 x half>, <4 x half>
+// CHECK:   [[VLD_DUP:%.*]] = call { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16>
 float16x4x4_t test_vld4_dup_f16(float16_t const * a) {
   return vld4_dup_f16(a);
 }
@@ -6568,11 +6574,11 @@ int32x4x4_t test_vld4q_lane_s32(int32_t const * a, int32x4x4_t b) {
 // CHECK:   [[ARRAYIDX6:%.*]] = getelementptr inbounds [4 x <8 x half>], [4 x <8 x half>]* [[VAL5]], i32 0, i32 3
 // CHECK:   [[TMP11:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX6]], align 16
 // CHECK:   [[TMP12:%.*]] = bitcast <8 x half> [[TMP11]] to <16 x i8>
-// CHECK:   [[TMP13:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x half>
-// CHECK:   [[TMP14:%.*]] = bitcast <16 x i8> [[TMP8]] to <8 x half>
-// CHECK:   [[TMP15:%.*]] = bitcast <16 x i8> [[TMP10]] to <8 x half>
-// CHECK:   [[TMP16:%.*]] = bitcast <16 x i8> [[TMP12]] to <8 x half>
-// CHECK:   [[VLD4Q_LANE_V:%.*]] = call { <8 x half>, <8 x half>, <8 x half>, <8 x half>
+// CHECK:   [[TMP13:%.*]] = bitcast <16 x i8> [[TMP6]] to <8 x i16>
+// CHECK:   [[TMP14:%.*]] = bitcast <16 x i8> [[TMP8]] to <8 x i16>
+// CHECK:   [[TMP15:%.*]] = bitcast <16 x i8> [[TMP10]] to <8 x i16>
+// CHECK:   [[TMP16:%.*]] = bitcast <16 x i8> [[TMP12]] to <8 x i16>
+// CHECK:   [[VLD4Q_LANE_V:%.*]] = call { <8 x i16>, <8 x i16>, <8 x i16>, <8 x i16>
 float16x8x4_t test_vld4q_lane_f16(float16_t const * a, float16x8x4_t b) {
   return vld4q_lane_f16(a, b, 7);
 }
@@ -6883,11 +6889,11 @@ int32x2x4_t test_vld4_lane_s32(int32_t const * a, int32x2x4_t b) {
 // CHECK:   [[ARRAYIDX6:%.*]] = getelementptr inbounds [4 x <4 x half>], [4 x <4 x half>]* [[VAL5]], i32 0, i32 3
 // CHECK:   [[TMP11:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX6]], align 8
 // CHECK:   [[TMP12:%.*]] = bitcast <4 x half> [[TMP11]] to <8 x i8>
-// CHECK:   [[TMP13:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x half>
-// CHECK:   [[TMP14:%.*]] = bitcast <8 x i8> [[TMP8]] to <4 x half>
-// CHECK:   [[TMP15:%.*]] = bitcast <8 x i8> [[TMP10]] to <4 x half>
-// CHECK:   [[TMP16:%.*]] = bitcast <8 x i8> [[TMP12]] to <4 x half>
-// CHECK:   [[VLD4_LANE_V:%.*]] = call { <4 x half>, <4 x half>, <4 x half>, <4 x half>
+// CHECK:   [[TMP13:%.*]] = bitcast <8 x i8> [[TMP6]] to <4 x i16>
+// CHECK:   [[TMP14:%.*]] = bitcast <8 x i8> [[TMP8]] to <4 x i16>
+// CHECK:   [[TMP15:%.*]] = bitcast <8 x i8> [[TMP10]] to <4 x i16>
+// CHECK:   [[TMP16:%.*]] = bitcast <8 x i8> [[TMP12]] to <4 x i16>
+// CHECK:   [[VLD4_LANE_V:%.*]] = call { <4 x i16>, <4 x i16>, <4 x i16>, <4 x i16>
 float16x4x4_t test_vld4_lane_f16(float16_t const * a, float16x4x4_t b) {
   return vld4_lane_f16(a, b, 3);
 }
@@ -15778,8 +15784,8 @@ void test_vst1q_s64(int64_t * a, int64x2_t b) {
 // CHECK-LABEL: @test_vst1q_f16(
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast <8 x half> %b to <16 x i8>
-// CHECK:   [[TMP2:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x half>
-// CHECK:   call void @llvm.arm.neon.vst1.p0i8.v8f16(i8* [[TMP0]], <8 x half> [[TMP2]], i32 2)
+// CHECK:   [[TMP2:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x i16>
+// CHECK:   call void @llvm.arm.neon.vst1.p0i8.v8i16(i8* [[TMP0]], <8 x i16> [[TMP2]], i32 2)
 // CHECK:   ret void
 void test_vst1q_f16(float16_t * a, float16x8_t b) {
   vst1q_f16(a, b);
@@ -15889,8 +15895,8 @@ void test_vst1_s64(int64_t * a, int64x1_t b) {
 // CHECK-LABEL: @test_vst1_f16(
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast <4 x half> %b to <8 x i8>
-// CHECK:   [[TMP2:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x half>
-// CHECK:   call void @llvm.arm.neon.vst1.p0i8.v4f16(i8* [[TMP0]], <4 x half> [[TMP2]], i32 2)
+// CHECK:   [[TMP2:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x i16>
+// CHECK:   call void @llvm.arm.neon.vst1.p0i8.v4i16(i8* [[TMP0]], <4 x i16> [[TMP2]], i32 2)
 // CHECK:   ret void
 void test_vst1_f16(float16_t * a, float16x4_t b) {
   vst1_f16(a, b);
@@ -16012,10 +16018,10 @@ void test_vst1q_lane_s64(int64_t * a, int64x2_t b) {
 // CHECK-LABEL: @test_vst1q_lane_f16(
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast <8 x half> %b to <16 x i8>
-// CHECK:   [[TMP2:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x half>
-// CHECK:   [[TMP3:%.*]] = extractelement <8 x half> [[TMP2]], i32 7
-// CHECK:   [[TMP4:%.*]] = bitcast i8* [[TMP0]] to half*
-// CHECK:   store half [[TMP3]], half* [[TMP4]], align 2
+// CHECK:   [[TMP2:%.*]] = bitcast <16 x i8> [[TMP1]] to <8 x i16>
+// CHECK:   [[TMP3:%.*]] = extractelement <8 x i16> [[TMP2]], i32 7
+// CHECK:   [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i16*
+// CHECK:   store i16 [[TMP3]], i16* [[TMP4]], align 2
 // CHECK:   ret void
 void test_vst1q_lane_f16(float16_t * a, float16x8_t b) {
   vst1q_lane_f16(a, b, 7);
@@ -16144,10 +16150,10 @@ void test_vst1_lane_s64(int64_t * a, int64x1_t b) {
 // CHECK-LABEL: @test_vst1_lane_f16(
 // CHECK:   [[TMP0:%.*]] = bitcast half* %a to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast <4 x half> %b to <8 x i8>
-// CHECK:   [[TMP2:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x half>
-// CHECK:   [[TMP3:%.*]] = extractelement <4 x half> [[TMP2]], i32 3
-// CHECK:   [[TMP4:%.*]] = bitcast i8* [[TMP0]] to half*
-// CHECK:   store half [[TMP3]], half* [[TMP4]], align 2
+// CHECK:   [[TMP2:%.*]] = bitcast <8 x i8> [[TMP1]] to <4 x i16>
+// CHECK:   [[TMP3:%.*]] = extractelement <4 x i16> [[TMP2]], i32 3
+// CHECK:   [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i16*
+// CHECK:   store i16 [[TMP3]], i16* [[TMP4]], align 2
 // CHECK:   ret void
 void test_vst1_lane_f16(float16_t * a, float16x4_t b) {
   vst1_lane_f16(a, b, 3);
@@ -16349,9 +16355,9 @@ void test_vst2q_s32(int32_t * a, int32x4x2_t b) {
 // CHECK:   [[ARRAYIDX2:%.*]] = getelementptr inbounds [2 x <8 x half>], [2 x <8 x half>]* [[VAL1]], i32 0, i32 1
 // CHECK:   [[TMP6:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX2]], align 16
 // CHECK:   [[TMP7:%.*]] = bitcast <8 x half> [[TMP6]] to <16 x i8>
-// CHECK:   [[TMP8:%.*]] = bitcast <16 x i8> [[TMP5]] to <8 x half>
-// CHECK:   [[TMP9:%.*]] = bitcast <16 x i8> [[TMP7]] to <8 x half>
-// CHECK:   call void @llvm.arm.neon.vst2.p0i8.v8f16(i8* [[TMP3]], <8 x half> [[TMP8]], <8 x half> [[TMP9]], i32 2)
+// CHECK:   [[TMP8:%.*]] = bitcast <16 x i8> [[TMP5]] to <8 x i16>
+// CHECK:   [[TMP9:%.*]] = bitcast <16 x i8> [[TMP7]] to <8 x i16>
+// CHECK:   call void @llvm.arm.neon.vst2.p0i8.v8i16(i8* [[TMP3]], <8 x i16> [[TMP8]], <8 x i16> [[TMP9]], i32 2)
 // CHECK:   ret void
 void test_vst2q_f16(float16_t * a, float16x8x2_t b) {
   vst2q_f16(a, b);
@@ -16646,9 +16652,9 @@ void test_vst2_s64(int64_t * a, int64x1x2_t b) {
 // CHECK:   [[ARRAYIDX2:%.*]] = getelementptr inbounds [2 x <4 x half>], [2 x <4 x half>]* [[VAL1]], i32 0, i32 1
 // CHECK:   [[TMP6:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX2]], align 8
 // CHECK:   [[TMP7:%.*]] = bitcast <4 x half> [[TMP6]] to <8 x i8>
-// CHECK:   [[TMP8:%.*]] = bitcast <8 x i8> [[TMP5]] to <4 x half>
-// CHECK:   [[TMP9:%.*]] = bitcast <8 x i8> [[TMP7]] to <4 x half>
-// CHECK:   call void @llvm.arm.neon.vst2.p0i8.v4f16(i8* [[TMP3]], <4 x half> [[TMP8]], <4 x half> [[TMP9]], i32 2)
+// CHECK:   [[TMP8:%.*]] = bitcast <8 x i8> [[TMP5]] to <4 x i16>
+// CHECK:   [[TMP9:%.*]] = bitcast <8 x i8> [[TMP7]] to <4 x i16>
+// CHECK:   call void @llvm.arm.neon.vst2.p0i8.v4i16(i8* [[TMP3]], <4 x i16> [[TMP8]], <4 x i16> [[TMP9]], i32 2)
 // CHECK:   ret void
 void test_vst2_f16(float16_t * a, float16x4x2_t b) {
   vst2_f16(a, b);
@@ -16849,9 +16855,9 @@ void test_vst2q_lane_s32(int32_t * a, int32x4x2_t b) {
 // CHECK:   [[ARRAYIDX2:%.*]] = getelementptr inbounds [2 x <8 x half>], [2 x <8 x half>]* [[VAL1]], i32 0, i32 1
 // CHECK:   [[TMP6:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX2]], align 16
 // CHECK:   [[TMP7:%.*]] = bitcast <8 x half> [[TMP6]] to <16 x i8>
-// CHECK:   [[TMP8:%.*]] = bitcast <16 x i8> [[TMP5]] to <8 x half>
-// CHECK:   [[TMP9:%.*]] = bitcast <16 x i8> [[TMP7]] to <8 x half>
-// CHECK:   call void @llvm.arm.neon.vst2lane.p0i8.v8f16(i8* [[TMP3]], <8 x half> [[TMP8]], <8 x half> [[TMP9]], i32 7, i32 2)
+// CHECK:   [[TMP8:%.*]] = bitcast <16 x i8> [[TMP5]] to <8 x i16>
+// CHECK:   [[TMP9:%.*]] = bitcast <16 x i8> [[TMP7]] to <8 x i16>
+// CHECK:   call void @llvm.arm.neon.vst2lane.p0i8.v8i16(i8* [[TMP3]], <8 x i16> [[TMP8]], <8 x i16> [[TMP9]], i32 7, i32 2)
 // CHECK:   ret void
 void test_vst2q_lane_f16(float16_t * a, float16x8x2_t b) {
   vst2q_lane_f16(a, b, 7);
@@ -17073,9 +17079,9 @@ void test_vst2_lane_s32(int32_t * a, int32x2x2_t b) {
 // CHECK:   [[ARRAYIDX2:%.*]] = getelementptr inbounds [2 x <4 x half>], [2 x <4 x half>]* [[VAL1]], i32 0, i32 1
 // CHECK:   [[TMP6:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX2]], align 8
 // CHECK:   [[TMP7:%.*]] = bitcast <4 x half> [[TMP6]] to <8 x i8>
-// CHECK:   [[TMP8:%.*]] = bitcast <8 x i8> [[TMP5]] to <4 x half>
-// CHECK:   [[TMP9:%.*]] = bitcast <8 x i8> [[TMP7]] to <4 x half>
-// CHECK:   call void @llvm.arm.neon.vst2lane.p0i8.v4f16(i8* [[TMP3]], <4 x half> [[TMP8]], <4 x half> [[TMP9]], i32 3, i32 2)
+// CHECK:   [[TMP8:%.*]] = bitcast <8 x i8> [[TMP5]] to <4 x i16>
+// CHECK:   [[TMP9:%.*]] = bitcast <8 x i8> [[TMP7]] to <4 x i16>
+// CHECK:   call void @llvm.arm.neon.vst2lane.p0i8.v4i16(i8* [[TMP3]], <4 x i16> [[TMP8]], <4 x i16> [[TMP9]], i32 3, i32 2)
 // CHECK:   ret void
 void test_vst2_lane_f16(float16_t * a, float16x4x2_t b) {
   vst2_lane_f16(a, b, 3);
@@ -17348,10 +17354,10 @@ void test_vst3q_s32(int32_t * a, int32x4x3_t b) {
 // CHECK:   [[ARRAYIDX4:%.*]] = getelementptr inbounds [3 x <8 x half>], [3 x <8 x half>]* [[VAL3]], i32 0, i32 2
 // CHECK:   [[TMP8:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX4]], align 16
 // CHECK:   [[TMP9:%.*]] = bitcast <8 x half> [[TMP8]] to <16 x i8>
-// CHECK:   [[TMP10:%.*]] = bitcast <16 x i8> [[TMP5]] to <8 x half>
-// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP7]] to <8 x half>
-// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP9]] to <8 x half>
-// CHECK:   call void @llvm.arm.neon.vst3.p0i8.v8f16(i8* [[TMP3]], <8 x half> [[TMP10]], <8 x half> [[TMP11]], <8 x half> [[TMP12]], i32 2)
+// CHECK:   [[TMP10:%.*]] = bitcast <16 x i8> [[TMP5]] to <8 x i16>
+// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP7]] to <8 x i16>
+// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP9]] to <8 x i16>
+// CHECK:   call void @llvm.arm.neon.vst3.p0i8.v8i16(i8* [[TMP3]], <8 x i16> [[TMP10]], <8 x i16> [[TMP11]], <8 x i16> [[TMP12]], i32 2)
 // CHECK:   ret void
 void test_vst3q_f16(float16_t * a, float16x8x3_t b) {
   vst3q_f16(a, b);
@@ -17699,10 +17705,10 @@ void test_vst3_s64(int64_t * a, int64x1x3_t b) {
 // CHECK:   [[ARRAYIDX4:%.*]] = getelementptr inbounds [3 x <4 x half>], [3 x <4 x half>]* [[VAL3]], i32 0, i32 2
 // CHECK:   [[TMP8:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX4]], align 8
 // CHECK:   [[TMP9:%.*]] = bitcast <4 x half> [[TMP8]] to <8 x i8>
-// CHECK:   [[TMP10:%.*]] = bitcast <8 x i8> [[TMP5]] to <4 x half>
-// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP7]] to <4 x half>
-// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP9]] to <4 x half>
-// CHECK:   call void @llvm.arm.neon.vst3.p0i8.v4f16(i8* [[TMP3]], <4 x half> [[TMP10]], <4 x half> [[TMP11]], <4 x half> [[TMP12]], i32 2)
+// CHECK:   [[TMP10:%.*]] = bitcast <8 x i8> [[TMP5]] to <4 x i16>
+// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP7]] to <4 x i16>
+// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP9]] to <4 x i16>
+// CHECK:   call void @llvm.arm.neon.vst3.p0i8.v4i16(i8* [[TMP3]], <4 x i16> [[TMP10]], <4 x i16> [[TMP11]], <4 x i16> [[TMP12]], i32 2)
 // CHECK:   ret void
 void test_vst3_f16(float16_t * a, float16x4x3_t b) {
   vst3_f16(a, b);
@@ -17940,10 +17946,10 @@ void test_vst3q_lane_s32(int32_t * a, int32x4x3_t b) {
 // CHECK:   [[ARRAYIDX4:%.*]] = getelementptr inbounds [3 x <8 x half>], [3 x <8 x half>]* [[VAL3]], i32 0, i32 2
 // CHECK:   [[TMP8:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX4]], align 16
 // CHECK:   [[TMP9:%.*]] = bitcast <8 x half> [[TMP8]] to <16 x i8>
-// CHECK:   [[TMP10:%.*]] = bitcast <16 x i8> [[TMP5]] to <8 x half>
-// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP7]] to <8 x half>
-// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP9]] to <8 x half>
-// CHECK:   call void @llvm.arm.neon.vst3lane.p0i8.v8f16(i8* [[TMP3]], <8 x half> [[TMP10]], <8 x half> [[TMP11]], <8 x half> [[TMP12]], i32 7, i32 2)
+// CHECK:   [[TMP10:%.*]] = bitcast <16 x i8> [[TMP5]] to <8 x i16>
+// CHECK:   [[TMP11:%.*]] = bitcast <16 x i8> [[TMP7]] to <8 x i16>
+// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP9]] to <8 x i16>
+// CHECK:   call void @llvm.arm.neon.vst3lane.p0i8.v8i16(i8* [[TMP3]], <8 x i16> [[TMP10]], <8 x i16> [[TMP11]], <8 x i16> [[TMP12]], i32 7, i32 2)
 // CHECK:   ret void
 void test_vst3q_lane_f16(float16_t * a, float16x8x3_t b) {
   vst3q_lane_f16(a, b, 7);
@@ -18205,10 +18211,10 @@ void test_vst3_lane_s32(int32_t * a, int32x2x3_t b) {
 // CHECK:   [[ARRAYIDX4:%.*]] = getelementptr inbounds [3 x <4 x half>], [3 x <4 x half>]* [[VAL3]], i32 0, i32 2
 // CHECK:   [[TMP8:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX4]], align 8
 // CHECK:   [[TMP9:%.*]] = bitcast <4 x half> [[TMP8]] to <8 x i8>
-// CHECK:   [[TMP10:%.*]] = bitcast <8 x i8> [[TMP5]] to <4 x half>
-// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP7]] to <4 x half>
-// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP9]] to <4 x half>
-// CHECK:   call void @llvm.arm.neon.vst3lane.p0i8.v4f16(i8* [[TMP3]], <4 x half> [[TMP10]], <4 x half> [[TMP11]], <4 x half> [[TMP12]], i32 3, i32 2)
+// CHECK:   [[TMP10:%.*]] = bitcast <8 x i8> [[TMP5]] to <4 x i16>
+// CHECK:   [[TMP11:%.*]] = bitcast <8 x i8> [[TMP7]] to <4 x i16>
+// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP9]] to <4 x i16>
+// CHECK:   call void @llvm.arm.neon.vst3lane.p0i8.v4i16(i8* [[TMP3]], <4 x i16> [[TMP10]], <4 x i16> [[TMP11]], <4 x i16> [[TMP12]], i32 3, i32 2)
 // CHECK:   ret void
 void test_vst3_lane_f16(float16_t * a, float16x4x3_t b) {
   vst3_lane_f16(a, b, 3);
@@ -18524,11 +18530,11 @@ void test_vst4q_s32(int32_t * a, int32x4x4_t b) {
 // CHECK:   [[ARRAYIDX6:%.*]] = getelementptr inbounds [4 x <8 x half>], [4 x <8 x half>]* [[VAL5]], i32 0, i32 3
 // CHECK:   [[TMP10:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX6]], align 16
 // CHECK:   [[TMP11:%.*]] = bitcast <8 x half> [[TMP10]] to <16 x i8>
-// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP5]] to <8 x half>
-// CHECK:   [[TMP13:%.*]] = bitcast <16 x i8> [[TMP7]] to <8 x half>
-// CHECK:   [[TMP14:%.*]] = bitcast <16 x i8> [[TMP9]] to <8 x half>
-// CHECK:   [[TMP15:%.*]] = bitcast <16 x i8> [[TMP11]] to <8 x half>
-// CHECK:   call void @llvm.arm.neon.vst4.p0i8.v8f16(i8* [[TMP3]], <8 x half> [[TMP12]], <8 x half> [[TMP13]], <8 x half> [[TMP14]], <8 x half> [[TMP15]], i32 2)
+// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP5]] to <8 x i16>
+// CHECK:   [[TMP13:%.*]] = bitcast <16 x i8> [[TMP7]] to <8 x i16>
+// CHECK:   [[TMP14:%.*]] = bitcast <16 x i8> [[TMP9]] to <8 x i16>
+// CHECK:   [[TMP15:%.*]] = bitcast <16 x i8> [[TMP11]] to <8 x i16>
+// CHECK:   call void @llvm.arm.neon.vst4.p0i8.v8i16(i8* [[TMP3]], <8 x i16> [[TMP12]], <8 x i16> [[TMP13]], <8 x i16> [[TMP14]], <8 x i16> [[TMP15]], i32 2)
 // CHECK:   ret void
 void test_vst4q_f16(float16_t * a, float16x8x4_t b) {
   vst4q_f16(a, b);
@@ -18929,11 +18935,11 @@ void test_vst4_s64(int64_t * a, int64x1x4_t b) {
 // CHECK:   [[ARRAYIDX6:%.*]] = getelementptr inbounds [4 x <4 x half>], [4 x <4 x half>]* [[VAL5]], i32 0, i32 3
 // CHECK:   [[TMP10:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX6]], align 8
 // CHECK:   [[TMP11:%.*]] = bitcast <4 x half> [[TMP10]] to <8 x i8>
-// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP5]] to <4 x half>
-// CHECK:   [[TMP13:%.*]] = bitcast <8 x i8> [[TMP7]] to <4 x half>
-// CHECK:   [[TMP14:%.*]] = bitcast <8 x i8> [[TMP9]] to <4 x half>
-// CHECK:   [[TMP15:%.*]] = bitcast <8 x i8> [[TMP11]] to <4 x half>
-// CHECK:   call void @llvm.arm.neon.vst4.p0i8.v4f16(i8* [[TMP3]], <4 x half> [[TMP12]], <4 x half> [[TMP13]], <4 x half> [[TMP14]], <4 x half> [[TMP15]], i32 2)
+// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP5]] to <4 x i16>
+// CHECK:   [[TMP13:%.*]] = bitcast <8 x i8> [[TMP7]] to <4 x i16>
+// CHECK:   [[TMP14:%.*]] = bitcast <8 x i8> [[TMP9]] to <4 x i16>
+// CHECK:   [[TMP15:%.*]] = bitcast <8 x i8> [[TMP11]] to <4 x i16>
+// CHECK:   call void @llvm.arm.neon.vst4.p0i8.v4i16(i8* [[TMP3]], <4 x i16> [[TMP12]], <4 x i16> [[TMP13]], <4 x i16> [[TMP14]], <4 x i16> [[TMP15]], i32 2)
 // CHECK:   ret void
 void test_vst4_f16(float16_t * a, float16x4x4_t b) {
   vst4_f16(a, b);
@@ -19208,11 +19214,11 @@ void test_vst4q_lane_s32(int32_t * a, int32x4x4_t b) {
 // CHECK:   [[ARRAYIDX6:%.*]] = getelementptr inbounds [4 x <8 x half>], [4 x <8 x half>]* [[VAL5]], i32 0, i32 3
 // CHECK:   [[TMP10:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX6]], align 16
 // CHECK:   [[TMP11:%.*]] = bitcast <8 x half> [[TMP10]] to <16 x i8>
-// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP5]] to <8 x half>
-// CHECK:   [[TMP13:%.*]] = bitcast <16 x i8> [[TMP7]] to <8 x half>
-// CHECK:   [[TMP14:%.*]] = bitcast <16 x i8> [[TMP9]] to <8 x half>
-// CHECK:   [[TMP15:%.*]] = bitcast <16 x i8> [[TMP11]] to <8 x half>
-// CHECK:   call void @llvm.arm.neon.vst4lane.p0i8.v8f16(i8* [[TMP3]], <8 x half> [[TMP12]], <8 x half> [[TMP13]], <8 x half> [[TMP14]], <8 x half> [[TMP15]], i32 7, i32 2)
+// CHECK:   [[TMP12:%.*]] = bitcast <16 x i8> [[TMP5]] to <8 x i16>
+// CHECK:   [[TMP13:%.*]] = bitcast <16 x i8> [[TMP7]] to <8 x i16>
+// CHECK:   [[TMP14:%.*]] = bitcast <16 x i8> [[TMP9]] to <8 x i16>
+// CHECK:   [[TMP15:%.*]] = bitcast <16 x i8> [[TMP11]] to <8 x i16>
+// CHECK:   call void @llvm.arm.neon.vst4lane.p0i8.v8i16(i8* [[TMP3]], <8 x i16> [[TMP12]], <8 x i16> [[TMP13]], <8 x i16> [[TMP14]], <8 x i16> [[TMP15]], i32 7, i32 2)
 // CHECK:   ret void
 void test_vst4q_lane_f16(float16_t * a, float16x8x4_t b) {
   vst4q_lane_f16(a, b, 7);
@@ -19514,11 +19520,11 @@ void test_vst4_lane_s32(int32_t * a, int32x2x4_t b) {
 // CHECK:   [[ARRAYIDX6:%.*]] = getelementptr inbounds [4 x <4 x half>], [4 x <4 x half>]* [[VAL5]], i32 0, i32 3
 // CHECK:   [[TMP10:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX6]], align 8
 // CHECK:   [[TMP11:%.*]] = bitcast <4 x half> [[TMP10]] to <8 x i8>
-// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP5]] to <4 x half>
-// CHECK:   [[TMP13:%.*]] = bitcast <8 x i8> [[TMP7]] to <4 x half>
-// CHECK:   [[TMP14:%.*]] = bitcast <8 x i8> [[TMP9]] to <4 x half>
-// CHECK:   [[TMP15:%.*]] = bitcast <8 x i8> [[TMP11]] to <4 x half>
-// CHECK:   call void @llvm.arm.neon.vst4lane.p0i8.v4f16(i8* [[TMP3]], <4 x half> [[TMP12]], <4 x half> [[TMP13]], <4 x half> [[TMP14]], <4 x half> [[TMP15]], i32 3, i32 2)
+// CHECK:   [[TMP12:%.*]] = bitcast <8 x i8> [[TMP5]] to <4 x i16>
+// CHECK:   [[TMP13:%.*]] = bitcast <8 x i8> [[TMP7]] to <4 x i16>
+// CHECK:   [[TMP14:%.*]] = bitcast <8 x i8> [[TMP9]] to <4 x i16>
+// CHECK:   [[TMP15:%.*]] = bitcast <8 x i8> [[TMP11]] to <4 x i16>
+// CHECK:   call void @llvm.arm.neon.vst4lane.p0i8.v4i16(i8* [[TMP3]], <4 x i16> [[TMP12]], <4 x i16> [[TMP13]], <4 x i16> [[TMP14]], <4 x i16> [[TMP15]], i32 3, i32 2)
 // CHECK:   ret void
 void test_vst4_lane_f16(float16_t * a, float16x4x4_t b) {
   vst4_lane_f16(a, b, 3);
diff --git a/test/CodeGen/bitscan-builtins.c b/test/CodeGen/bitscan-builtins.c
index 71e49845f896..5fd3f13fbc73 100644
--- a/test/CodeGen/bitscan-builtins.c
+++ b/test/CodeGen/bitscan-builtins.c
@@ -1,5 +1,8 @@
 // RUN: %clang_cc1 -ffreestanding -triple x86_64-unknown-unknown -emit-llvm -o - %s | FileCheck %s
 
+// PR33722
+// RUN: %clang_cc1 -ffreestanding -triple x86_64-unknown-unknown -D_MSC_VER -emit-llvm -o - %s | FileCheck %s
+
 #include <immintrin.h>
 
 int test_bit_scan_forward(int a) {
diff --git a/test/CodeGen/default-address-space.c b/test/CodeGen/default-address-space.c
index fc5f55ffd6f4..b7f40585b267 100644
--- a/test/CodeGen/default-address-space.c
+++ b/test/CodeGen/default-address-space.c
@@ -1,24 +1,24 @@
 // RUN: %clang_cc1 -triple amdgcn -emit-llvm < %s | FileCheck -check-prefixes=PIZ,COM %s
 // RUN: %clang_cc1 -triple amdgcn---amdgiz -emit-llvm < %s | FileCheck -check-prefixes=CHECK,COM %s
 
-// PIZ-DAG: @foo = common addrspace(4) global i32 0
-// CHECK-DAG: @foo = common global i32 0
+// PIZ-DAG: @foo = common addrspace(1) global i32 0
+// CHECK-DAG: @foo = common addrspace(1) global i32 0
 int foo;
 
-// PIZ-DAG: @ban = common addrspace(4) global [10 x i32] zeroinitializer
-// CHECK-DAG: @ban = common global [10 x i32] zeroinitializer
+// PIZ-DAG: @ban = common addrspace(1) global [10 x i32] zeroinitializer
+// CHECK-DAG: @ban = common addrspace(1) global [10 x i32] zeroinitializer
 int ban[10];
 
-// PIZ-DAG: @A = common addrspace(4) global i32 addrspace(4)* null
-// PIZ-DAG: @B = common addrspace(4) global i32 addrspace(4)* null
-// CHECK-DAG: @A = common global i32* null
-// CHECK-DAG: @B = common global i32* null
+// PIZ-DAG: @A = common addrspace(1) global i32 addrspace(4)* null
+// PIZ-DAG: @B = common addrspace(1) global i32 addrspace(4)* null
+// CHECK-DAG: @A = common addrspace(1) global i32* null
+// CHECK-DAG: @B = common addrspace(1) global i32* null
 int *A;
 int *B;
 
 // COM-LABEL: define i32 @test1()
-// PIZ: load i32, i32 addrspace(4)* @foo
-// CHECK: load i32, i32* @foo
+// PIZ: load i32, i32 addrspace(4)* addrspacecast{{[^@]+}} @foo
+// CHECK: load i32, i32* addrspacecast{{[^@]+}} @foo
 int test1() { return foo; }
 
 // COM-LABEL: define i32 @test2(i32 %i)
@@ -30,13 +30,13 @@ int test1() { return foo; }
 int test2(int i) { return ban[i]; }
 
 // COM-LABEL: define void @test3()
-// PIZ: load i32 addrspace(4)*, i32 addrspace(4)* addrspace(4)* @B
+// PIZ: load i32 addrspace(4)*, i32 addrspace(4)* addrspace(4)* addrspacecast{{[^@]+}} @B
 // PIZ: load i32, i32 addrspace(4)*
-// PIZ: load i32 addrspace(4)*, i32 addrspace(4)* addrspace(4)* @A
+// PIZ: load i32 addrspace(4)*, i32 addrspace(4)* addrspace(4)* addrspacecast{{[^@]+}} @A
 // PIZ: store i32 {{.*}}, i32 addrspace(4)*
-// CHECK: load i32*, i32** @B
+// CHECK: load i32*, i32** addrspacecast{{.*}} @B
 // CHECK: load i32, i32*
-// CHECK: load i32*, i32** @A
+// CHECK: load i32*, i32** addrspacecast{{.*}} @A
 // CHECK: store i32 {{.*}}, i32*
 void test3() {
   *A = *B;
@@ -50,7 +50,7 @@ void test3() {
 // PIZ: %[[arrayidx:.*]] = getelementptr inbounds i32, i32 addrspace(4)* %[[r0]]
 // PIZ: store i32 0, i32 addrspace(4)* %[[arrayidx]]
 // CHECK-LABEL: define void @test4(i32* %a)
-// CHECK: %[[alloca:.*]] = alloca i32*, align 4, addrspace(5)
+// CHECK: %[[alloca:.*]] = alloca i32*, align 8, addrspace(5)
 // CHECK: %[[a_addr:.*]] = addrspacecast{{.*}} %[[alloca]] to i32**
 // CHECK: store i32* %a, i32** %[[a_addr]]
 // CHECK: %[[r0:.*]] = load i32*, i32** %[[a_addr]]
diff --git a/test/CodeGen/mcount.c b/test/CodeGen/mcount.c
index 2839d8ef6af3..2284acac0f8e 100644
--- a/test/CodeGen/mcount.c
+++ b/test/CodeGen/mcount.c
@@ -8,6 +8,10 @@
 // RUN: %clang_cc1 -pg -triple arm-netbsd-eabi -emit-llvm -o - %s | FileCheck -check-prefixes=CHECK-PREFIXED,NO-MCOUNT1 %s
 // RUN: %clang_cc1 -pg -triple aarch64-netbsd -emit-llvm -o - %s | FileCheck -check-prefixes=CHECK-PREFIXED,NO-MCOUNT1 %s
 // RUN: %clang_cc1 -pg -triple mips-netbsd -emit-llvm -o - %s | FileCheck -check-prefixes=CHECK-PREFIXED,NO-MCOUNT1 %s
+// RUN: %clang_cc1 -pg -triple mips-unknown-gnu-linux -emit-llvm -o - %s | FileCheck -check-prefixes=CHECK-PREFIXED,NO-MCOUNT1 %s
+// RUN: %clang_cc1 -pg -triple mipsel-unknown-gnu-linux -emit-llvm -o - %s | FileCheck -check-prefixes=CHECK-PREFIXED,NO-MCOUNT1 %s
+// RUN: %clang_cc1 -pg -triple mips64-unknown-gnu-linux -emit-llvm -o - %s | FileCheck -check-prefixes=CHECK-PREFIXED,NO-MCOUNT1 %s
+// RUN: %clang_cc1 -pg -triple mips64el-unknown-gnu-linux -emit-llvm -o - %s | FileCheck -check-prefixes=CHECK-PREFIXED,NO-MCOUNT1 %s
 // RUN: %clang_cc1 -pg -triple powerpc-netbsd -emit-llvm -o - %s | FileCheck -check-prefixes=CHECK-PREFIXED,NO-MCOUNT1 %s
 // RUN: %clang_cc1 -pg -triple powerpc64-netbsd -emit-llvm -o - %s | FileCheck -check-prefixes=CHECK-PREFIXED,NO-MCOUNT1 %s
 // RUN: %clang_cc1 -pg -triple powerpc64le-netbsd -emit-llvm -o - %s | FileCheck -check-prefixes=CHECK-PREFIXED,NO-MCOUNT1 %s
diff --git a/test/CodeGen/ms-barriers-intrinsics.c b/test/CodeGen/ms-barriers-intrinsics.c
index b0dfc3042a6e..c7da50cd0d83 100644
--- a/test/CodeGen/ms-barriers-intrinsics.c
+++ b/test/CodeGen/ms-barriers-intrinsics.c
@@ -13,19 +13,19 @@ typedef __SIZE_TYPE__ size_t;
 
 void test_ReadWriteBarrier() { _ReadWriteBarrier(); }
 // CHECK-LABEL: define void @test_ReadWriteBarrier
-// CHECK:   fence singlethread seq_cst
+// CHECK:   fence syncscope("singlethread") seq_cst
 // CHECK:   ret void
 // CHECK: }
 
 void test_ReadBarrier() { _ReadBarrier(); }
 // CHECK-LABEL: define void @test_ReadBarrier
-// CHECK:   fence singlethread seq_cst
+// CHECK:   fence syncscope("singlethread") seq_cst
 // CHECK:   ret void
 // CHECK: }
 
 void test_WriteBarrier() { _WriteBarrier(); }
 // CHECK-LABEL: define void @test_WriteBarrier
-// CHECK:   fence singlethread seq_cst
+// CHECK:   fence syncscope("singlethread") seq_cst
 // CHECK:   ret void
 // CHECK: }
 
diff --git a/test/CodeGen/no-devirt.cpp b/test/CodeGen/no-devirt.cpp
index 4333b7cde7c6..544b1394f421 100644
--- a/test/CodeGen/no-devirt.cpp
+++ b/test/CodeGen/no-devirt.cpp
@@ -21,7 +21,7 @@ public:
 struct Wrapper {
   TmplWithArray<bool, 10> data;
   bool indexIt(int a) {
-    if (a > 6) return data[a] ;      // Should not devirtualize
+    if (a > 6) return data[a] ;      // Should devirtualize
     if (a > 4) return data.func1(a); // Should devirtualize
     return data.func2(a);            // Should devirtualize
   }
@@ -53,7 +53,7 @@ bool stuff(int p)
 }
 #endif
 
-// CHECK-NOT: call {{.*}} @_ZN13TmplWithArrayIbLi10EEixEi
+// CHECK-DAG: call {{.*}} @_ZN13TmplWithArrayIbLi10EEixEi
 // CHECK-DAG: call {{.*}} @_ZN13TmplWithArrayIbLi10EE5func1Ei
 // CHECK-DAG: call {{.*}} @_ZN13TmplWithArrayIbLi10EE5func2Ei
 
diff --git a/test/CodeGen/pgo-sample-thinlto-summary.c b/test/CodeGen/pgo-sample-thinlto-summary.c
index 51c8faa6be6d..7045db08f22e 100644
--- a/test/CodeGen/pgo-sample-thinlto-summary.c
+++ b/test/CodeGen/pgo-sample-thinlto-summary.c
@@ -1,9 +1,7 @@
 // RUN: %clang_cc1 -O2 -fprofile-sample-use=%S/Inputs/pgo-sample-thinlto-summary.prof %s -emit-llvm -o - 2>&1 | FileCheck %s -check-prefix=SAMPLEPGO
 // RUN: %clang_cc1 -O2 -fprofile-sample-use=%S/Inputs/pgo-sample-thinlto-summary.prof %s -emit-llvm -flto=thin -o - 2>&1 | FileCheck %s -check-prefix=THINLTO
 // RUN: %clang_cc1 -O2 -fexperimental-new-pass-manager -fprofile-sample-use=%S/Inputs/pgo-sample-thinlto-summary.prof %s -emit-llvm -o - 2>&1 | FileCheck %s -check-prefix=SAMPLEPGO
-// FIXME: Run the following command once LTOPreLinkDefaultPipeline is
-//        customized.
-// %clang_cc1 -O2 -fexperimental-new-pass-manager -fprofile-sample-use=%S/Inputs/pgo-sample-thinlto-summary.prof %s -emit-llvm -flto=thin -o - 2>&1 | FileCheck %s -check-prefix=THINLTO
+// RUN: %clang_cc1 -O2 -fexperimental-new-pass-manager -fprofile-sample-use=%S/Inputs/pgo-sample-thinlto-summary.prof %s -emit-llvm -flto=thin -o - 2>&1 | FileCheck %s -check-prefix=THINLTO
 // Checks if hot call is inlined by normal compile, but not inlined by
 // thinlto compile.
 
diff --git a/test/CodeGenCXX/amdgcn-automatic-variable.cpp b/test/CodeGenCXX/amdgcn-automatic-variable.cpp
index 7df27c28e6d2..a82275206c6d 100644
--- a/test/CodeGenCXX/amdgcn-automatic-variable.cpp
+++ b/test/CodeGenCXX/amdgcn-automatic-variable.cpp
@@ -15,8 +15,8 @@ void func2(void) {
   // CHECK: %lv1 = alloca i32, align 4, addrspace(5)
   // CHECK: %lv2 = alloca i32, align 4, addrspace(5)
   // CHECK: %la = alloca [100 x i32], align 4, addrspace(5)
-  // CHECK: %lp1 = alloca i32*, align 4, addrspace(5)
-  // CHECK: %lp2 = alloca i32*, align 4, addrspace(5)
+  // CHECK: %lp1 = alloca i32*, align 8, addrspace(5)
+  // CHECK: %lp2 = alloca i32*, align 8, addrspace(5)
   // CHECK: %lvc = alloca i32, align 4, addrspace(5)
 
   // CHECK: %[[r0:.*]] = addrspacecast i32 addrspace(5)* %lv1 to i32*
@@ -34,12 +34,12 @@ void func2(void) {
   la[0] = 3;
 
   // CHECK: %[[r3:.*]] = addrspacecast i32* addrspace(5)* %lp1 to i32**
-  // CHECK: store i32* %[[r0]], i32** %[[r3]], align 4
+  // CHECK: store i32* %[[r0]], i32** %[[r3]], align 8
   int *lp1 = &lv1;
 
   // CHECK: %[[r4:.*]] = addrspacecast i32* addrspace(5)* %lp2 to i32**
   // CHECK: %[[arraydecay:.*]] = getelementptr inbounds [100 x i32], [100 x i32]* %[[r2]], i32 0, i32 0
-  // CHECK: store i32* %[[arraydecay]], i32** %[[r4]], align 4
+  // CHECK: store i32* %[[arraydecay]], i32** %[[r4]], align 8
   int *lp2 = la;
 
   // CHECK: call void @_Z5func1Pi(i32* %[[r0]])
@@ -80,3 +80,5 @@ void func4(int x) {
   // CHECK: call void @_Z5func1Pi(i32* %[[r0]])
   func1(&x);
 }
+
+// CHECK-NOT: !opencl.ocl.version
diff --git a/test/CodeGenCXX/cxx0x-initializer-stdinitializerlist.cpp b/test/CodeGenCXX/cxx0x-initializer-stdinitializerlist.cpp
index 7bab11488ad4..67215ef48fb3 100644
--- a/test/CodeGenCXX/cxx0x-initializer-stdinitializerlist.cpp
+++ b/test/CodeGenCXX/cxx0x-initializer-stdinitializerlist.cpp
@@ -1,4 +1,5 @@
-// RUN: %clang_cc1 -std=c++11 -triple x86_64-none-linux-gnu -emit-llvm -o - %s | FileCheck %s
+// RUN: %clang_cc1 -std=c++11 -triple x86_64-none-linux-gnu -emit-llvm -o - %s | FileCheck -check-prefixes=X86,CHECK %s
+// RUN: %clang_cc1 -std=c++11 -triple amdgcn-amd-amdhsa-amdgiz -DNO_TLS -emit-llvm -o - %s | FileCheck -check-prefixes=AMD,CHECK %s
 
 namespace std {
   typedef decltype(sizeof(int)) size_t;
@@ -46,54 +47,72 @@ struct wantslist1 {
   wantslist1(std::initializer_list<destroyme1>);
   ~wantslist1();
 };
-
-// CHECK: @_ZGR15globalInitList1_ = internal constant [3 x i32] [i32 1, i32 2, i32 3]
-// CHECK: @globalInitList1 = global %{{[^ ]+}} { i32* getelementptr inbounds ([3 x i32], [3 x i32]* @_ZGR15globalInitList1_, i32 0, i32 0), i{{32|64}} 3 }
+// X86: @_ZGR15globalInitList1_ = internal constant [3 x i32] [i32 1, i32 2, i32 3]
+// X86: @globalInitList1 = global %{{[^ ]+}} { i32* getelementptr inbounds ([3 x i32], [3 x i32]* @_ZGR15globalInitList1_, i32 0, i32 0), i{{32|64}} 3 }
+// AMD: @_ZGR15globalInitList1_ = internal addrspace(1) constant [3 x i32] [i32 1, i32 2, i32 3]
+// AMD: @globalInitList1 = addrspace(1) global %{{[^ ]+}} { i32* addrspacecast (i32 addrspace(1)* getelementptr inbounds ([3 x i32], [3 x i32] addrspace(1)* @_ZGR15globalInitList1_, i32 0, i32 0) to i32*), i{{32|64}} 3 }
 std::initializer_list<int> globalInitList1 = {1, 2, 3};
 
+#ifndef NO_TLS
 namespace thread_local_global_array {
-  // FIXME: We should be able to constant-evaluate this even though the
-  // initializer is not a constant expression (pointers to thread_local
-  // objects aren't really a problem).
-  //
-  // CHECK: @_ZN25thread_local_global_array1xE = thread_local global
-  // CHECK: @_ZGRN25thread_local_global_array1xE_ = internal thread_local constant [4 x i32] [i32 1, i32 2, i32 3, i32 4]
-  std::initializer_list<int> thread_local x = { 1, 2, 3, 4 };
+// FIXME: We should be able to constant-evaluate this even though the
+// initializer is not a constant expression (pointers to thread_local
+// objects aren't really a problem).
+//
+// X86: @_ZN25thread_local_global_array1xE = thread_local global
+// X86: @_ZGRN25thread_local_global_array1xE_ = internal thread_local constant [4 x i32] [i32 1, i32 2, i32 3, i32 4]
+std::initializer_list<int> thread_local x = {1, 2, 3, 4};
 }
-
-// CHECK: @globalInitList2 = global %{{[^ ]+}} zeroinitializer
-// CHECK: @_ZGR15globalInitList2_ = internal global [2 x %[[WITHARG:[^ ]*]]] zeroinitializer
-
-// CHECK: @_ZN15partly_constant1kE = global i32 0, align 4
-// CHECK: @_ZN15partly_constant2ilE = global {{.*}} null, align 8
-// CHECK: @[[PARTLY_CONSTANT_OUTER:_ZGRN15partly_constant2ilE.*]] = internal global {{.*}} zeroinitializer, align 8
-// CHECK: @[[PARTLY_CONSTANT_INNER:_ZGRN15partly_constant2ilE.*]] = internal global [3 x {{.*}}] zeroinitializer, align 8
-// CHECK: @[[PARTLY_CONSTANT_FIRST:_ZGRN15partly_constant2ilE.*]] = internal constant [3 x i32] [i32 1, i32 2, i32 3], align 4
-// CHECK: @[[PARTLY_CONSTANT_SECOND:_ZGRN15partly_constant2ilE.*]] = internal global [2 x i32] zeroinitializer, align 4
-// CHECK: @[[PARTLY_CONSTANT_THIRD:_ZGRN15partly_constant2ilE.*]] = internal constant [4 x i32] [i32 5, i32 6, i32 7, i32 8], align 4
-
-// CHECK: @[[REFTMP1:.*]] = private constant [2 x i32] [i32 42, i32 43], align 4
-// CHECK: @[[REFTMP2:.*]] = private constant [3 x %{{.*}}] [%{{.*}} { i32 1 }, %{{.*}} { i32 2 }, %{{.*}} { i32 3 }], align 4
+#endif
+
+// X86: @globalInitList2 = global %{{[^ ]+}} zeroinitializer
+// X86: @_ZGR15globalInitList2_ = internal global [2 x %[[WITHARG:[^ ]*]]] zeroinitializer
+// AMD: @globalInitList2 = addrspace(1) global %{{[^ ]+}} zeroinitializer
+// AMD: @_ZGR15globalInitList2_ = internal addrspace(1) global [2 x %[[WITHARG:[^ ]*]]] zeroinitializer
+
+// X86: @_ZN15partly_constant1kE = global i32 0, align 4
+// X86: @_ZN15partly_constant2ilE = global {{.*}} null, align 8
+// X86: @[[PARTLY_CONSTANT_OUTER:_ZGRN15partly_constant2ilE.*]] = internal global {{.*}} zeroinitializer, align 8
+// X86: @[[PARTLY_CONSTANT_INNER:_ZGRN15partly_constant2ilE.*]] = internal global [3 x {{.*}}] zeroinitializer, align 8
+// X86: @[[PARTLY_CONSTANT_FIRST:_ZGRN15partly_constant2ilE.*]] = internal constant [3 x i32] [i32 1, i32 2, i32 3], align 4
+// X86: @[[PARTLY_CONSTANT_SECOND:_ZGRN15partly_constant2ilE.*]] = internal global [2 x i32] zeroinitializer, align 4
+// X86: @[[PARTLY_CONSTANT_THIRD:_ZGRN15partly_constant2ilE.*]] = internal constant [4 x i32] [i32 5, i32 6, i32 7, i32 8], align 4
+// AMD: @_ZN15partly_constant1kE = addrspace(1) global i32 0, align 4
+// AMD: @_ZN15partly_constant2ilE = addrspace(2) global {{.*}} null, align 8
+// AMD: @[[PARTLY_CONSTANT_OUTER:_ZGRN15partly_constant2ilE.*]] = internal addrspace(2) global {{.*}} zeroinitializer, align 8
+// AMD: @[[PARTLY_CONSTANT_INNER:_ZGRN15partly_constant2ilE.*]] = internal addrspace(2) global [3 x {{.*}}] zeroinitializer, align 8
+// AMD: @[[PARTLY_CONSTANT_FIRST:_ZGRN15partly_constant2ilE.*]] = internal addrspace(2) constant [3 x i32] [i32 1, i32 2, i32 3], align 4
+// AMD: @[[PARTLY_CONSTANT_SECOND:_ZGRN15partly_constant2ilE.*]] = internal addrspace(2) global [2 x i32] zeroinitializer, align 4
+// AMD: @[[PARTLY_CONSTANT_THIRD:_ZGRN15partly_constant2ilE.*]] = internal addrspace(2) constant [4 x i32] [i32 5, i32 6, i32 7, i32 8], align 4
+
+// X86: @[[REFTMP1:.*]] = private constant [2 x i32] [i32 42, i32 43], align 4
+// X86: @[[REFTMP2:.*]] = private constant [3 x %{{.*}}] [%{{.*}} { i32 1 }, %{{.*}} { i32 2 }, %{{.*}} { i32 3 }], align 4
+// AMD: @[[REFTMP1:.*]] = private addrspace(2) constant [2 x i32] [i32 42, i32 43], align 4
+// AMD: @[[REFTMP2:.*]] = private addrspace(2) constant [3 x %{{.*}}] [%{{.*}} { i32 1 }, %{{.*}} { i32 2 }, %{{.*}} { i32 3 }], align 4
 
 // CHECK: appending global
 
-
 // thread_local initializer:
-// CHECK-LABEL: define internal void
-// CHECK: store i32* getelementptr inbounds ([4 x i32], [4 x i32]* @_ZGRN25thread_local_global_array1xE_, i64 0, i64 0),
-// CHECK:       i32** getelementptr inbounds ({{.*}}, {{.*}}* @_ZN25thread_local_global_array1xE, i32 0, i32 0), align 8
-// CHECK: store i64 4, i64* getelementptr inbounds ({{.*}}, {{.*}}* @_ZN25thread_local_global_array1xE, i32 0, i32 1), align 8
-
-
-// CHECK-LABEL: define internal void
-// CHECK: call void @_ZN8witharg1C1ERK10destroyme1(%[[WITHARG]]* getelementptr inbounds ([2 x %[[WITHARG]]], [2 x %[[WITHARG]]]* @_ZGR15globalInitList2_, i{{32|64}} 0, i{{32|64}} 0
-// CHECK: call void @_ZN8witharg1C1ERK10destroyme1(%[[WITHARG]]* getelementptr inbounds ([2 x %[[WITHARG]]], [2 x %[[WITHARG]]]* @_ZGR15globalInitList2_, i{{32|64}} 0, i{{32|64}} 1
-// CHECK: __cxa_atexit
-// CHECK: store %[[WITHARG]]* getelementptr inbounds ([2 x %[[WITHARG]]], [2 x %[[WITHARG]]]* @_ZGR15globalInitList2_, i64 0, i64 0),
-// CHECK:       %[[WITHARG]]** getelementptr inbounds (%{{.*}}, %{{.*}}* @globalInitList2, i32 0, i32 0), align 8
-// CHECK: store i64 2, i64* getelementptr inbounds (%{{.*}}, %{{.*}}* @globalInitList2, i32 0, i32 1), align 8
-// CHECK: call void @_ZN10destroyme1D1Ev
+// X86-LABEL: define internal void @__cxx_global_var_init
+// X86: store i32* getelementptr inbounds ([4 x i32], [4 x i32]* @_ZGRN25thread_local_global_array1xE_, i64 0, i64 0),
+// X86:       i32** getelementptr inbounds ({{.*}}, {{.*}}* @_ZN25thread_local_global_array1xE, i32 0, i32 0), align 8
+// X86: store i64 4, i64* getelementptr inbounds ({{.*}}, {{.*}}* @_ZN25thread_local_global_array1xE, i32 0, i32 1), align 8
+
+// CHECK-LABEL: define internal void @__cxx_global_var_init
+// X86: call void @_ZN8witharg1C1ERK10destroyme1(%[[WITHARG]]* getelementptr inbounds ([2 x %[[WITHARG]]], [2 x %[[WITHARG]]]* @_ZGR15globalInitList2_, i{{32|64}} 0, i{{32|64}} 0
+// X86: call void @_ZN8witharg1C1ERK10destroyme1(%[[WITHARG]]* getelementptr inbounds ([2 x %[[WITHARG]]], [2 x %[[WITHARG]]]* @_ZGR15globalInitList2_, i{{32|64}} 0, i{{32|64}} 1
+// AMD: call void @_ZN8witharg1C1ERK10destroyme1(%[[WITHARG]]* getelementptr inbounds ([2 x %[[WITHARG]]], [2 x %[[WITHARG]]]* addrspacecast ({{[^@]+}} @_ZGR15globalInitList2_ {{[^)]+}}), i{{32|64}} 0, i{{32|64}} 0
+// AMD: call void @_ZN8witharg1C1ERK10destroyme1(%[[WITHARG]]* getelementptr inbounds ([2 x %[[WITHARG]]], [2 x %[[WITHARG]]]* addrspacecast ({{[^@]+}} @_ZGR15globalInitList2_ {{[^)]+}}), i{{32|64}} 0, i{{32|64}} 1
+// CHECK: call i32 @__cxa_atexit
+// X86: store %[[WITHARG]]* getelementptr inbounds ([2 x %[[WITHARG]]], [2 x %[[WITHARG]]]* @_ZGR15globalInitList2_, i64 0, i64 0),
+// X86:       %[[WITHARG]]** getelementptr inbounds (%{{.*}}, %{{.*}}* @globalInitList2, i32 0, i32 0), align 8
+// X86: store i64 2, i64* getelementptr inbounds (%{{.*}}, %{{.*}}* @globalInitList2, i32 0, i32 1), align 8
+// AMD: store %[[WITHARG]]* getelementptr inbounds ([2 x %[[WITHARG]]], [2 x %[[WITHARG]]]* addrspacecast ({{[^@]+}} @_ZGR15globalInitList2_ {{[^)]+}}), i64 0, i64 0),
+// AMD:       %[[WITHARG]]** getelementptr inbounds (%{{.*}}, %{{.*}}* addrspacecast ({{[^@]+}} @globalInitList2 {{[^)]+}}), i32 0, i32 0), align 8
+// AMD: store i64 2, i64* getelementptr inbounds (%{{.*}}, %{{.*}}* addrspacecast ({{[^@]+}} @globalInitList2 {{[^)]+}}), i32 0, i32 1), align 8
 // CHECK: call void @_ZN10destroyme1D1Ev
+// CHECK-NEXT: call void @_ZN10destroyme1D1Ev
+// CHECK-NEXT: ret void
 std::initializer_list<witharg1> globalInitList2 = {
   witharg1(destroyme1()), witharg1(destroyme1())
 };
@@ -101,7 +120,9 @@ std::initializer_list<witharg1> globalInitList2 = {
 void fn1(int i) {
   // CHECK-LABEL: define void @_Z3fn1i
   // temporary array
-  // CHECK: [[array:%[^ ]+]] = alloca [3 x i32]
+  // X86: [[array:%[^ ]+]] = alloca [3 x i32]
+  // AMD: [[alloca:%[^ ]+]] = alloca [3 x i32], align 4, addrspace(5)
+  // AMD: [[array:%[^ ]+]] = addrspacecast [3 x i32] addrspace(5)* [[alloca]] to [3 x i32]*
   // CHECK: getelementptr inbounds [3 x i32], [3 x i32]* [[array]], i{{32|64}} 0
   // CHECK-NEXT: store i32 1, i32*
   // CHECK-NEXT: getelementptr
@@ -175,7 +196,6 @@ void fn6() {
   destroyme2 dm2;
   // CHECK: call void @_ZN10destroyme2D1Ev
 }
-
 void fn7() {
   // CHECK-LABEL: define void @_Z3fn7v
   // temps should be destroyed before dm2
@@ -366,37 +386,36 @@ namespace partly_constant {
   std::initializer_list<std::initializer_list<int>> &&il = { { 1, 2, 3 }, { 4, k }, { 5, 6, 7, 8 } };
   // First init list.
   // CHECK-NOT: @[[PARTLY_CONSTANT_FIRST]],
-  // CHECK: store i32* getelementptr inbounds ({{.*}}, {{.*}}* @[[PARTLY_CONSTANT_FIRST]], i64 0, i64 0),
-  // CHECK:       i32** getelementptr inbounds ({{.*}}, {{.*}}* @[[PARTLY_CONSTANT_INNER]], i64 0, i64 0, i32 0)
-  // CHECK: store i64 3, i64* getelementptr inbounds ({{.*}}, {{.*}}* @[[PARTLY_CONSTANT_INNER]], i64 0, i64 0, i32 1)
+  // CHECK: store i32* getelementptr inbounds ({{.*}}, {{.*}}* {{.*}}@[[PARTLY_CONSTANT_FIRST]]{{.*}}, i64 0, i64 0),
+  // CHECK:       i32** getelementptr inbounds ({{.*}}, {{.*}}* {{.*}}@[[PARTLY_CONSTANT_INNER]]{{.*}}, i64 0, i64 0, i32 0)
+  // CHECK: store i64 3, i64* getelementptr inbounds ({{.*}}, {{.*}}* {{.*}}@[[PARTLY_CONSTANT_INNER]]{{.*}}, i64 0, i64 0, i32 1)
   // CHECK-NOT: @[[PARTLY_CONSTANT_FIRST]],
   //
   // Second init list array (non-constant).
-  // CHECK: store i32 4, i32* getelementptr inbounds ({{.*}}, {{.*}}* @[[PARTLY_CONSTANT_SECOND]], i64 0, i64 0)
-  // CHECK: load i32, i32* @_ZN15partly_constant1kE
-  // CHECK: store i32 {{.*}}, i32* getelementptr inbounds ({{.*}}, {{.*}}* @[[PARTLY_CONSTANT_SECOND]], i64 0, i64 1)
+  // CHECK: store i32 4, i32* getelementptr inbounds ({{.*}}, {{.*}}* {{.*}}@[[PARTLY_CONSTANT_SECOND]]{{.*}}, i64 0, i64 0)
+  // CHECK: load i32, i32* {{.*}}@_ZN15partly_constant1kE
+  // CHECK: store i32 {{.*}}, i32* getelementptr inbounds ({{.*}}, {{.*}}* {{.*}}@[[PARTLY_CONSTANT_SECOND]]{{.*}}, i64 0, i64 1)
   //
   // Second init list.
-  // CHECK: store i32* getelementptr inbounds ({{.*}}, {{.*}}* @[[PARTLY_CONSTANT_SECOND]], i64 0, i64 0),
-  // CHECK:       i32** getelementptr inbounds ({{.*}}, {{.*}}* @[[PARTLY_CONSTANT_INNER]], i64 0, i64 1, i32 0)
-  // CHECK: store i64 2, i64* getelementptr inbounds ({{.*}}, {{.*}}* @[[PARTLY_CONSTANT_INNER]], i64 0, i64 1, i32 1)
+  // CHECK: store i32* getelementptr inbounds ({{.*}}, {{.*}}* {{.*}}@[[PARTLY_CONSTANT_SECOND]]{{.*}}, i64 0, i64 0),
+  // CHECK:       i32** getelementptr inbounds ({{.*}}, {{.*}}* {{.*}}@[[PARTLY_CONSTANT_INNER]]{{.*}}, i64 0, i64 1, i32 0)
+  // CHECK: store i64 2, i64* getelementptr inbounds ({{.*}}, {{.*}}* {{.*}}@[[PARTLY_CONSTANT_INNER]]{{.*}}, i64 0, i64 1, i32 1)
   //
   // Third init list.
   // CHECK-NOT: @[[PARTLY_CONSTANT_THIRD]],
-  // CHECK: store i32* getelementptr inbounds ({{.*}}, {{.*}}* @[[PARTLY_CONSTANT_THIRD]], i64 0, i64 0),
-  // CHECK:       i32** getelementptr inbounds ({{.*}}, {{.*}}* @[[PARTLY_CONSTANT_INNER]], i64 0, i64 2, i32 0)
-  // CHECK: store i64 4, i64* getelementptr inbounds ({{.*}}, {{.*}}* @_ZGRN15partly_constant2ilE4_, i64 0, i64 2, i32 1)
+  // CHECK: store i32* getelementptr inbounds ({{.*}}, {{.*}}* {{.*}}@[[PARTLY_CONSTANT_THIRD]]{{.*}}, i64 0, i64 0),
+  // CHECK:       i32** getelementptr inbounds ({{.*}}, {{.*}}* {{.*}}@[[PARTLY_CONSTANT_INNER]]{{.*}}, i64 0, i64 2, i32 0)
+  // CHECK: store i64 4, i64* getelementptr inbounds ({{.*}}, {{.*}}* {{.*}}@_ZGRN15partly_constant2ilE4_{{.*}}, i64 0, i64 2, i32 1)
   // CHECK-NOT: @[[PARTLY_CONSTANT_THIRD]],
   //
   // Outer init list.
-  // CHECK: store {{.*}}* getelementptr inbounds ({{.*}}, {{.*}}* @[[PARTLY_CONSTANT_INNER]], i64 0, i64 0),
-  // CHECK:       {{.*}}** getelementptr inbounds ({{.*}}, {{.*}}* @[[PARTLY_CONSTANT_OUTER]], i32 0, i32 0)
-  // CHECK: store i64 3, i64* getelementptr inbounds ({{.*}}, {{.*}}* @[[PARTLY_CONSTANT_OUTER]], i32 0, i32 1)
+  // CHECK: store {{.*}}* getelementptr inbounds ({{.*}}, {{.*}}* {{.*}}@[[PARTLY_CONSTANT_INNER]]{{.*}}, i64 0, i64 0),
+  // CHECK:       {{.*}}** getelementptr inbounds ({{.*}}, {{.*}}* {{.*}}@[[PARTLY_CONSTANT_OUTER]]{{.*}}, i32 0, i32 0)
+  // CHECK: store i64 3, i64* getelementptr inbounds ({{.*}}, {{.*}}* {{.*}}@[[PARTLY_CONSTANT_OUTER]]{{.*}}, i32 0, i32 1)
   //
   // 'il' reference.
-  // CHECK: store {{.*}}* @[[PARTLY_CONSTANT_OUTER]], {{.*}}** @_ZN15partly_constant2ilE, align 8
+  // CHECK: store {{.*}}* {{.*}}@[[PARTLY_CONSTANT_OUTER]]{{.*}}, {{.*}}** {{.*}}@_ZN15partly_constant2ilE{{.*}}, align 8
 }
-
 namespace nested {
   struct A { A(); ~A(); };
   struct B { const A &a; ~B(); };
@@ -463,7 +482,7 @@ namespace PR20445 {
   template<int x> void f() { new MyClass({42, 43}); }
   template void f<0>();
   // CHECK-LABEL: define {{.*}} @_ZN7PR204451fILi0EEEvv(
-  // CHECK: store i32* getelementptr inbounds ([2 x i32], [2 x i32]* @[[REFTMP1]], i64 0, i64 0)
+  // CHECK: store i32* getelementptr inbounds ([2 x i32], [2 x i32]* {{.*}}@[[REFTMP1]]{{.*}}, i64 0, i64 0)
   // CHECK: call void @_ZN7PR204456vectorC1ESt16initializer_listIiE(
   // CHECK: call void @_ZN7PR204457MyClassC1ERKNS_6vectorE(
 }
@@ -476,9 +495,9 @@ namespace ConstExpr {
   };
   void f(std::initializer_list<C>);
   void g() {
-// CHECK-LABEL: _ZN9ConstExpr1gEv
-// CHECK: store %"class.ConstExpr::C"* getelementptr inbounds ([3 x %"class.ConstExpr::C"], [3 x %"class.ConstExpr::C"]* @[[REFTMP2]], i64 0, i64 0)
-// CHECK: call void @_ZN9ConstExpr1fESt16initializer_listINS_1CEE
+    // CHECK-LABEL: _ZN9ConstExpr1gEv
+    // CHECK: store %"class.ConstExpr::C"* getelementptr inbounds ([3 x %"class.ConstExpr::C"], [3 x %"class.ConstExpr::C"]* {{.*}}@[[REFTMP2]]{{.*}}, i64 0, i64 0)
+    // CHECK: call void @_ZN9ConstExpr1fESt16initializer_listINS_1CEE
     f({C(1), C(2), C(3)});
   }
 }
@@ -498,11 +517,13 @@ namespace B19773010 {
   void f1() {
     // CHECK-LABEL: @_ZN9B197730102f1Ev
     testcase a{{"", ENUM_CONSTANT}};
-    // CHECK: store %"struct.B19773010::pair"* getelementptr inbounds ([1 x %"struct.B19773010::pair"], [1 x %"struct.B19773010::pair"]* bitcast ([1 x { i8*, i32 }]* @.ref.tmp{{.*}} to [1 x %"struct.B19773010::pair"]*), i64 0, i64 0), %"struct.B19773010::pair"** %{{.*}}, align 8
+    // X86: store %"struct.B19773010::pair"* getelementptr inbounds ([1 x %"struct.B19773010::pair"], [1 x %"struct.B19773010::pair"]* bitcast ([1 x { i8*, i32 }]* @.ref.tmp{{.*}} to [1 x %"struct.B19773010::pair"]*), i64 0, i64 0), %"struct.B19773010::pair"** %{{.*}}, align 8
+    // AMD: store %"struct.B19773010::pair"* getelementptr inbounds ([1 x %"struct.B19773010::pair"], [1 x %"struct.B19773010::pair"]* addrspacecast{{.*}} bitcast ([1 x { i8*, i32 }] addrspace(2)* @.ref.tmp{{.*}} to [1 x %"struct.B19773010::pair"] addrspace(2)*){{.*}}, i64 0, i64 0), %"struct.B19773010::pair"** %{{.*}}, align 8
   }
   void f2() {
     // CHECK-LABEL: @_ZN9B197730102f2Ev
-    // CHECK: store %"struct.B19773010::pair"* getelementptr inbounds ([1 x %"struct.B19773010::pair"], [1 x %"struct.B19773010::pair"]* bitcast ([1 x { i8*, i32 }]* @_ZGRZN9B197730102f2EvE1p_ to [1 x %"struct.B19773010::pair"]*), i64 0, i64 0), %"struct.B19773010::pair"** getelementptr inbounds ([2 x %"class.std::initializer_list.10"], [2 x %"class.std::initializer_list.10"]* @_ZZN9B197730102f2EvE1p, i64 0, i64 1, i32 0), align 16
+    // X86: store %"struct.B19773010::pair"* getelementptr inbounds ([1 x %"struct.B19773010::pair"], [1 x %"struct.B19773010::pair"]* bitcast ([1 x { i8*, i32 }]* @_ZGRZN9B197730102f2EvE1p_ to [1 x %"struct.B19773010::pair"]*), i64 0, i64 0), %"struct.B19773010::pair"** getelementptr inbounds ([2 x %"class.std::initializer_list.10"], [2 x %"class.std::initializer_list.10"]* @_ZZN9B197730102f2EvE1p, i64 0, i64 1, i32 0), align 16
+    // AMD: store %"struct.B19773010::pair"* getelementptr inbounds ([1 x %"struct.B19773010::pair"], [1 x %"struct.B19773010::pair"]* addrspacecast{{.*}} bitcast ([1 x { i8*, i32 }] addrspace(1)* @_ZGRZN9B197730102f2EvE1p_ to [1 x %"struct.B19773010::pair"] addrspace(1)*){{.*}}, i64 0, i64 0), %"struct.B19773010::pair"** getelementptr inbounds ([2 x %"class.std::initializer_list.10"], [2 x %"class.std::initializer_list.10"]* addrspacecast{{.*}}@_ZZN9B197730102f2EvE1p{{.*}}, i64 0, i64 1, i32 0), align 8
     static std::initializer_list<pair<const char *, E>> a, p[2] =
         {a, {{"", ENUM_CONSTANT}}};
   }
diff --git a/test/CodeGenCXX/devirtualize-virtual-function-calls-final.cpp b/test/CodeGenCXX/devirtualize-virtual-function-calls-final.cpp
index b90620ab600f..2ab2f759cfe3 100644
--- a/test/CodeGenCXX/devirtualize-virtual-function-calls-final.cpp
+++ b/test/CodeGenCXX/devirtualize-virtual-function-calls-final.cpp
@@ -241,3 +241,53 @@ namespace Test10 {
     return static_cast<A *>(b)->f();
   }
 }
+
+namespace Test11 {
+  // Check that the definitions of Derived's operators are emitted.
+
+  // CHECK-LABEL: define linkonce_odr void @_ZN6Test111SIiE4foo1Ev(
+  // CHECK: call void @_ZN6Test111SIiE7DerivedclEv(
+  // CHECK: call zeroext i1 @_ZN6Test111SIiE7DerivedeqERKNS_4BaseE(
+  // CHECK: call zeroext i1 @_ZN6Test111SIiE7DerivedntEv(
+  // CHECK: call dereferenceable(4) %"class.Test11::Base"* @_ZN6Test111SIiE7DerivedixEi(
+  // CHECK: define linkonce_odr void @_ZN6Test111SIiE7DerivedclEv(
+  // CHECK: define linkonce_odr zeroext i1 @_ZN6Test111SIiE7DerivedeqERKNS_4BaseE(
+  // CHECK: define linkonce_odr zeroext i1 @_ZN6Test111SIiE7DerivedntEv(
+  // CHECK: define linkonce_odr dereferenceable(4) %"class.Test11::Base"* @_ZN6Test111SIiE7DerivedixEi(
+  class Base {
+  public:
+    virtual void operator()() {}
+    virtual bool operator==(const Base &other) { return false; }
+    virtual bool operator!() { return false; }
+    virtual Base &operator[](int i) { return *this; }
+  };
+
+  template<class T>
+  struct S {
+    class Derived final : public Base {
+    public:
+      void operator()() override {}
+      bool operator==(const Base &other) override { return true; }
+      bool operator!() override { return true; }
+      Base &operator[](int i) override { return *this; }
+    };
+
+    Derived *ptr = nullptr, *ptr2 = nullptr;
+
+    void foo1() {
+      if (ptr && ptr2) {
+        // These calls get devirtualized. Linkage fails if the definitions of
+        // the called functions are not emitted.
+        (*ptr)();
+        (void)(*ptr == *ptr2);
+        (void)(!(*ptr));
+        (void)((*ptr)[1]);
+      }
+    }
+  };
+
+  void foo2() {
+    S<int> *s = new S<int>;
+    s->foo1();
+  }
+}
diff --git a/test/CodeGenCXX/dllimport-memptr-global.cpp b/test/CodeGenCXX/dllimport-memptr-global.cpp
new file mode 100644
index 000000000000..e64537b8b9f9
--- /dev/null
+++ b/test/CodeGenCXX/dllimport-memptr-global.cpp
@@ -0,0 +1,58 @@
+// Also check that -Wglobal-constructors does the right thing. Strictly
+// speaking, this is a Sema test, but this avoids test case duplication.
+// RUN: %clang_cc1 -Wglobal-constructors %s -verify -triple i686-windows-msvc -fms-extensions -std=c++11
+//
+// RUN: %clang_cc1 %s -emit-llvm -o - -triple i686-windows-msvc -fms-extensions -std=c++11 | FileCheck %s
+
+struct __declspec(dllimport) Single {
+  void nonvirt();
+  virtual void virt();
+};
+
+struct A { int a; };
+struct B { int b; };
+struct __declspec(dllimport) Multi : A, B {
+  void nonvirt();
+  virtual void virt();
+};
+
+struct __declspec(dllimport) Virtual : virtual A {
+  void nonvirt();
+  virtual void virt();
+};
+
+struct General;
+static_assert(sizeof(void (General::*)()) == 16, "force general memptr model");
+struct __declspec(dllimport) General {
+  void nonvirt();
+  virtual void virt();
+};
+
+auto mp_single_nv = &Single::nonvirt; // expected-warning {{global constructor}}
+auto mp_multi_nv = &Multi::nonvirt; // expected-warning {{global constructor}}
+auto mp_virtual_nv = &Virtual::nonvirt; // expected-warning {{global constructor}}
+auto mp_general_nv = &General::nonvirt; // expected-warning {{global constructor}}
+
+auto mp_single_v = &Single::virt;
+auto mp_multi_v = &Multi::virt;
+auto mp_virtual_v = &Virtual::virt;
+auto mp_general_v = &General::virt;
+
+// All of the non-virtual globals need dynamic initializers.
+
+// CHECK: @"\01?mp_single_nv@@3P8Single@@AEXXZQ1@" = global i8* null, align 4
+// CHECK: @"\01?mp_multi_nv@@3P8Multi@@AEXXZQ1@" = global { i8*, i32 } zeroinitializer, align 4
+// CHECK: @"\01?mp_virtual_nv@@3P8Virtual@@AEXXZQ1@" = global { i8*, i32, i32 } zeroinitializer, align 4
+// CHECK: @"\01?mp_general_nv@@3P8General@@AEXXZQ1@" = global { i8*, i32, i32, i32 } zeroinitializer, align 4
+
+// CHECK: @"\01?mp_single_v@@3P8Single@@AEXXZQ1@" = global i8* bitcast (void (%struct.Single*, ...)* @"\01??_9Single@@$BA@AE" to i8*), align 4
+// CHECK: @"\01?mp_multi_v@@3P8Multi@@AEXXZQ1@" = global { i8*, i32 } { i8* bitcast (void (%struct.Multi*, ...)* @"\01??_9Multi@@$BA@AE" to i8*), i32 0 }, align 4
+// CHECK: @"\01?mp_virtual_v@@3P8Virtual@@AEXXZQ1@" = global { i8*, i32, i32 } { i8* bitcast (void (%struct.Virtual*, ...)* @"\01??_9Virtual@@$BA@AE" to i8*), i32 0, i32 0 }, align 4
+// CHECK: @"\01?mp_general_v@@3P8General@@AEXXZQ1@" = global { i8*, i32, i32, i32 } { i8* bitcast (void (%struct.General*, ...)* @"\01??_9General@@$BA@AE" to i8*), i32 0, i32 0, i32 0 }, align 4
+
+// CHECK: define internal void @_GLOBAL__sub_I{{.*}}() {{.*}} {
+// CHECK:   call void @"\01??__Emp_single_nv@@YAXXZ"()
+// CHECK:   call void @"\01??__Emp_multi_nv@@YAXXZ"()
+// CHECK:   call void @"\01??__Emp_virtual_nv@@YAXXZ"()
+// CHECK:   call void @"\01??__Emp_general_nv@@YAXXZ"()
+// CHECK: }
diff --git a/test/CodeGenCXX/vtable-available-externally.cpp b/test/CodeGenCXX/vtable-available-externally.cpp
index db99f73d9e72..2e2cdbbfeff5 100644
--- a/test/CodeGenCXX/vtable-available-externally.cpp
+++ b/test/CodeGenCXX/vtable-available-externally.cpp
@@ -275,9 +275,8 @@ struct C {
   virtual D& operator=(const D&);
 };
 
-// Cannot emit D's vtable available_externally, because we cannot create
-// a reference to the inline virtual D::operator= function.
-// CHECK-TEST11: @_ZTVN6Test111DE = external unnamed_addr constant
+// Can emit D's vtable available_externally.
+// CHECK-TEST11: @_ZTVN6Test111DE = available_externally unnamed_addr constant
 struct D : C {
   virtual void key();
 };
diff --git a/test/CodeGenCXX/windows-itanium-type-info.cpp b/test/CodeGenCXX/windows-itanium-type-info.cpp
index ad89318f599e..285b59815da2 100644
--- a/test/CodeGenCXX/windows-itanium-type-info.cpp
+++ b/test/CodeGenCXX/windows-itanium-type-info.cpp
@@ -32,9 +32,15 @@ void f() {
 // CHECK-DAG: @_ZTV7derived = dllexport unnamed_addr constant
 
 // CHECK-DAG: @_ZTI4base = external dllimport constant
-// CHECK-DAG: @_ZTS4base = external dllimport constant
-// CHECK-NOT: @_ZTV4base = external dllimport constant
 
 // CHECK-EH-IMPORT: @_ZTS4base = linkonce_odr constant
 // CHECK-EH-IMPORT: @_ZTI4base = linkonce_odr constant
 
+struct __declspec(dllimport) gatekeeper {};
+struct zuul : gatekeeper {
+  virtual ~zuul();
+};
+zuul::~zuul() {}
+
+// CHECK-DAG: @_ZTI10gatekeeper = linkonce_odr constant
+// CHECK-DAG: @_ZTS10gatekeeper = linkonce_odr constant
diff --git a/test/CodeGenOpenCL/address-spaces.cl b/test/CodeGenOpenCL/address-spaces.cl
index 7c665286547a..488b8f9d480e 100644
--- a/test/CodeGenOpenCL/address-spaces.cl
+++ b/test/CodeGenOpenCL/address-spaces.cl
@@ -4,6 +4,8 @@
 // RUN: %clang_cc1 %s -O0 -triple amdgcn-amd-amdhsa-opencl -DCL20 -cl-std=CL2.0 -emit-llvm -o - | FileCheck %s --check-prefixes=CL20,CL20SPIR
 // RUN: %clang_cc1 %s -O0 -triple amdgcn-amd-amdhsa-amdgizcl -emit-llvm -o - | FileCheck %s -check-prefixes=CHECK,GIZ
 // RUN: %clang_cc1 %s -O0 -triple amdgcn-amd-amdhsa-amdgizcl -DCL20 -cl-std=CL2.0 -emit-llvm -o - | FileCheck %s --check-prefixes=CL20,CL20GIZ
+// RUN: %clang_cc1 %s -O0 -triple amdgcn-mesa-mesa3d -emit-llvm -o - | FileCheck --check-prefixes=CHECK,SPIR %s
+// RUN: %clang_cc1 %s -O0 -triple r600-- -emit-llvm -o - | FileCheck --check-prefixes=CHECK,SPIR %s
 
 // SPIR: i32* %arg
 // GIZ: i32 addrspace(5)* %arg
diff --git a/test/CodeGenOpenCL/amdgcn-automatic-variable.cl b/test/CodeGenOpenCL/amdgcn-automatic-variable.cl
index 29309567ec71..19287c7d8998 100644
--- a/test/CodeGenOpenCL/amdgcn-automatic-variable.cl
+++ b/test/CodeGenOpenCL/amdgcn-automatic-variable.cl
@@ -22,8 +22,8 @@ void func2(void) {
   // CHECK: %la = alloca [100 x i32], align 4, addrspace(5)
   // CL12: %lp1 = alloca i32 addrspace(5)*, align 4, addrspace(5)
   // CL12: %lp2 = alloca i32 addrspace(5)*, align 4, addrspace(5)
-  // CL20: %lp1 = alloca i32*, align 4, addrspace(5)
-  // CL20: %lp2 = alloca i32*, align 4, addrspace(5)
+  // CL20: %lp1 = alloca i32*, align 8, addrspace(5)
+  // CL20: %lp2 = alloca i32*, align 8, addrspace(5)
   // CHECK: %lvc = alloca i32, align 4, addrspace(5)
 
   // CHECK: store i32 1, i32 addrspace(5)* %lv1
@@ -39,13 +39,13 @@ void func2(void) {
 
   // CL12: store i32 addrspace(5)* %lv1, i32 addrspace(5)* addrspace(5)* %lp1, align 4
   // CL20: %[[r0:.*]] = addrspacecast i32 addrspace(5)* %lv1 to i32*
-  // CL20: store i32* %[[r0]], i32* addrspace(5)* %lp1, align 4
+  // CL20: store i32* %[[r0]], i32* addrspace(5)* %lp1, align 8
   int *lp1 = &lv1;
 
   // CHECK: %[[arraydecay:.*]] = getelementptr inbounds [100 x i32], [100 x i32] addrspace(5)* %la, i32 0, i32 0
   // CL12: store i32 addrspace(5)* %[[arraydecay]], i32 addrspace(5)* addrspace(5)* %lp2, align 4
   // CL20: %[[r1:.*]] = addrspacecast i32 addrspace(5)* %[[arraydecay]] to i32*
-  // CL20: store i32* %[[r1]], i32* addrspace(5)* %lp2, align 4
+  // CL20: store i32* %[[r1]], i32* addrspace(5)* %lp2, align 8
   int *lp2 = la;
 
   // CL12: call void @func1(i32 addrspace(5)* %lv1)
diff --git a/test/CodeGenOpenCL/amdgpu-nullptr.cl b/test/CodeGenOpenCL/amdgpu-nullptr.cl
index 402be5760cf7..69f54fcaa483 100644
--- a/test/CodeGenOpenCL/amdgpu-nullptr.cl
+++ b/test/CodeGenOpenCL/amdgpu-nullptr.cl
@@ -27,13 +27,13 @@ private char *private_p = 0;
 // CHECK: @local_p = local_unnamed_addr addrspace(1) global i8 addrspace(3)* addrspacecast (i8 addrspace(4)* null to i8 addrspace(3)*), align 4
 local char *local_p = 0;
 
-// CHECK: @global_p = local_unnamed_addr addrspace(1) global i8 addrspace(1)* null, align 4
+// CHECK: @global_p = local_unnamed_addr addrspace(1) global i8 addrspace(1)* null, align 8
 global char *global_p = 0;
 
-// CHECK: @constant_p = local_unnamed_addr addrspace(1) global i8 addrspace(2)* null, align 4
+// CHECK: @constant_p = local_unnamed_addr addrspace(1) global i8 addrspace(2)* null, align 8
 constant char *constant_p = 0;
 
-// CHECK: @generic_p = local_unnamed_addr addrspace(1) global i8 addrspace(4)* null, align 4
+// CHECK: @generic_p = local_unnamed_addr addrspace(1) global i8 addrspace(4)* null, align 8
 generic char *generic_p = 0;
 
 // Test NULL as initializer.
@@ -44,19 +44,19 @@ private char *private_p_NULL = NULL;
 // CHECK: @local_p_NULL = local_unnamed_addr addrspace(1) global i8 addrspace(3)* addrspacecast (i8 addrspace(4)* null to i8 addrspace(3)*), align 4
 local char *local_p_NULL = NULL;
 
-// CHECK: @global_p_NULL = local_unnamed_addr addrspace(1) global i8 addrspace(1)* null, align 4
+// CHECK: @global_p_NULL = local_unnamed_addr addrspace(1) global i8 addrspace(1)* null, align 8
 global char *global_p_NULL = NULL;
 
-// CHECK: @constant_p_NULL = local_unnamed_addr addrspace(1) global i8 addrspace(2)* null, align 4
+// CHECK: @constant_p_NULL = local_unnamed_addr addrspace(1) global i8 addrspace(2)* null, align 8
 constant char *constant_p_NULL = NULL;
 
-// CHECK: @generic_p_NULL = local_unnamed_addr addrspace(1) global i8 addrspace(4)* null, align 4
+// CHECK: @generic_p_NULL = local_unnamed_addr addrspace(1) global i8 addrspace(4)* null, align 8
 generic char *generic_p_NULL = NULL;
 
 // Test constant folding of null pointer.
 // A null pointer should be folded to a null pointer in the target address space.
 
-// CHECK: @fold_generic = local_unnamed_addr addrspace(1) global i32 addrspace(4)* null, align 4
+// CHECK: @fold_generic = local_unnamed_addr addrspace(1) global i32 addrspace(4)* null, align 8
 generic int *fold_generic = (global int*)(generic float*)(private char*)0;
 
 // CHECK: @fold_priv = local_unnamed_addr addrspace(1) global i16* null, align 4
@@ -104,8 +104,8 @@ int fold_int5_local = (int) &((local StructTy1*)0)->p2;
 // NOOPT: @test_static_var_private.sp3 = internal addrspace(1) global i8* null, align 4
 // NOOPT: @test_static_var_private.sp4 = internal addrspace(1) global i8* null, align 4
 // NOOPT: @test_static_var_private.sp5 = internal addrspace(1) global i8* null, align 4
-// NOOPT: @test_static_var_private.SS1 = internal addrspace(1) global %struct.StructTy1 { i8* null, i8 addrspace(3)* addrspacecast (i8 addrspace(4)* null to i8 addrspace(3)*), i8 addrspace(2)* null, i8 addrspace(1)* null, i8 addrspace(4)* null }, align 4
-// NOOPT: @test_static_var_private.SS2 = internal addrspace(1) global %struct.StructTy2 zeroinitializer, align 4
+// NOOPT: @test_static_var_private.SS1 = internal addrspace(1) global %struct.StructTy1 { i8* null, i8 addrspace(3)* addrspacecast (i8 addrspace(4)* null to i8 addrspace(3)*), i8 addrspace(2)* null, i8 addrspace(1)* null, i8 addrspace(4)* null }, align 8
+// NOOPT: @test_static_var_private.SS2 = internal addrspace(1) global %struct.StructTy2 zeroinitializer, align 8
 
 void test_static_var_private(void) {
   static private char *sp1 = 0;
@@ -123,8 +123,8 @@ void test_static_var_private(void) {
 // NOOPT: @test_static_var_local.sp3 = internal addrspace(1) global i8 addrspace(3)* addrspacecast (i8 addrspace(4)* null to i8 addrspace(3)*), align 4
 // NOOPT: @test_static_var_local.sp4 = internal addrspace(1) global i8 addrspace(3)* null, align 4
 // NOOPT: @test_static_var_local.sp5 = internal addrspace(1) global i8 addrspace(3)* null, align 4
-// NOOPT: @test_static_var_local.SS1 = internal addrspace(1) global %struct.StructTy1 { i8* null, i8 addrspace(3)* addrspacecast (i8 addrspace(4)* null to i8 addrspace(3)*), i8 addrspace(2)* null, i8 addrspace(1)* null, i8 addrspace(4)* null }, align 4
-// NOOPT: @test_static_var_local.SS2 = internal addrspace(1) global %struct.StructTy2 zeroinitializer, align 4
+// NOOPT: @test_static_var_local.SS1 = internal addrspace(1) global %struct.StructTy1 { i8* null, i8 addrspace(3)* addrspacecast (i8 addrspace(4)* null to i8 addrspace(3)*), i8 addrspace(2)* null, i8 addrspace(1)* null, i8 addrspace(4)* null }, align 8
+// NOOPT: @test_static_var_local.SS2 = internal addrspace(1) global %struct.StructTy2 zeroinitializer, align 8
 void test_static_var_local(void) {
   static local char *sp1 = 0;
   static local char *sp2 = NULL;
@@ -143,9 +143,9 @@ void test_static_var_local(void) {
 // NOOPT: store i8* null, i8** %sp3, align 4
 // NOOPT: store i8* null, i8** %sp4, align 4
 // NOOPT: %[[SS1:.*]] = bitcast %struct.StructTy1* %SS1 to i8*
-// NOOPT: call void @llvm.memcpy.p0i8.p2i8.i64(i8* %[[SS1]], i8 addrspace(2)* bitcast (%struct.StructTy1 addrspace(2)* @test_func_scope_var_private.SS1 to i8 addrspace(2)*), i64 32, i32 4, i1 false)
+// NOOPT: call void @llvm.memcpy.p0i8.p2i8.i64(i8* %[[SS1]], i8 addrspace(2)* bitcast (%struct.StructTy1 addrspace(2)* @test_func_scope_var_private.SS1 to i8 addrspace(2)*), i64 32, i32 8, i1 false)
 // NOOPT: %[[SS2:.*]] = bitcast %struct.StructTy2* %SS2 to i8*
-// NOOPT: call void @llvm.memset.p0i8.i64(i8* %[[SS2]], i8 0, i64 24, i32 4, i1 false)
+// NOOPT: call void @llvm.memset.p0i8.i64(i8* %[[SS2]], i8 0, i64 24, i32 8, i1 false)
 void test_func_scope_var_private(void) {
   private char *sp1 = 0;
   private char *sp2 = NULL;
@@ -163,9 +163,9 @@ void test_func_scope_var_private(void) {
 // NOOPT: store i8 addrspace(3)* null, i8 addrspace(3)** %sp3, align 4
 // NOOPT: store i8 addrspace(3)* null, i8 addrspace(3)** %sp4, align 4
 // NOOPT: %[[SS1:.*]] = bitcast %struct.StructTy1* %SS1 to i8*
-// NOOPT: call void @llvm.memcpy.p0i8.p2i8.i64(i8* %[[SS1]], i8 addrspace(2)* bitcast (%struct.StructTy1 addrspace(2)* @test_func_scope_var_local.SS1 to i8 addrspace(2)*), i64 32, i32 4, i1 false)
+// NOOPT: call void @llvm.memcpy.p0i8.p2i8.i64(i8* %[[SS1]], i8 addrspace(2)* bitcast (%struct.StructTy1 addrspace(2)* @test_func_scope_var_local.SS1 to i8 addrspace(2)*), i64 32, i32 8, i1 false)
 // NOOPT: %[[SS2:.*]] = bitcast %struct.StructTy2* %SS2 to i8*
-// NOOPT: call void @llvm.memset.p0i8.i64(i8* %[[SS2]], i8 0, i64 24, i32 4, i1 false)
+// NOOPT: call void @llvm.memset.p0i8.i64(i8* %[[SS2]], i8 0, i64 24, i32 8, i1 false)
 void test_func_scope_var_local(void) {
   local char *sp1 = 0;
   local char *sp2 = NULL;
@@ -189,28 +189,28 @@ private char *p1;
 // CHECK: @p2 = weak local_unnamed_addr addrspace(1) global i8 addrspace(3)* addrspacecast (i8 addrspace(4)* null to i8 addrspace(3)*), align 4
 local char *p2;
 
-// CHECK: @p3 = common local_unnamed_addr addrspace(1) global i8 addrspace(2)* null, align 4
+// CHECK: @p3 = common local_unnamed_addr addrspace(1) global i8 addrspace(2)* null, align 8
 constant char *p3;
 
-// CHECK: @p4 = common local_unnamed_addr addrspace(1) global i8 addrspace(1)* null, align 4
+// CHECK: @p4 = common local_unnamed_addr addrspace(1) global i8 addrspace(1)* null, align 8
 global char *p4;
 
-// CHECK: @p5 = common local_unnamed_addr addrspace(1) global i8 addrspace(4)* null, align 4
+// CHECK: @p5 = common local_unnamed_addr addrspace(1) global i8 addrspace(4)* null, align 8
 generic char *p5;
 
 // Test default initialization of sturcture.
 
-// CHECK: @S1 = weak local_unnamed_addr addrspace(1) global %struct.StructTy1 { i8* null, i8 addrspace(3)* addrspacecast (i8 addrspace(4)* null to i8 addrspace(3)*), i8 addrspace(2)* null, i8 addrspace(1)* null, i8 addrspace(4)* null }, align 4
+// CHECK: @S1 = weak local_unnamed_addr addrspace(1) global %struct.StructTy1 { i8* null, i8 addrspace(3)* addrspacecast (i8 addrspace(4)* null to i8 addrspace(3)*), i8 addrspace(2)* null, i8 addrspace(1)* null, i8 addrspace(4)* null }, align 8
 StructTy1 S1;
 
-// CHECK: @S2 = common local_unnamed_addr addrspace(1) global %struct.StructTy2 zeroinitializer, align 4
+// CHECK: @S2 = common local_unnamed_addr addrspace(1) global %struct.StructTy2 zeroinitializer, align 8
 StructTy2 S2;
 
 // Test default initialization of array.
-// CHECK: @A1 = weak local_unnamed_addr addrspace(1) global [2 x %struct.StructTy1] [%struct.StructTy1 { i8* null, i8 addrspace(3)* addrspacecast (i8 addrspace(4)* null to i8 addrspace(3)*), i8 addrspace(2)* null, i8 addrspace(1)* null, i8 addrspace(4)* null }, %struct.StructTy1 { i8* null, i8 addrspace(3)* addrspacecast (i8 addrspace(4)* null to i8 addrspace(3)*), i8 addrspace(2)* null, i8 addrspace(1)* null, i8 addrspace(4)* null }], align 4
+// CHECK: @A1 = weak local_unnamed_addr addrspace(1) global [2 x %struct.StructTy1] [%struct.StructTy1 { i8* null, i8 addrspace(3)* addrspacecast (i8 addrspace(4)* null to i8 addrspace(3)*), i8 addrspace(2)* null, i8 addrspace(1)* null, i8 addrspace(4)* null }, %struct.StructTy1 { i8* null, i8 addrspace(3)* addrspacecast (i8 addrspace(4)* null to i8 addrspace(3)*), i8 addrspace(2)* null, i8 addrspace(1)* null, i8 addrspace(4)* null }], align 8
 StructTy1 A1[2];
 
-// CHECK: @A2 = common local_unnamed_addr addrspace(1) global [2 x %struct.StructTy2] zeroinitializer, align 4
+// CHECK: @A2 = common local_unnamed_addr addrspace(1) global [2 x %struct.StructTy2] zeroinitializer, align 8
 StructTy2 A2[2];
 
 // Test comparison with 0.
@@ -597,7 +597,7 @@ int test_and_ptr(private char* p1, local char* p2) {
 // Test folding of null pointer in function scope.
 // NOOPT-LABEL: test_fold_private
 // NOOPT: call void @test_fold_callee
-// NOOPT: store i32 addrspace(1)* null, i32 addrspace(1)** %glob, align 4
+// NOOPT: store i32 addrspace(1)* null, i32 addrspace(1)** %glob, align 8
 // NOOPT: %{{.*}} = sub i64 %{{.*}}, 0
 // NOOPT: call void @test_fold_callee
 // NOOPT: %{{.*}} = add nsw i64 %{{.*}}, 0
@@ -612,7 +612,7 @@ void test_fold_private(void) {
 
 // NOOPT-LABEL: test_fold_local
 // NOOPT: call void @test_fold_callee
-// NOOPT: store i32 addrspace(1)* null, i32 addrspace(1)** %glob, align 4
+// NOOPT: store i32 addrspace(1)* null, i32 addrspace(1)** %glob, align 8
 // NOOPT: %{{.*}} = sub i64 %{{.*}}, 0
 // NOOPT: call void @test_fold_callee
 // NOOPT: %{{.*}} = add nsw i64 %{{.*}}, sext (i32 ptrtoint (i32 addrspace(3)* addrspacecast (i32 addrspace(4)* null to i32 addrspace(3)*) to i32) to i64)
diff --git a/test/CodeGenOpenCL/amdgpu-sizeof-alignof.cl b/test/CodeGenOpenCL/amdgpu-sizeof-alignof.cl
new file mode 100644
index 000000000000..a5d438933fa4
--- /dev/null
+++ b/test/CodeGenOpenCL/amdgpu-sizeof-alignof.cl
@@ -0,0 +1,70 @@
+// RUN: %clang_cc1 -triple r600 -cl-std=CL1.2 %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -triple amdgcn-mesa-mesa3d -cl-std=CL1.2 %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -triple amdgcn---opencl -cl-std=CL1.2 %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -triple amdgcn---opencl -cl-std=CL2.0 %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -triple amdgcn---amdgizcl -cl-std=CL1.2 %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -triple amdgcn---amdgizcl -cl-std=CL2.0 %s -emit-llvm -o - | FileCheck %s
+
+#ifdef __AMDGCN__
+#define PTSIZE 8
+#else
+#define PTSIZE 4
+#endif
+
+#ifdef cl_khr_fp64
+#pragma OPENCL EXTENSION cl_khr_fp64 : enable
+#endif
+#ifdef cl_khr_fp16
+#pragma OPENCL EXTENSION cl_khr_fp16 : enable
+#endif
+
+typedef __SIZE_TYPE__ size_t;
+typedef __PTRDIFF_TYPE__ ptrdiff_t;
+typedef __INTPTR_TYPE__ intptr_t;
+typedef __UINTPTR_TYPE__ uintptr_t;
+typedef global void *global_ptr_t;
+typedef constant void *constant_ptr_t;
+typedef local void *local_ptr_t;
+typedef private void *private_ptr_t;
+
+void check(bool);
+
+void test() {
+  // CHECK-NOT: call void @check(i1 zeroext false)
+  check(sizeof(size_t) == PTSIZE);
+  check(__alignof__(size_t) == PTSIZE);
+  check(sizeof(intptr_t) == PTSIZE);
+  check(__alignof__(intptr_t) == PTSIZE);
+  check(sizeof(uintptr_t) == PTSIZE);
+  check(__alignof__(uintptr_t) == PTSIZE);
+  check(sizeof(ptrdiff_t) == PTSIZE);
+  check(__alignof__(ptrdiff_t) == PTSIZE);
+
+  check(sizeof(char) == 1);
+  check(__alignof__(char) == 1);
+  check(sizeof(short) == 2);
+  check(__alignof__(short) == 2);
+  check(sizeof(int) == 4);
+  check(__alignof__(int) == 4);
+  check(sizeof(long) == 8);
+  check(__alignof__(long) == 8);
+#ifdef cl_khr_fp16
+  check(sizeof(half) == 2);
+  check(__alignof__(half) == 2);
+#endif
+  check(sizeof(float) == 4);
+  check(__alignof__(float) == 4);
+#ifdef cl_khr_fp64
+  check(sizeof(double) == 8);
+  check(__alignof__(double) == 8);
+#endif
+
+  check(sizeof(void*) == (__OPENCL_C_VERSION__ >= 200 ? 8 : 4));
+  check(__alignof__(void*) == (__OPENCL_C_VERSION__ >= 200 ? 8 : 4));
+  check(sizeof(global_ptr_t) == PTSIZE);
+  check(__alignof__(global_ptr_t) == PTSIZE);
+  check(sizeof(constant_ptr_t) == PTSIZE);
+  check(__alignof__(constant_ptr_t) == PTSIZE);
+  check(sizeof(local_ptr_t) == 4);
+  check(__alignof__(private_ptr_t) == 4);
+}
diff --git a/test/Driver/autocomplete.c b/test/Driver/autocomplete.c
index f0bdcce5707c..e9e2e992bae9 100644
--- a/test/Driver/autocomplete.c
+++ b/test/Driver/autocomplete.c
@@ -36,3 +36,7 @@
 // MTHREADMODELALL: posix single
 // RUN: %clang --autocomplete=-mrelocation-model, | FileCheck %s -check-prefix=MRELOCMODELALL
 // MRELOCMODELALL: dynamic-no-pic pic ropi ropi-rwpi rwpi static
+// RUN: %clang --autocomplete=-mrelocation-mode | FileCheck %s -check-prefix=MRELOCMODEL_CLANG
+// MRELOCMODEL_CLANG-NOT: -mrelocation-model
+// RUN: %clang --autocomplete=#-mrelocation-mode | FileCheck %s -check-prefix=MRELOCMODEL_CC1
+// MRELOCMODEL_CC1: -mrelocation-model
diff --git a/test/Driver/clang_f_opts.c b/test/Driver/clang_f_opts.c
index e4b72d69ca3f..c17cec6eba9b 100644
--- a/test/Driver/clang_f_opts.c
+++ b/test/Driver/clang_f_opts.c
@@ -356,6 +356,8 @@
 // RUN: -ftree-vrp                                                            \
 // RUN: -fno-devirtualize                                                     \
 // RUN: -fno-devirtualize-speculatively                                       \
+// RUN: -fslp-vectorize-aggressive                                            \
+// RUN: -fno-slp-vectorize-aggressive                                         \
 // RUN: %s 2>&1 | FileCheck --check-prefix=CHECK-WARNING %s
 // CHECK-WARNING-DAG: optimization flag '-finline-limit=1000' is not supported
 // CHECK-WARNING-DAG: optimization flag '-finline-limit' is not supported
@@ -422,6 +424,8 @@
 // CHECK-WARNING-DAG: optimization flag '-ftree-vrp' is not supported
 // CHECK-WARNING-DAG: optimization flag '-fno-devirtualize' is not supported
 // CHECK-WARNING-DAG: optimization flag '-fno-devirtualize-speculatively' is not supported
+// CHECK-WARNING-DAG: the flag '-fslp-vectorize-aggressive' has been deprecated and will be ignored
+// CHECK-WARNING-DAG: the flag '-fno-slp-vectorize-aggressive' has been deprecated and will be ignored
 
 // Test that we mute the warning on these
 // RUN: %clang -### -finline-limit=1000 -Wno-invalid-command-line-argument              \
diff --git a/test/Driver/crash report spaces.c b/test/Driver/crash-report-spaces.c
index 9bc4626b48b8..3e95a0de2516 100644
--- a/test/Driver/crash report spaces.c
+++ b/test/Driver/crash-report-spaces.c
@@ -1,6 +1,7 @@
 // RUN: rm -rf "%t"
 // RUN: mkdir "%t"
-// RUN: not env TMPDIR="%t" TEMP="%t" TMP="%t" RC_DEBUG_OPTIONS=1 %clang -fsyntax-only "%s" 2>&1 | FileCheck "%s"
+// RUN: cp "%s" "%t/crash report spaces.c"
+// RUN: not env TMPDIR="%t" TEMP="%t" TMP="%t" RC_DEBUG_OPTIONS=1 %clang -fsyntax-only "%t/crash report spaces.c" 2>&1 | FileCheck "%s"
 // RUN: cat "%t/crash report spaces"-*.c | FileCheck --check-prefix=CHECKSRC "%s"
 // RUN: cat "%t/crash report spaces"-*.sh | FileCheck --check-prefix=CHECKSH "%s"
 // REQUIRES: crash-recovery
diff --git a/test/Driver/darwin-sdk-vs-os-version.c b/test/Driver/darwin-sdk-vs-os-version.c
new file mode 100644
index 000000000000..391f4d5a7305
--- /dev/null
+++ b/test/Driver/darwin-sdk-vs-os-version.c
@@ -0,0 +1,10 @@
+// REQUIRES: system-darwin
+
+// Ensure that we never pick a version that's based on the SDK that's newer than
+// the system version:
+// RUN: rm -rf %t/SDKs/MacOSX10.99.99.sdk
+// RUN: mkdir -p %t/SDKs/MacOSX10.99.99.sdk
+// RUN: %clang -target x86_64-apple-darwin -isysroot %t/SDKs/MacOSX10.99.99.sdk %s -### 2>&1 \
+// RUN:   | FileCheck --check-prefix=CHECK-MACOSX-SYSTEM-VERSION %s
+
+// CHECK-MACOSX-SYSTEM-VERSION-NOT: 10.99.99"
diff --git a/test/Driver/fuchsia.c b/test/Driver/fuchsia.c
index 58c2bbe533c9..64d31cc0d38c 100644
--- a/test/Driver/fuchsia.c
+++ b/test/Driver/fuchsia.c
@@ -12,7 +12,7 @@
 // CHECK: Scrt1.o
 // CHECK-NOT: crti.o
 // CHECK-NOT: crtbegin.o
-// CHECK: "-L[[SYSROOT]]/lib"
+// CHECK: "-L[[SYSROOT]]{{/|\\\\}}lib"
 // CHECK: "{{.*[/\\]}}libclang_rt.builtins-x86_64.a"
 // CHECK: "-lc"
 // CHECK-NOT: crtend.o
diff --git a/test/Driver/fuchsia.cpp b/test/Driver/fuchsia.cpp
index 4490f94d0715..4c3f7d29acd8 100644
--- a/test/Driver/fuchsia.cpp
+++ b/test/Driver/fuchsia.cpp
@@ -1,9 +1,10 @@
 // RUN: %clangxx %s -### -no-canonical-prefixes --target=x86_64-unknown-fuchsia \
 // RUN:     --sysroot=%S/platform 2>&1 -fuse-ld=ld | FileCheck %s
 // CHECK: {{.*}}clang{{.*}}" "-cc1"
+// CHECK: "-triple" "x86_64-fuchsia"
 // CHECK: "-fuse-init-array"
 // CHECK: "-isysroot" "[[SYSROOT:[^"]+]]"
-// CHECK: "-internal-isystem" "[[SYSROOT]]{{/|\\\\}}include{{/|\\\\}}c++{{/|\\\\}}v1"
+// CHECK: "-internal-isystem" "{{.*[/\\]}}x86_64-fuchsia{{/|\\\\}}include{{/|\\\\}}c++{{/|\\\\}}v1"
 // CHECK: "-internal-externc-isystem" "[[SYSROOT]]{{/|\\\\}}include"
 // CHECK: {{.*}}lld{{.*}}" "-flavor" "gnu"
 // CHECK: "--sysroot=[[SYSROOT]]"
@@ -13,7 +14,7 @@
 // CHECK: Scrt1.o
 // CHECK-NOT: crti.o
 // CHECK-NOT: crtbegin.o
-// CHECK: "-L[[SYSROOT]]/lib"
+// CHECK: "-L[[SYSROOT]]{{/|\\\\}}lib"
 // CHECK: "-lc++" "-lc++abi" "-lunwind" "-lm"
 // CHECK: "{{.*[/\\]}}libclang_rt.builtins-x86_64.a"
 // CHECK: "-lc"
diff --git a/test/Driver/mips-features.c b/test/Driver/mips-features.c
index 69fc20e1f245..b228a2d5781d 100644
--- a/test/Driver/mips-features.c
+++ b/test/Driver/mips-features.c
@@ -70,6 +70,18 @@
 // RUN:   | FileCheck --check-prefix=CHECK-NOMMSA %s
 // CHECK-NOMMSA: "-target-feature" "-msa"
 //
+// -mmt
+// RUN: %clang -target mips-linux-gnu -### -c %s \
+// RUN:     -mno-mt -mmt 2>&1 \
+// RUN:   | FileCheck --check-prefix=CHECK-MMT %s
+// CHECK-MMT: "-target-feature" "+mt"
+//
+// -mno-mt
+// RUN: %clang -target mips-linux-gnu -### -c %s \
+// RUN:     -mmt -mno-mt 2>&1 \
+// RUN:   | FileCheck --check-prefix=CHECK-NOMMT %s
+// CHECK-NOMMT: "-target-feature" "-mt"
+//
 // -modd-spreg
 // RUN: %clang -target mips-linux-gnu -### -c %s -mno-odd-spreg -modd-spreg 2>&1 \
 // RUN:   | FileCheck --check-prefix=CHECK-MODDSPREG %s
@@ -247,3 +259,14 @@
 // RUN:   | FileCheck --check-prefix=CHECK-IMG-SINGLEFLOAT-FPXX %s
 // CHECK-IMG-SINGLEFLOAT-FPXX: "-target-feature" "+single-float"
 // CHECK-IMG-SINGLEFLOAT-FPXX: "-target-feature" "+fpxx"
+
+// -mlong-call
+// RUN: %clang -target mips-img-linux-gnu -### -c %s -mlong-calls 2>&1 \
+// RUN:   | FileCheck --check-prefix=LONG-CALLS-ON %s
+// RUN: %clang -target mips-img-linux-gnu -### -c %s -mno-long-calls 2>&1 \
+// RUN:   | FileCheck --check-prefix=LONG-CALLS-OFF %s
+// RUN: %clang -target mips-img-linux-gnu -### -c %s 2>&1 \
+// RUN:   | FileCheck --check-prefix=LONG-CALLS-DEF %s
+// LONG-CALLS-ON: "-target-feature" "+long-calls"
+// LONG-CALLS-OFF: "-target-feature" "-long-calls"
+// LONG-CALLS-DEF-NOT: "long-calls"
diff --git a/test/FixIt/fixit-add-synthesize-to-property.m b/test/FixIt/fixit-add-synthesize-to-property.m
new file mode 100644
index 000000000000..19b2f4b73f01
--- /dev/null
+++ b/test/FixIt/fixit-add-synthesize-to-property.m
@@ -0,0 +1,14 @@
+// RUN: %clang_cc1 -fdiagnostics-parseable-fixits %s 2>&1 | FileCheck %s
+
+@protocol P1
+
+@property int prop;
+
+@end
+
+@interface I <P1>
+
+@end
+
+@implementation I
+@end // CHECK: fix-it:{{.*}}:{[[@LINE]]:1-[[@LINE]]:1}:"@synthesize prop;\n\n"
diff --git a/test/Import/direct/Inputs/S.c b/test/Import/direct/Inputs/S.c
new file mode 100644
index 000000000000..b0876d27df44
--- /dev/null
+++ b/test/Import/direct/Inputs/S.c
@@ -0,0 +1,3 @@
+struct S {
+  int a;
+};
diff --git a/test/Import/direct/test.c b/test/Import/direct/test.c
new file mode 100644
index 000000000000..5fac24c015dc
--- /dev/null
+++ b/test/Import/direct/test.c
@@ -0,0 +1,5 @@
+// RUN: clang-import-test -direct -import %S/Inputs/S.c -expression %s
+void expr() {
+  struct S MyS;
+  MyS.a = 3;
+}
diff --git a/test/Import/enum/Inputs/S.cpp b/test/Import/enum/Inputs/S.cpp
new file mode 100644
index 000000000000..8506c68512db
--- /dev/null
+++ b/test/Import/enum/Inputs/S.cpp
@@ -0,0 +1,4 @@
+enum E {
+  a = 1,
+  b = 2
+};
diff --git a/test/Import/enum/test.cpp b/test/Import/enum/test.cpp
new file mode 100644
index 000000000000..aa68063cb207
--- /dev/null
+++ b/test/Import/enum/test.cpp
@@ -0,0 +1,4 @@
+// RUN: clang-import-test -import %S/Inputs/S.cpp -expression %s
+void expr() {
+  static_assert(E::a + E::b == 3);
+}
diff --git a/test/Import/import-overrides/Inputs/Hierarchy.cpp b/test/Import/import-overrides/Inputs/Hierarchy.cpp
new file mode 100644
index 000000000000..3f91be765d18
--- /dev/null
+++ b/test/Import/import-overrides/Inputs/Hierarchy.cpp
@@ -0,0 +1,9 @@
+class Base {
+public:
+  virtual void foo() {}
+};
+
+class Derived : public Base {
+public:
+  void foo() override {}
+};
diff --git a/test/Import/import-overrides/test.cpp b/test/Import/import-overrides/test.cpp
new file mode 100644
index 000000000000..ded29cded2ee
--- /dev/null
+++ b/test/Import/import-overrides/test.cpp
@@ -0,0 +1,7 @@
+// RUN: clang-import-test -dump-ast -import %S/Inputs/Hierarchy.cpp -expression %s | FileCheck %s
+
+// CHECK: Overrides:{{.*}}Base::foo
+
+void foo() {
+  Derived d;
+}
diff --git a/test/Index/Core/index-source-invalid-name.cpp b/test/Index/Core/index-source-invalid-name.cpp
new file mode 100644
index 000000000000..1b4b059cd1b3
--- /dev/null
+++ b/test/Index/Core/index-source-invalid-name.cpp
@@ -0,0 +1,13 @@
+// RUN: c-index-test core -print-source-symbols -- %s -std=c++1z -target x86_64-apple-macosx10.7 | FileCheck %s
+
+namespace rdar32474406 {
+// CHECK: [[@LINE+1]]:6 | function/C | foo | c:@N@rdar32474406@F@foo# | __ZN12rdar324744063fooEv | Decl,RelChild | rel: 1
+void foo();
+// CHECK: [[@LINE+1]]:16 | type-alias/C | Func_t | c:index-source-invalid-name.cpp@N@rdar32474406@T@Func_t | <no-cgname> | Def,RelChild | rel: 1
+typedef void (*Func_t)();
+// CHECK: [[@LINE+4]]:1 | type-alias/C | Func_t | c:index-source-invalid-name.cpp@N@rdar32474406@T@Func_t | <no-cgname> | Ref,RelCont | rel: 1
+// CHECK-NEXT: RelCont | rdar32474406 | c:@N@rdar32474406
+// CHECK: [[@LINE+2]]:14 | function/C | foo | c:@N@rdar32474406@F@foo# | __ZN12rdar324744063fooEv | Ref,RelCont | rel: 1
+// CHECK-NEXT: RelCont | rdar32474406 | c:@N@rdar32474406
+Func_t[] = { foo }; // invalid decomposition
+}
diff --git a/test/Index/Core/index-source.cpp b/test/Index/Core/index-source.cpp
index 6d20fdd48e50..4864d6cf0150 100644
--- a/test/Index/Core/index-source.cpp
+++ b/test/Index/Core/index-source.cpp
@@ -496,3 +496,19 @@ void localStructuredBindingAndRef() {
 }
 
 }
+
+namespace rd33122110 {
+
+struct Outer {
+    template<typename T>
+    struct Nested { };
+};
+
+}
+
+template<>
+struct rd33122110::Outer::Nested<int>;
+// CHECK: [[@LINE-1]]:8 | namespace/C++ | rd33122110 | c:@N@rd33122110 | <no-cgname> | Ref,RelCont | rel: 1
+// CHECK-NEXT: RelCont | Nested | c:@N@rd33122110@S@Outer@S@Nested>#I
+// CHECK: [[@LINE-3]]:20 | struct/C++ | Outer | c:@N@rd33122110@S@Outer | <no-cgname> | Ref,RelCont | rel: 1
+// CHECK-NEXT: RelCont | Nested | c:@N@rd33122110@S@Outer@S@Nested>#I
diff --git a/test/Index/Core/index-source.m b/test/Index/Core/index-source.m
index a64c34ad2ac1..c911973a70d6 100644
--- a/test/Index/Core/index-source.m
+++ b/test/Index/Core/index-source.m
@@ -413,3 +413,28 @@ void classReceivers() {
   (void)ClassReceivers.implicit;
 // CHECK: [[@LINE-1]]:9 | class/ObjC | ClassReceivers | c:objc(cs)ClassReceivers | _OBJC_CLASS_$_ClassReceivers | Ref,RelCont | rel: 1
 }
+
+@interface ImplicitProperties
+
+- (int)implicit;
+- (void)setImplicit:(int)x;
+
++ (int)classImplicit;
++ (void)setClassImplicit:(int)y;
+
+@end
+
+void testImplicitProperties(ImplicitProperties *c) {
+  c.implicit = 0;
+// CHECK: [[@LINE-1]]:5 | instance-method/ObjC | setImplicit: | c:objc(cs)ImplicitProperties(im)setImplicit: | -[ImplicitProperties setImplicit:] | Ref,Call,Dyn,RelRec,RelCall,RelCont | rel: 2
+// CHECK-NEXT: RelCall,RelCont | testImplicitProperties | c:@F@testImplicitProperties
+  c.implicit;
+// CHECK: [[@LINE-1]]:5 | instance-method/ObjC | implicit | c:objc(cs)ImplicitProperties(im)implicit | -[ImplicitProperties implicit] | Ref,Call,Dyn,RelRec,RelCall,RelCont | rel: 2
+// CHECK-NEXT: RelCall,RelCont | testImplicitProperties | c:@F@testImplicitProperties
+  ImplicitProperties.classImplicit = 1;
+// CHECK: [[@LINE-1]]:22 | class-method/ObjC | setClassImplicit: | c:objc(cs)ImplicitProperties(cm)setClassImplicit: | +[ImplicitProperties setClassImplicit:] | Ref,Call,RelCall,RelCont | rel: 1
+// CHECK-NEXT: RelCall,RelCont | testImplicitProperties | c:@F@testImplicitProperties
+  ImplicitProperties.classImplicit;
+// CHECK: [[@LINE-1]]:22 | class-method/ObjC | classImplicit | c:objc(cs)ImplicitProperties(cm)classImplicit | +[ImplicitProperties classImplicit] | Ref,Call,RelCall,RelCont | rel: 1
+// CHECK-NEXT: RelCall,RelCont | testImplicitProperties | c:@F@testImplicitProperties
+}
diff --git a/test/Index/Inputs/empty.dia b/test/Index/Inputs/empty.dia
new file mode 100644
index 000000000000..e69de29bb2d1
--- /dev/null
+++ b/test/Index/Inputs/empty.dia
diff --git a/test/Index/pipe-size.cl b/test/Index/pipe-size.cl
index d07d7067da8c..cfe97935ee63 100644
--- a/test/Index/pipe-size.cl
+++ b/test/Index/pipe-size.cl
@@ -11,6 +11,6 @@ __kernel void testPipe( pipe int test )
     // SPIR: store i32 4, i32* %s, align 4
     // SPIR64: store %opencl.pipe_t addrspace(1)* %test, %opencl.pipe_t addrspace(1)** %test.addr, align 8
     // SPIR64: store i32 8, i32* %s, align 4
-    // AMD: store %opencl.pipe_t addrspace(1)* %test, %opencl.pipe_t addrspace(1)* addrspace(5)* %test.addr, align 4
+    // AMD: store %opencl.pipe_t addrspace(1)* %test, %opencl.pipe_t addrspace(1)* addrspace(5)* %test.addr, align 8
     // AMD: store i32 8, i32 addrspace(5)* %s, align 4
 }
diff --git a/test/Index/print-type-declaration.cpp b/test/Index/print-type-declaration.cpp
index 31c0a73fcd06..a0953d1e5638 100644
--- a/test/Index/print-type-declaration.cpp
+++ b/test/Index/print-type-declaration.cpp
@@ -7,6 +7,13 @@ int main()
   auto b = a;
 }
 
+enum RegularEnum {};
+
+enum class ScopedEnum {};
+
 // RUN: c-index-test -test-print-type-declaration -std=c++11 %s | FileCheck %s
 // CHECK: VarDecl=a:6:8 (Definition) [typedeclaration=Test] [typekind=Record]
 // CHECK: VarDecl=b:7:8 (Definition) [typedeclaration=Test] [typekind=Record]
+// CHECK: EnumDecl=RegularEnum:10:6 (Definition) [typedeclaration=RegularEnum] [typekind=Enum]
+// CHECK: EnumDecl=ScopedEnum:12:12 (Definition) (scoped) [typedeclaration=ScopedEnum] [typekind=Enum]
+
diff --git a/test/Index/read-empty-diags.test b/test/Index/read-empty-diags.test
new file mode 100644
index 000000000000..cef751c920ab
--- /dev/null
+++ b/test/Index/read-empty-diags.test
@@ -0,0 +1,2 @@
+// RUN: not c-index-test -read-diagnostics %S/Inputs/empty.dia 2>&1 | FileCheck %s
+// CHECK: Trouble deserializing file (Invalid File): Invalid diagnostics signature
diff --git a/test/Index/usrs.m b/test/Index/usrs.m
index 92c3a3fafee2..d0a860e1afad 100644
--- a/test/Index/usrs.m
+++ b/test/Index/usrs.m
@@ -119,7 +119,7 @@ int test_multi_declaration(void) {
 // CHECK: usrs.m c:@SA@MyStruct Extent=[15:9 - 18:2]
 // CHECK: usrs.m c:@SA@MyStruct@FI@wa Extent=[16:3 - 16:9]
 // CHECK: usrs.m c:@SA@MyStruct@FI@moo Extent=[17:3 - 17:10]
-// CHECK: usrs.m c:usrs.m@T@MyStruct Extent=[15:1 - 18:11]
+// CHECK: usrs.m c:@T@MyStruct Extent=[15:1 - 18:11]
 // CHECK: usrs.m c:@E@Pizza Extent=[20:1 - 23:2]
 // CHECK: usrs.m c:@E@Pizza@CHEESE Extent=[21:3 - 21:9]
 // CHECK: usrs.m c:@E@Pizza@MUSHROOMS Extent=[22:3 - 22:12]
diff --git a/test/Misc/find-diagnostic-id.c b/test/Misc/find-diagnostic-id.c
new file mode 100644
index 000000000000..bef66178f96f
--- /dev/null
+++ b/test/Misc/find-diagnostic-id.c
@@ -0,0 +1,5 @@
+// RUN: diagtool find-diagnostic-id warn_unused_variable | FileCheck %s
+// RUN: not diagtool find-diagnostic-id warn_unused_vars 2>&1 | FileCheck --check-prefix=ERROR %s
+
+// CHECK: {{^[0-9]+$}}
+// ERROR: error: invalid diagnostic 'warn_unused_vars'
diff --git a/test/Modules/missing-flag.cpp b/test/Modules/missing-flag.cpp
new file mode 100644
index 000000000000..560183be59e8
--- /dev/null
+++ b/test/Modules/missing-flag.cpp
@@ -0,0 +1,4 @@
+// RUN: not %clang_cc1 -x c++-module-map %s -emit-module -fmodule-name=Foo -o %t 2>&1 | FileCheck %s
+// CHECK: module compilation requires '-fmodules'
+module Foo {}
+#pragma clang module contents
diff --git a/test/Modules/odr_hash.cpp b/test/Modules/odr_hash.cpp
index 6e01e989a374..ee45ae529961 100644
--- a/test/Modules/odr_hash.cpp
+++ b/test/Modules/odr_hash.cpp
@@ -1339,6 +1339,84 @@ Bravo<char> golf;
 #endif
 }
 
+namespace Friend {
+#if defined(FIRST)
+struct T1 {};
+struct S1 {
+  friend class T1;
+};
+#elif defined(SECOND)
+struct T1 {};
+struct S1 {
+  friend T1;
+};
+#else
+S1 s1;
+// expected-error@second.h:* {{'Friend::S1' has different definitions in different modules; first difference is definition in module 'SecondModule' found friend 'Friend::T1'}}
+// expected-note@first.h:* {{but in 'FirstModule' found friend 'class T1'}}
+#endif
+
+#if defined(FIRST)
+struct T2 {};
+struct S2 {
+  friend class T2;
+};
+#elif defined(SECOND)
+struct T2 {};
+struct S2 {
+  friend struct T2;
+};
+#else
+S2 s2;
+// expected-error@second.h:* {{'Friend::S2' has different definitions in different modules; first difference is definition in module 'SecondModule' found friend 'struct T2'}}
+// expected-note@first.h:* {{but in 'FirstModule' found friend 'class T2'}}
+#endif
+
+#if defined(FIRST)
+struct T3 {};
+struct S3 {
+  friend const T3;
+};
+#elif defined(SECOND)
+struct T3 {};
+struct S3 {
+  friend T3;
+};
+#else
+S3 s3;
+// expected-error@second.h:* {{'Friend::S3' has different definitions in different modules; first difference is definition in module 'SecondModule' found friend 'Friend::T3'}}
+// expected-note@first.h:* {{but in 'FirstModule' found friend 'const Friend::T3'}}
+#endif
+
+#if defined(FIRST)
+struct T4 {};
+struct S4 {
+  friend T4;
+};
+#elif defined(SECOND)
+struct S4 {
+  friend void T4();
+};
+#else
+S4 s4;
+// expected-error@second.h:* {{'Friend::S4' has different definitions in different modules; first difference is definition in module 'SecondModule' found friend function}}
+// expected-note@first.h:* {{but in 'FirstModule' found friend class}}
+#endif
+
+#if defined(FIRST)
+struct S5 {
+  friend void T5a();
+};
+#elif defined(SECOND)
+struct S5 {
+  friend void T5b();
+};
+#else
+S5 s5;
+// expected-error@second.h:* {{'Friend::S5' has different definitions in different modules; first difference is definition in module 'SecondModule' found friend function 'T5b'}}
+// expected-note@first.h:* {{but in 'FirstModule' found friend function 'T5a'}}
+#endif
+}
 // Interesting cases that should not cause errors.  struct S should not error
 // while struct T should error at the access specifier mismatch at the end.
 namespace AllDecls {
diff --git a/test/Modules/preprocess-build.cpp b/test/Modules/preprocess-build.cpp
index bf9f16162be2..36a0740a0a76 100644
--- a/test/Modules/preprocess-build.cpp
+++ b/test/Modules/preprocess-build.cpp
@@ -1,4 +1,4 @@
-// RUN: %clang_cc1 -std=c++1z %s -verify
+// RUN: %clang_cc1 -std=c++1z -fmodules %s -verify
 
 #pragma clang module build baz
   module baz {}
diff --git a/test/Modules/relative-dep-gen.cpp b/test/Modules/relative-dep-gen.cpp
index e58dd6eb3596..bfa9471859b8 100644
--- a/test/Modules/relative-dep-gen.cpp
+++ b/test/Modules/relative-dep-gen.cpp
@@ -2,17 +2,17 @@
 // RUN: rm -rf %t
 // RUN: mkdir %t
 //
-// RUN: %clang_cc1 -cc1 -fno-implicit-modules -fmodule-name=relative-dep-gen -emit-module -x c++ Inputs/relative-dep-gen.modulemap -dependency-file %t/build.d -MT mod.pcm -o %t/mod.pcm
-// RUN: %clang_cc1 -cc1 -fno-implicit-modules -fmodule-map-file=Inputs/relative-dep-gen.modulemap -fmodule-file=%t/mod.pcm -dependency-file %t/use-explicit.d -MT use.o relative-dep-gen.cpp -fsyntax-only
-// RUN: %clang_cc1 -cc1 -fno-implicit-modules -fmodule-map-file=Inputs/relative-dep-gen.modulemap -dependency-file %t/use-implicit.d relative-dep-gen.cpp -MT use.o -fsyntax-only
+// RUN: %clang_cc1 -fmodules -fno-implicit-modules -fmodule-name=relative-dep-gen -emit-module -x c++ Inputs/relative-dep-gen.modulemap -dependency-file %t/build.d -MT mod.pcm -o %t/mod.pcm
+// RUN: %clang_cc1 -fmodules -fno-implicit-modules -fmodule-map-file=Inputs/relative-dep-gen.modulemap -fmodule-file=%t/mod.pcm -dependency-file %t/use-explicit.d -MT use.o relative-dep-gen.cpp -fsyntax-only
+// RUN: %clang_cc1 -fmodules -fmodules-cache-path=%t -fmodule-map-file=Inputs/relative-dep-gen.modulemap -dependency-file %t/use-implicit.d relative-dep-gen.cpp -MT use.o -fsyntax-only
 //
 // RUN: FileCheck --check-prefix=CHECK-BUILD %s < %t/build.d
-// RUN: FileCheck --check-prefix=CHECK-USE %s < %t/use-explicit.d
-// RUN: FileCheck --check-prefix=CHECK-USE %s < %t/use-implicit.d
+// RUN: FileCheck --check-prefix=CHECK-USE --check-prefix=CHECK-EXPLICIT %s < %t/use-explicit.d
+// RUN: FileCheck --check-prefix=CHECK-USE --check-prefix=CHECK-IMPLICIT %s < %t/use-implicit.d
 //
-// RUN: %clang_cc1 -cc1 -fno-implicit-modules -fmodule-name=relative-dep-gen -emit-module -x c++ Inputs/relative-dep-gen-cwd.modulemap -dependency-file %t/build-cwd.d -MT mod.pcm -o %t/mod-cwd.pcm -fmodule-map-file-home-is-cwd
-// RUN: %clang_cc1 -cc1 -fno-implicit-modules -fmodule-map-file=Inputs/relative-dep-gen-cwd.modulemap -fmodule-file=%t/mod-cwd.pcm -dependency-file %t/use-explicit-cwd.d -MT use.o relative-dep-gen.cpp -fsyntax-only -fmodule-map-file-home-is-cwd
-// RUN: %clang_cc1 -cc1 -fno-implicit-modules -fmodule-map-file=Inputs/relative-dep-gen-cwd.modulemap -dependency-file %t/use-implicit-cwd.d relative-dep-gen.cpp -MT use.o -fsyntax-only -fmodule-map-file-home-is-cwd
+// RUN: %clang_cc1 -fmodules -fno-implicit-modules -fmodule-name=relative-dep-gen -emit-module -x c++ Inputs/relative-dep-gen-cwd.modulemap -dependency-file %t/build-cwd.d -MT mod.pcm -o %t/mod-cwd.pcm -fmodule-map-file-home-is-cwd
+// RUN: %clang_cc1 -fmodules -fno-implicit-modules -fmodule-map-file=Inputs/relative-dep-gen-cwd.modulemap -fmodule-file=%t/mod-cwd.pcm -dependency-file %t/use-explicit-cwd.d -MT use.o relative-dep-gen.cpp -fsyntax-only -fmodule-map-file-home-is-cwd
+// RUN: %clang_cc1 -fmodules -fmodules-cache-path=%t -fmodule-map-file=Inputs/relative-dep-gen-cwd.modulemap -dependency-file %t/use-implicit-cwd.d relative-dep-gen.cpp -MT use.o -fsyntax-only -fmodule-map-file-home-is-cwd
 //
 // RUN: FileCheck --check-prefix=CHECK-BUILD %s < %t/build-cwd.d
 // RUN: FileCheck --check-prefix=CHECK-USE %s < %t/use-explicit-cwd.d
@@ -23,7 +23,7 @@
 // RUN: cp %S/Inputs/relative-dep-gen-1.h %t/Inputs
 // RUN: cp %s %t
 // RUN: cd %t
-// RUN: %clang_cc1 -cc1 -fno-implicit-modules -fmodule-file=%t/mod-cwd.pcm -dependency-file %t/use-explicit-no-map-cwd.d -MT use.o relative-dep-gen.cpp -fsyntax-only -fmodule-map-file-home-is-cwd
+// RUN: %clang_cc1 -fmodules -fno-implicit-modules -fmodule-file=%t/mod-cwd.pcm -dependency-file %t/use-explicit-no-map-cwd.d -MT use.o relative-dep-gen.cpp -fsyntax-only -fmodule-map-file-home-is-cwd
 // RUN: cat %t/use-explicit-no-map-cwd.d
 // RUN: FileCheck --check-prefix=CHECK-USE %s < %t/use-explicit-no-map-cwd.d
 
@@ -35,4 +35,5 @@
 // CHECK-BUILD:   {{[ \t]}}Inputs/relative-dep-gen-2.h
 // CHECK-USE: use.o:
 // CHECK-USE-DAG:   {{[ \t]}}relative-dep-gen.cpp
-// CHECK-USE-DAG:   {{[ \t]}}Inputs{{[/\\]}}relative-dep-gen-1.h
+// CHECK-EXPLICIT-DAG:   mod.pcm
+// CHECK-IMPLICIT-DAG:   {{[ \t]}}Inputs{{[/\\]}}relative-dep-gen-1.h
diff --git a/test/OpenMP/taskloop_ast_print.cpp b/test/OpenMP/taskloop_ast_print.cpp
index 36f90c95dd33..dd0af85916cf 100644
--- a/test/OpenMP/taskloop_ast_print.cpp
+++ b/test/OpenMP/taskloop_ast_print.cpp
@@ -20,7 +20,7 @@ T tmain(T argc) {
 // CHECK-NEXT: for (int i = 0; i < 2; ++i)
 // CHECK-NEXT: a = 2;
 #pragma omp parallel
-#pragma omp taskloop private(argc, b), firstprivate(c, d), lastprivate(d, f) collapse(N) shared(g) if (c) final(d) mergeable priority(f) nogroup num_tasks(N) reduction(min:a)
+#pragma omp taskloop private(argc, b), firstprivate(c, d), lastprivate(d, f) collapse(N) shared(g) if (c) final(d) mergeable priority(f) nogroup num_tasks(N)
   for (int i = 0; i < 2; ++i)
     for (int j = 0; j < 2; ++j)
       for (int j = 0; j < 2; ++j)
@@ -33,7 +33,7 @@ T tmain(T argc) {
           for (int j = 0; j < 2; ++j)
             foo();
   // CHECK-NEXT: #pragma omp parallel
-  // CHECK-NEXT: #pragma omp taskloop private(argc,b) firstprivate(c,d) lastprivate(d,f) collapse(N) shared(g) if(c) final(d) mergeable priority(f) nogroup num_tasks(N) reduction(min: a)
+  // CHECK-NEXT: #pragma omp taskloop private(argc,b) firstprivate(c,d) lastprivate(d,f) collapse(N) shared(g) if(c) final(d) mergeable priority(f) nogroup num_tasks(N)
   // CHECK-NEXT: for (int i = 0; i < 2; ++i)
   // CHECK-NEXT: for (int j = 0; j < 2; ++j)
   // CHECK-NEXT: for (int j = 0; j < 2; ++j)
@@ -53,8 +53,8 @@ int main(int argc, char **argv) {
   int b = argc, c, d, e, f, g;
   static int a;
 // CHECK: static int a;
-#pragma omp taskloop if(taskloop: a) default(none) shared(a) final(b) priority(5) nogroup num_tasks(argc) reduction(*: g)
-  // CHECK-NEXT: #pragma omp taskloop if(taskloop: a) default(none) shared(a) final(b) priority(5) nogroup num_tasks(argc) reduction(*: g)
+#pragma omp taskloop if(taskloop: a) default(none) shared(a) final(b) priority(5) num_tasks(argc) reduction(*: g)
+  // CHECK-NEXT: #pragma omp taskloop if(taskloop: a) default(none) shared(a) final(b) priority(5) num_tasks(argc) reduction(*: g)
   for (int i = 0; i < 2; ++i)
     a = 2;
 // CHECK-NEXT: for (int i = 0; i < 2; ++i)
diff --git a/test/OpenMP/taskloop_codegen.cpp b/test/OpenMP/taskloop_codegen.cpp
index bc7367ce0216..b1b45455802a 100644
--- a/test/OpenMP/taskloop_codegen.cpp
+++ b/test/OpenMP/taskloop_codegen.cpp
@@ -8,6 +8,7 @@
 // CHECK-LABEL: @main
 int main(int argc, char **argv) {
 // CHECK: [[GTID:%.+]] = call i32 @__kmpc_global_thread_num(%ident_t* [[DEFLOC:@.+]])
+// CHECK: call void @__kmpc_taskgroup(%ident_t* [[DEFLOC]], i32 [[GTID]])
 // CHECK: [[TASKV:%.+]] = call i8* @__kmpc_omp_task_alloc(%ident_t* [[DEFLOC]], i32 [[GTID]], i32 33, i64 72, i64 1, i32 (i32, i8*)* bitcast (i32 (i32, [[TDP_TY:%.+]]*)* [[TASK1:@.+]] to i32 (i32, i8*)*))
 // CHECK: [[TASK:%.+]] = bitcast i8* [[TASKV]] to [[TDP_TY]]*
 // CHECK: [[TASK_DATA:%.+]] = getelementptr inbounds [[TDP_TY]], [[TDP_TY]]* [[TASK]], i32 0, i32 0
@@ -19,6 +20,7 @@ int main(int argc, char **argv) {
 // CHECK: store i64 1, i64* [[ST]],
 // CHECK: [[ST_VAL:%.+]] = load i64, i64* [[ST]],
 // CHECK: call void @__kmpc_taskloop(%ident_t* [[DEFLOC]], i32 [[GTID]], i8* [[TASKV]], i32 1, i64* [[DOWN]], i64* [[UP]], i64 [[ST_VAL]], i32 0, i32 0, i64 0, i8* null)
+// CHECK: call void @__kmpc_end_taskgroup(%ident_t* [[DEFLOC]], i32 [[GTID]])
 #pragma omp taskloop priority(argc)
   for (int i = 0; i < 10; ++i)
     ;
@@ -33,10 +35,11 @@ int main(int argc, char **argv) {
 // CHECK: store i64 1, i64* [[ST]],
 // CHECK: [[ST_VAL:%.+]] = load i64, i64* [[ST]],
 // CHECK: [[GRAINSIZE:%.+]] = zext i32 %{{.+}} to i64
-// CHECK: call void @__kmpc_taskloop(%ident_t* [[DEFLOC]], i32 [[GTID]], i8* [[TASKV]], i32 1, i64* [[DOWN]], i64* [[UP]], i64 [[ST_VAL]], i32 1, i32 1, i64 [[GRAINSIZE]], i8* null)
+// CHECK: call void @__kmpc_taskloop(%ident_t* [[DEFLOC]], i32 [[GTID]], i8* [[TASKV]], i32 1, i64* [[DOWN]], i64* [[UP]], i64 [[ST_VAL]], i32 0, i32 1, i64 [[GRAINSIZE]], i8* null)
 #pragma omp taskloop nogroup grainsize(argc)
   for (int i = 0; i < 10; ++i)
     ;
+// CHECK: call void @__kmpc_taskgroup(%ident_t* [[DEFLOC]], i32 [[GTID]])
 // CHECK: [[TASKV:%.+]] = call i8* @__kmpc_omp_task_alloc(%ident_t* [[DEFLOC]], i32 [[GTID]], i32 1, i64 72, i64 24, i32 (i32, i8*)* bitcast (i32 (i32, [[TDP_TY:%.+]]*)* [[TASK3:@.+]] to i32 (i32, i8*)*))
 // CHECK: [[TASK:%.+]] = bitcast i8* [[TASKV]] to [[TDP_TY]]*
 // CHECK: [[TASK_DATA:%.+]] = getelementptr inbounds [[TDP_TY]], [[TDP_TY]]* [[TASK]], i32 0, i32 0
@@ -50,6 +53,7 @@ int main(int argc, char **argv) {
 // CHECK: store i64 1, i64* [[ST]],
 // CHECK: [[ST_VAL:%.+]] = load i64, i64* [[ST]],
 // CHECK: call void @__kmpc_taskloop(%ident_t* [[DEFLOC]], i32 [[GTID]], i8* [[TASKV]], i32 [[IF_INT]], i64* [[DOWN]], i64* [[UP]], i64 [[ST_VAL]], i32 0, i32 2, i64 4, i8* null)
+// CHECK: call void @__kmpc_end_taskgroup(%ident_t* [[DEFLOC]], i32 [[GTID]])
   int i;
 #pragma omp taskloop if(argc) shared(argc, argv) collapse(2) num_tasks(4)
   for (i = 0; i < argc; ++i)
diff --git a/test/OpenMP/taskloop_reduction_messages.cpp b/test/OpenMP/taskloop_reduction_messages.cpp
new file mode 100644
index 000000000000..7295075c3cb8
--- /dev/null
+++ b/test/OpenMP/taskloop_reduction_messages.cpp
@@ -0,0 +1,331 @@
+// RUN: %clang_cc1 -verify -fopenmp -ferror-limit 150 -o - %s
+// RUN: %clang_cc1 -verify -fopenmp -std=c++98 -ferror-limit 150 -o - %s
+// RUN: %clang_cc1 -verify -fopenmp -std=c++11 -ferror-limit 150 -o - %s
+
+void foo() {
+}
+
+bool foobool(int argc) {
+  return argc;
+}
+
+void foobar(int &ref) {
+#pragma omp taskloop reduction(+:ref)
+  for (int i = 0; i < 10; ++i)
+    foo();
+}
+
+struct S1; // expected-note {{declared here}} expected-note 4 {{forward declaration of 'S1'}}
+extern S1 a;
+class S2 {
+  mutable int a;
+  S2 &operator+(const S2 &arg) { return (*this); } // expected-note 3 {{implicitly declared private here}}
+
+public:
+  S2() : a(0) {}
+  S2(S2 &s2) : a(s2.a) {}
+  static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
+  static const float S2sc;
+};
+const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+S2 b;                     // expected-note 3 {{'b' defined here}}
+const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
+class S3 {
+  int a;
+
+public:
+  int b;
+  S3() : a(0) {}
+  S3(const S3 &s3) : a(s3.a) {}
+  S3 operator+(const S3 &arg1) { return arg1; }
+};
+int operator+(const S3 &arg1, const S3 &arg2) { return 5; }
+S3 c;               // expected-note 3 {{'c' defined here}}
+const S3 ca[5];     // expected-note 2 {{'ca' defined here}}
+extern const int f; // expected-note 4 {{'f' declared here}}
+class S4 {
+  int a;
+  S4(); // expected-note {{implicitly declared private here}}
+  S4(const S4 &s4);
+  S4 &operator+(const S4 &arg) { return (*this); }
+
+public:
+  S4(int v) : a(v) {}
+};
+S4 &operator&=(S4 &arg1, S4 &arg2) { return arg1; }
+class S5 {
+  int a:32;
+  S5() : a(0) {} // expected-note {{implicitly declared private here}}
+  S5(const S5 &s5) : a(s5.a) {}
+  S5 &operator+(const S5 &arg);
+
+public:
+  S5(int v) : a(v) {}
+};
+class S6 { // expected-note 3 {{candidate function (the implicit copy assignment operator) not viable: no known conversion from 'int' to 'const S6' for 1st argument}}
+#if __cplusplus >= 201103L // C++11 or later
+// expected-note@-2 3 {{candidate function (the implicit move assignment operator) not viable}}
+#endif
+  int a;
+
+public:
+  S6() : a(6) {}
+  operator int() { return 6; }
+} o;
+
+struct S7 {
+  int a: 32;
+  S7() {
+#pragma omp taskloop reduction(+:a) // expected-error {{expected addressable reduction item for the task-based directives}}
+    for (int i = 0; i < 10; ++i)
+      ++a;
+  }
+};
+
+S3 h, k;
+#pragma omp threadprivate(h) // expected-note 2 {{defined as threadprivate or thread local}}
+
+template <class T>       // expected-note {{declared here}}
+T tmain(T argc) {
+  const T d = T();       // expected-note 4 {{'d' defined here}}
+  const T da[5] = {T()}; // expected-note 2 {{'da' defined here}}
+  T qa[5] = {T()};
+  T i;
+  T &j = i;                        // expected-note 4 {{'j' defined here}}
+  S3 &p = k;                       // expected-note 2 {{'p' defined here}}
+  const T &r = da[(int)i];         // expected-note 2 {{'r' defined here}}
+  T &q = qa[(int)i];               // expected-note 2 {{'q' defined here}}
+  T fl;
+#pragma omp taskloop reduction // expected-error {{expected '(' after 'reduction'}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction + // expected-error {{expected '(' after 'reduction'}} expected-warning {{extra tokens at the end of '#pragma omp taskloop' are ignored}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction( // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(- // expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction() // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(*) // expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected expression}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(\) // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(& : argc // expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(| : argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(|| : argc ? i : argc) // expected-error 2 {{expected variable name, array element or array section}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(foo : argc) //expected-error {{incorrect reduction identifier, expected one of '+', '-', '*', '&', '|', '^', '&&', '||', 'min' or 'max' or declare reduction for type 'float'}} expected-error {{incorrect reduction identifier, expected one of '+', '-', '*', '&', '|', '^', '&&', '||', 'min' or 'max' or declare reduction for type 'int'}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(&& : argc)
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(^ : T) // expected-error {{'T' does not refer to a value}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(+ : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 3 {{const-qualified list item cannot be reduction}} expected-error 2 {{'operator+' is a private member of 'S2'}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(min : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 4 {{arguments of OpenMP clause 'reduction' for 'min' or 'max' must be of arithmetic type}} expected-error 3 {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(max : h.b) // expected-error {{expected variable name, array element or array section}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(+ : ba) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(* : ca) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(- : da) // expected-error {{const-qualified list item cannot be reduction}} expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(^ : fl) // expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(&& : S2::S2s) // expected-error {{shared variable cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(&& : S2::S2sc) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(+ : h, k) // expected-error {{threadprivate or thread local variable cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(+ : o) // expected-error 2 {{no viable overloaded '='}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop private(i), reduction(+ : j), reduction(+ : q) // expected-error 4 {{argument of OpenMP clause 'reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp parallel private(k)
+#pragma omp taskloop reduction(+ : p), reduction(+ : p) // expected-error 2 {{argument of OpenMP clause 'reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(+ : p), reduction(+ : p) // expected-error 2 {{variable can appear only once in OpenMP 'reduction' clause}} expected-note 2 {{previously referenced here}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(+ : r) // expected-error 2 {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp parallel shared(i)
+#pragma omp parallel reduction(min : i)
+#pragma omp taskloop reduction(max : j) // expected-error 2 {{argument of OpenMP clause 'reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp parallel private(fl)
+#pragma omp taskloop reduction(+ : fl)
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp parallel reduction(* : fl)
+#pragma omp taskloop reduction(+ : fl)
+  for (int i = 0; i < 10; ++i)
+    foo();
+
+  return T();
+}
+
+namespace A {
+double x;
+#pragma omp threadprivate(x) // expected-note {{defined as threadprivate or thread local}}
+}
+namespace B {
+using A::x;
+}
+
+int main(int argc, char **argv) {
+  const int d = 5;       // expected-note 2 {{'d' defined here}}
+  const int da[5] = {0}; // expected-note {{'da' defined here}}
+  int qa[5] = {0};
+  S4 e(4);
+  S5 g(5);
+  int i;
+  int &j = i;                      // expected-note 2 {{'j' defined here}}
+  S3 &p = k;                       // expected-note 2 {{'p' defined here}}
+  const int &r = da[i];            // expected-note {{'r' defined here}}
+  int &q = qa[i];                  // expected-note {{'q' defined here}}
+  float fl;
+#pragma omp taskloop reduction // expected-error {{expected '(' after 'reduction'}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction + // expected-error {{expected '(' after 'reduction'}} expected-warning {{extra tokens at the end of '#pragma omp taskloop' are ignored}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction( // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(- // expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction() // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(*) // expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected expression}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(\) // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(foo : argc // expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{incorrect reduction identifier, expected one of '+', '-', '*', '&', '|', '^', '&&', '||', 'min' or 'max'}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(| : argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(|| : argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name, array element or array section}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(~ : argc) // expected-error {{expected unqualified-id}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(&& : argc)
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(^ : S1) // expected-error {{'S1' does not refer to a value}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(+ : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 2 {{const-qualified list item cannot be reduction}} expected-error {{'operator+' is a private member of 'S2'}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(min : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 2 {{arguments of OpenMP clause 'reduction' for 'min' or 'max' must be of arithmetic type}} expected-error 2 {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(max : h.b) // expected-error {{expected variable name, array element or array section}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(+ : ba) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(* : ca) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(- : da) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(^ : fl) // expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(&& : S2::S2s) // expected-error {{shared variable cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(&& : S2::S2sc) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(& : e, g) // expected-error {{calling a private constructor of class 'S4'}} expected-error {{invalid operands to binary expression ('S4' and 'S4')}} expected-error {{calling a private constructor of class 'S5'}} expected-error {{invalid operands to binary expression ('S5' and 'S5')}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(+ : h, k, B::x) // expected-error 2 {{threadprivate or thread local variable cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(+ : o) // expected-error {{no viable overloaded '='}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop private(i), reduction(+ : j), reduction(+ : q) // expected-error 2 {{argument of OpenMP clause 'reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp parallel private(k)
+#pragma omp taskloop reduction(+ : p), reduction(+ : p) // expected-error 2 {{argument of OpenMP clause 'reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(+ : p), reduction(+ : p) // expected-error {{variable can appear only once in OpenMP 'reduction' clause}} expected-note {{previously referenced here}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop reduction(+ : r) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp parallel shared(i)
+#pragma omp parallel reduction(min : i)
+#pragma omp taskloop reduction(max : j) // expected-error {{argument of OpenMP clause 'reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp parallel private(fl)
+#pragma omp taskloop reduction(+ : fl)
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp parallel reduction(* : fl)
+#pragma omp taskloop reduction(+ : fl)
+  for (int i = 0; i < 10; ++i)
+    foo();
+  static int m;
+#pragma omp taskloop reduction(+ : m) // OK
+  for (int i = 0; i < 10; ++i)
+    m++;
+#pragma omp taskloop nogroup reduction(+ : m) // expected-error {{'reduction' clause cannot be used with 'nogroup' clause}}
+  for (int i = 0; i < 10; ++i)
+    m++;
+
+  return tmain(argc) + tmain(fl); // expected-note {{in instantiation of function template specialization 'tmain<int>' requested here}} expected-note {{in instantiation of function template specialization 'tmain<float>' requested here}}
+}
diff --git a/test/OpenMP/taskloop_simd_ast_print.cpp b/test/OpenMP/taskloop_simd_ast_print.cpp
index 9bee8f7f93fd..8678b9122e29 100644
--- a/test/OpenMP/taskloop_simd_ast_print.cpp
+++ b/test/OpenMP/taskloop_simd_ast_print.cpp
@@ -21,7 +21,7 @@ T tmain(T argc) {
 // CHECK-NEXT: for (int i = 0; i < 2; ++i)
 // CHECK-NEXT: a = 2;
 #pragma omp parallel
-#pragma omp taskloop simd private(argc, b), firstprivate(c, d), lastprivate(d, f) collapse(N) shared(g) if (c) final(d) mergeable priority(f) simdlen(N) nogroup num_tasks(N) reduction(min:a)
+#pragma omp taskloop simd private(argc, b), firstprivate(c, d), lastprivate(d, f) collapse(N) shared(g) if (c) final(d) mergeable priority(f) simdlen(N) nogroup num_tasks(N)
   for (int i = 0; i < 2; ++i)
     for (int j = 0; j < 2; ++j)
       for (int j = 0; j < 2; ++j)
@@ -34,7 +34,7 @@ T tmain(T argc) {
           for (int j = 0; j < 2; ++j)
             foo();
   // CHECK-NEXT: #pragma omp parallel
-  // CHECK-NEXT: #pragma omp taskloop simd private(argc,b) firstprivate(c,d) lastprivate(d,f) collapse(N) shared(g) if(c) final(d) mergeable priority(f) simdlen(N) nogroup num_tasks(N) reduction(min: a)
+  // CHECK-NEXT: #pragma omp taskloop simd private(argc,b) firstprivate(c,d) lastprivate(d,f) collapse(N) shared(g) if(c) final(d) mergeable priority(f) simdlen(N) nogroup num_tasks(N)
   // CHECK-NEXT: for (int i = 0; i < 2; ++i)
   // CHECK-NEXT: for (int j = 0; j < 2; ++j)
   // CHECK-NEXT: for (int j = 0; j < 2; ++j)
@@ -54,8 +54,8 @@ int main(int argc, char **argv) {
   int b = argc, c, d, e, f, g;
   static int a;
 // CHECK: static int a;
-#pragma omp taskloop simd if(taskloop: a) default(none) shared(a) final(b) priority(5) safelen(8) linear(b), aligned(argv) nogroup num_tasks(argc) reduction(*: g)
-  // CHECK-NEXT: #pragma omp taskloop simd if(taskloop: a) default(none) shared(a) final(b) priority(5) safelen(8) linear(b) aligned(argv) nogroup num_tasks(argc) reduction(*: g)
+#pragma omp taskloop simd if(taskloop: a) default(none) shared(a) final(b) priority(5) safelen(8) linear(b), aligned(argv) num_tasks(argc) reduction(*: g)
+  // CHECK-NEXT: #pragma omp taskloop simd if(taskloop: a) default(none) shared(a) final(b) priority(5) safelen(8) linear(b) aligned(argv) num_tasks(argc) reduction(*: g)
   for (int i = 0; i < 2; ++i)
     a = 2;
 // CHECK-NEXT: for (int i = 0; i < 2; ++i)
diff --git a/test/OpenMP/taskloop_simd_codegen.cpp b/test/OpenMP/taskloop_simd_codegen.cpp
index 60ba5f22129b..48c6f479e9fa 100644
--- a/test/OpenMP/taskloop_simd_codegen.cpp
+++ b/test/OpenMP/taskloop_simd_codegen.cpp
@@ -8,6 +8,7 @@
 // CHECK-LABEL: @main
 int main(int argc, char **argv) {
 // CHECK: [[GTID:%.+]] = call i32 @__kmpc_global_thread_num(%ident_t* [[DEFLOC:@.+]])
+// CHECK: call void @__kmpc_taskgroup(%ident_t* [[DEFLOC]], i32 [[GTID]])
 // CHECK: [[TASKV:%.+]] = call i8* @__kmpc_omp_task_alloc(%ident_t* [[DEFLOC]], i32 [[GTID]], i32 33, i64 72, i64 1, i32 (i32, i8*)* bitcast (i32 (i32, [[TDP_TY:%.+]]*)* [[TASK1:@.+]] to i32 (i32, i8*)*))
 // CHECK: [[TASK:%.+]] = bitcast i8* [[TASKV]] to [[TDP_TY]]*
 // CHECK: [[TASK_DATA:%.+]] = getelementptr inbounds [[TDP_TY]], [[TDP_TY]]* [[TASK]], i32 0, i32 0
@@ -19,6 +20,7 @@ int main(int argc, char **argv) {
 // CHECK: store i64 1, i64* [[ST]],
 // CHECK: [[ST_VAL:%.+]] = load i64, i64* [[ST]],
 // CHECK: call void @__kmpc_taskloop(%ident_t* [[DEFLOC]], i32 [[GTID]], i8* [[TASKV]], i32 1, i64* [[DOWN]], i64* [[UP]], i64 [[ST_VAL]], i32 0, i32 0, i64 0, i8* null)
+// CHECK: call void @__kmpc_end_taskgroup(%ident_t* [[DEFLOC]], i32 [[GTID]])
 #pragma omp taskloop simd priority(argc)
   for (int i = 0; i < 10; ++i)
     ;
@@ -33,10 +35,11 @@ int main(int argc, char **argv) {
 // CHECK: store i64 1, i64* [[ST]],
 // CHECK: [[ST_VAL:%.+]] = load i64, i64* [[ST]],
 // CHECK: [[GRAINSIZE:%.+]] = zext i32 %{{.+}} to i64
-// CHECK: call void @__kmpc_taskloop(%ident_t* [[DEFLOC]], i32 [[GTID]], i8* [[TASKV]], i32 1, i64* [[DOWN]], i64* [[UP]], i64 [[ST_VAL]], i32 1, i32 1, i64 [[GRAINSIZE]], i8* null)
+// CHECK: call void @__kmpc_taskloop(%ident_t* [[DEFLOC]], i32 [[GTID]], i8* [[TASKV]], i32 1, i64* [[DOWN]], i64* [[UP]], i64 [[ST_VAL]], i32 0, i32 1, i64 [[GRAINSIZE]], i8* null)
 #pragma omp taskloop simd nogroup grainsize(argc) simdlen(4)
   for (int i = 0; i < 10; ++i)
     ;
+// CHECK: call void @__kmpc_taskgroup(%ident_t* [[DEFLOC]], i32 [[GTID]])
 // CHECK: [[TASKV:%.+]] = call i8* @__kmpc_omp_task_alloc(%ident_t* [[DEFLOC]], i32 [[GTID]], i32 1, i64 72, i64 24, i32 (i32, i8*)* bitcast (i32 (i32, [[TDP_TY:%.+]]*)* [[TASK3:@.+]] to i32 (i32, i8*)*))
 // CHECK: [[TASK:%.+]] = bitcast i8* [[TASKV]] to [[TDP_TY]]*
 // CHECK: [[TASK_DATA:%.+]] = getelementptr inbounds [[TDP_TY]], [[TDP_TY]]* [[TASK]], i32 0, i32 0
@@ -50,6 +53,7 @@ int main(int argc, char **argv) {
 // CHECK: store i64 1, i64* [[ST]],
 // CHECK: [[ST_VAL:%.+]] = load i64, i64* [[ST]],
 // CHECK: call void @__kmpc_taskloop(%ident_t* [[DEFLOC]], i32 [[GTID]], i8* [[TASKV]], i32 [[IF_INT]], i64* [[DOWN]], i64* [[UP]], i64 [[ST_VAL]], i32 0, i32 2, i64 4, i8* null)
+// CHECK: call void @__kmpc_end_taskgroup(%ident_t* [[DEFLOC]], i32 [[GTID]])
   int i;
 #pragma omp taskloop simd if(argc) shared(argc, argv) collapse(2) num_tasks(4) safelen(32)
   for (i = 0; i < argc; ++i)
diff --git a/test/OpenMP/taskloop_simd_reduction_messages.cpp b/test/OpenMP/taskloop_simd_reduction_messages.cpp
new file mode 100644
index 000000000000..f9739f708a28
--- /dev/null
+++ b/test/OpenMP/taskloop_simd_reduction_messages.cpp
@@ -0,0 +1,331 @@
+// RUN: %clang_cc1 -verify -fopenmp -ferror-limit 150 -o - %s
+// RUN: %clang_cc1 -verify -fopenmp -std=c++98 -ferror-limit 150 -o - %s
+// RUN: %clang_cc1 -verify -fopenmp -std=c++11 -ferror-limit 150 -o - %s
+
+void foo() {
+}
+
+bool foobool(int argc) {
+  return argc;
+}
+
+void foobar(int &ref) {
+#pragma omp taskloop simd reduction(+:ref)
+  for (int i = 0; i < 10; ++i)
+    foo();
+}
+
+struct S1; // expected-note {{declared here}} expected-note 4 {{forward declaration of 'S1'}}
+extern S1 a;
+class S2 {
+  mutable int a;
+  S2 &operator+(const S2 &arg) { return (*this); } // expected-note 3 {{implicitly declared private here}}
+
+public:
+  S2() : a(0) {}
+  S2(S2 &s2) : a(s2.a) {}
+  static float S2s; // expected-note 2 {{static data member is predetermined as shared}}
+  static const float S2sc;
+};
+const float S2::S2sc = 0; // expected-note 2 {{'S2sc' defined here}}
+S2 b;                     // expected-note 3 {{'b' defined here}}
+const S2 ba[5];           // expected-note 2 {{'ba' defined here}}
+class S3 {
+  int a;
+
+public:
+  int b;
+  S3() : a(0) {}
+  S3(const S3 &s3) : a(s3.a) {}
+  S3 operator+(const S3 &arg1) { return arg1; }
+};
+int operator+(const S3 &arg1, const S3 &arg2) { return 5; }
+S3 c;               // expected-note 3 {{'c' defined here}}
+const S3 ca[5];     // expected-note 2 {{'ca' defined here}}
+extern const int f; // expected-note 4 {{'f' declared here}}
+class S4 {
+  int a;
+  S4(); // expected-note {{implicitly declared private here}}
+  S4(const S4 &s4);
+  S4 &operator+(const S4 &arg) { return (*this); }
+
+public:
+  S4(int v) : a(v) {}
+};
+S4 &operator&=(S4 &arg1, S4 &arg2) { return arg1; }
+class S5 {
+  int a;
+  S5() : a(0) {} // expected-note {{implicitly declared private here}}
+  S5(const S5 &s5) : a(s5.a) {}
+  S5 &operator+(const S5 &arg);
+
+public:
+  S5(int v) : a(v) {}
+};
+class S6 { // expected-note 3 {{candidate function (the implicit copy assignment operator) not viable: no known conversion from 'int' to 'const S6' for 1st argument}}
+#if __cplusplus >= 201103L // C++11 or later
+// expected-note@-2 3 {{candidate function (the implicit move assignment operator) not viable}}
+#endif
+  int a;
+
+public:
+  S6() : a(6) {}
+  operator int() { return 6; }
+} o;
+
+struct S7 {
+  int a: 32;
+  S7() {
+#pragma omp taskloop reduction(+:a) // expected-error {{expected addressable reduction item for the task-based directives}}
+    for (int i = 0; i < 10; ++i)
+      ++a;
+  }
+};
+
+S3 h, k;
+#pragma omp threadprivate(h) // expected-note 2 {{defined as threadprivate or thread local}}
+
+template <class T>       // expected-note {{declared here}}
+T tmain(T argc) {
+  const T d = T();       // expected-note 4 {{'d' defined here}}
+  const T da[5] = {T()}; // expected-note 2 {{'da' defined here}}
+  T qa[5] = {T()};
+  T i;
+  T &j = i;                        // expected-note 4 {{'j' defined here}}
+  S3 &p = k;                       // expected-note 2 {{'p' defined here}}
+  const T &r = da[(int)i];         // expected-note 2 {{'r' defined here}}
+  T &q = qa[(int)i];               // expected-note 2 {{'q' defined here}}
+  T fl;
+#pragma omp taskloop simd reduction // expected-error {{expected '(' after 'reduction'}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction + // expected-error {{expected '(' after 'reduction'}} expected-warning {{extra tokens at the end of '#pragma omp taskloop simd' are ignored}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction( // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(- // expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction() // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(*) // expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected expression}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(\) // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(& : argc // expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(| : argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(|| : argc ? i : argc) // expected-error 2 {{expected variable name, array element or array section}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(foo : argc) //expected-error {{incorrect reduction identifier, expected one of '+', '-', '*', '&', '|', '^', '&&', '||', 'min' or 'max' or declare reduction for type 'float'}} expected-error {{incorrect reduction identifier, expected one of '+', '-', '*', '&', '|', '^', '&&', '||', 'min' or 'max' or declare reduction for type 'int'}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(&& : argc)
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(^ : T) // expected-error {{'T' does not refer to a value}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(+ : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 3 {{const-qualified list item cannot be reduction}} expected-error 2 {{'operator+' is a private member of 'S2'}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(min : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 4 {{arguments of OpenMP clause 'reduction' for 'min' or 'max' must be of arithmetic type}} expected-error 3 {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(max : h.b) // expected-error {{expected variable name, array element or array section}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(+ : ba) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(* : ca) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(- : da) // expected-error {{const-qualified list item cannot be reduction}} expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(^ : fl) // expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(&& : S2::S2s) // expected-error {{shared variable cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(&& : S2::S2sc) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(+ : h, k) // expected-error {{threadprivate or thread local variable cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(+ : o) // expected-error 2 {{no viable overloaded '='}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd private(i), reduction(+ : j), reduction(+ : q) // expected-error 4 {{argument of OpenMP clause 'reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp parallel private(k)
+#pragma omp taskloop simd reduction(+ : p), reduction(+ : p) // expected-error 2 {{argument of OpenMP clause 'reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(+ : p), reduction(+ : p) // expected-error 2 {{variable can appear only once in OpenMP 'reduction' clause}} expected-note 2 {{previously referenced here}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(+ : r) // expected-error 2 {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp parallel shared(i)
+#pragma omp parallel reduction(min : i)
+#pragma omp taskloop simd reduction(max : j) // expected-error 2 {{argument of OpenMP clause 'reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp parallel private(fl)
+#pragma omp taskloop simd reduction(+ : fl)
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp parallel reduction(* : fl)
+#pragma omp taskloop simd reduction(+ : fl)
+  for (int i = 0; i < 10; ++i)
+    foo();
+
+  return T();
+}
+
+namespace A {
+double x;
+#pragma omp threadprivate(x) // expected-note {{defined as threadprivate or thread local}}
+}
+namespace B {
+using A::x;
+}
+
+int main(int argc, char **argv) {
+  const int d = 5;       // expected-note 2 {{'d' defined here}}
+  const int da[5] = {0}; // expected-note {{'da' defined here}}
+  int qa[5] = {0};
+  S4 e(4);
+  S5 g(5);
+  int i;
+  int &j = i;                      // expected-note 2 {{'j' defined here}}
+  S3 &p = k;                       // expected-note 2 {{'p' defined here}}
+  const int &r = da[i];            // expected-note {{'r' defined here}}
+  int &q = qa[i];                  // expected-note {{'q' defined here}}
+  float fl;
+#pragma omp taskloop simd reduction // expected-error {{expected '(' after 'reduction'}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction + // expected-error {{expected '(' after 'reduction'}} expected-warning {{extra tokens at the end of '#pragma omp taskloop simd' are ignored}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction( // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(- // expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction() // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(*) // expected-warning {{missing ':' after reduction identifier - ignoring}} expected-error {{expected expression}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(\) // expected-error {{expected unqualified-id}} expected-warning {{missing ':' after reduction identifier - ignoring}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(foo : argc // expected-error {{expected ')'}} expected-note {{to match this '('}} expected-error {{incorrect reduction identifier, expected one of '+', '-', '*', '&', '|', '^', '&&', '||', 'min' or 'max'}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(| : argc, // expected-error {{expected expression}} expected-error {{expected ')'}} expected-note {{to match this '('}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(|| : argc > 0 ? argv[1] : argv[2]) // expected-error {{expected variable name, array element or array section}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(~ : argc) // expected-error {{expected unqualified-id}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(&& : argc)
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(^ : S1) // expected-error {{'S1' does not refer to a value}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(+ : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 2 {{const-qualified list item cannot be reduction}} expected-error {{'operator+' is a private member of 'S2'}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(min : a, b, c, d, f) // expected-error {{a reduction list item with incomplete type 'S1'}} expected-error 2 {{arguments of OpenMP clause 'reduction' for 'min' or 'max' must be of arithmetic type}} expected-error 2 {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(max : h.b) // expected-error {{expected variable name, array element or array section}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(+ : ba) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(* : ca) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(- : da) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(^ : fl) // expected-error {{invalid operands to binary expression ('float' and 'float')}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(&& : S2::S2s) // expected-error {{shared variable cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(&& : S2::S2sc) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(& : e, g) // expected-error {{calling a private constructor of class 'S4'}} expected-error {{invalid operands to binary expression ('S4' and 'S4')}} expected-error {{calling a private constructor of class 'S5'}} expected-error {{invalid operands to binary expression ('S5' and 'S5')}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(+ : h, k, B::x) // expected-error 2 {{threadprivate or thread local variable cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(+ : o) // expected-error {{no viable overloaded '='}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd private(i), reduction(+ : j), reduction(+ : q) // expected-error 2 {{argument of OpenMP clause 'reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp parallel private(k)
+#pragma omp taskloop simd reduction(+ : p), reduction(+ : p) // expected-error 2 {{argument of OpenMP clause 'reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(+ : p), reduction(+ : p) // expected-error {{variable can appear only once in OpenMP 'reduction' clause}} expected-note {{previously referenced here}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp taskloop simd reduction(+ : r) // expected-error {{const-qualified list item cannot be reduction}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp parallel shared(i)
+#pragma omp parallel reduction(min : i)
+#pragma omp taskloop simd reduction(max : j) // expected-error {{argument of OpenMP clause 'reduction' must reference the same object in all threads}}
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp parallel private(fl)
+#pragma omp taskloop simd reduction(+ : fl)
+  for (int i = 0; i < 10; ++i)
+    foo();
+#pragma omp parallel reduction(* : fl)
+#pragma omp taskloop simd reduction(+ : fl)
+  for (int i = 0; i < 10; ++i)
+    foo();
+  static int m;
+#pragma omp taskloop simd reduction(+ : m) // OK
+  for (int i = 0; i < 10; ++i)
+    m++;
+#pragma omp taskloop simd reduction(+ : m) nogroup // expected-error {{'reduction' clause cannot be used with 'nogroup' clause}}
+  for (int i = 0; i < 10; ++i)
+    m++;
+
+  return tmain(argc) + tmain(fl); // expected-note {{in instantiation of function template specialization 'tmain<int>' requested here}} expected-note {{in instantiation of function template specialization 'tmain<float>' requested here}}
+}
diff --git a/test/Preprocessor/init.c b/test/Preprocessor/init.c
index 0bd9e9784599..0ee7edecb3c4 100644
--- a/test/Preprocessor/init.c
+++ b/test/Preprocessor/init.c
@@ -2185,13 +2185,13 @@
 // ARMV6-CLOUDABI:#define __CloudABI__ 1
 // ARMV6-CLOUDABI:#define __arm__ 1
 
-// RUN: %clang_cc1 -E -dM -ffreestanding -triple=arm-netbsd-eabi < /dev/null | FileCheck -match-full-lines -check-prefix ARM-NETBSD %s
-//
+// RUN: %clang -E -dM -ffreestanding -target arm-netbsd-eabi %s -o - | FileCheck -match-full-lines -check-prefix ARM-NETBSD %s
+
 // ARM-NETBSD-NOT:#define _LP64
 // ARM-NETBSD:#define __APCS_32__ 1
 // ARM-NETBSD-NOT:#define __ARMEB__ 1
 // ARM-NETBSD:#define __ARMEL__ 1
-// ARM-NETBSD:#define __ARM_ARCH_4T__ 1
+// ARM-NETBSD:#define __ARM_ARCH_5TE__ 1
 // ARM-NETBSD:#define __ARM_DWARF_EH__ 1
 // ARM-NETBSD:#define __ARM_EABI__ 1
 // ARM-NETBSD-NOT:#define __ARM_BIG_ENDIAN 1
@@ -2333,6 +2333,7 @@
 // ARM-NETBSD:#define __SIZE_MAX__ 4294967295UL
 // ARM-NETBSD:#define __SIZE_TYPE__ long unsigned int
 // ARM-NETBSD:#define __SIZE_WIDTH__ 32
+// ARM-NETBSD:#define __SOFTFP__ 1
 // ARM-NETBSD:#define __UINT16_C_SUFFIX__
 // ARM-NETBSD:#define __UINT16_MAX__ 65535
 // ARM-NETBSD:#define __UINT16_TYPE__ unsigned short
@@ -2377,6 +2378,11 @@
 // ARM-NETBSD:#define __arm 1
 // ARM-NETBSD:#define __arm__ 1
 
+// RUN: %clang -E -dM -ffreestanding -target arm-netbsd-eabihf %s -o - | FileCheck -match-full-lines -check-prefix ARMHF-NETBSD %s
+// ARMHF-NETBSD:#define __SIZE_WIDTH__ 32
+// ARMHF-NETBSD-NOT:#define __SOFTFP__ 1
+// ARMHF-NETBSD:#define __UINT16_C_SUFFIX__
+
 // RUN: %clang_cc1 -E -dM -ffreestanding -triple=arm-none-eabi < /dev/null | FileCheck -match-full-lines -check-prefix ARM-NONE-EABI %s
 // RUN: %clang_cc1 -E -dM -ffreestanding -triple=arm-none-eabihf < /dev/null | FileCheck -match-full-lines -check-prefix ARM-NONE-EABI %s
 // RUN: %clang_cc1 -E -dM -ffreestanding -triple=aarch64-none-eabi < /dev/null | FileCheck -match-full-lines -check-prefix ARM-NONE-EABI %s
diff --git a/test/Rewriter/objc-modern-metadata-visibility2.mm b/test/Rewriter/objc-modern-metadata-visibility2.mm
new file mode 100644
index 000000000000..4e64ac4beb46
--- /dev/null
+++ b/test/Rewriter/objc-modern-metadata-visibility2.mm
@@ -0,0 +1,45 @@
+// REQUIRES: abi-breaking-checks
+// NOTE: This test has been split from objc-modern-metadata-visibility.mm in
+// order to test with -reverse-iterate as this flag is only present with
+// ABI_BREAKING_CHECKS.
+
+// RUN: %clang_cc1 -E %s -o %t.mm -mllvm -reverse-iterate
+// RUN: %clang_cc1 -x objective-c++ -fblocks -fms-extensions -rewrite-objc %t.mm -mllvm -reverse-iterate -o - | FileCheck %s
+// rdar://11144048
+
+@class NSString;
+
+@interface NSObject {
+    Class isa;
+}
+@end
+
+@interface Sub : NSObject {
+    int subIvar;
+    NSString *nsstring;
+@private
+    id PrivateIvar;
+}
+@end
+
+@implementation Sub
+- (id) MyNSString { return subIvar ? PrivateIvar : nsstring; }
+@end
+
+@interface NSString @end
+@implementation NSString @end
+
+// CHECK: __declspec(allocate(".objc_ivar$B")) extern "C" __declspec(dllimport) unsigned long OBJC_IVAR_$_Sub$subIvar;
+// CHECK: __declspec(allocate(".objc_ivar$B")) extern "C" unsigned long OBJC_IVAR_$_Sub$PrivateIvar;
+// CHECK: __declspec(allocate(".objc_ivar$B")) extern "C" __declspec(dllimport) unsigned long OBJC_IVAR_$_Sub$nsstring;
+// CHECK: #pragma warning(disable:4273)
+// CHECK: __declspec(allocate(".objc_ivar$B")) extern "C" __declspec(dllexport) unsigned long int OBJC_IVAR_$_Sub$subIvar
+// CHECK: __declspec(allocate(".objc_ivar$B")) extern "C" __declspec(dllexport) unsigned long int OBJC_IVAR_$_Sub$nsstring
+// CHECK: __declspec(allocate(".objc_ivar$B")) extern "C" unsigned long int OBJC_IVAR_$_Sub$PrivateIvar
+// CHECK: extern "C" __declspec(dllimport) struct _class_t OBJC_METACLASS_$_NSObject;
+// CHECK: extern "C" __declspec(dllexport) struct _class_t OBJC_METACLASS_$_Sub
+// CHECK: extern "C" __declspec(dllimport) struct _class_t OBJC_CLASS_$_NSObject;
+// CHECK: extern "C" __declspec(dllexport) struct _class_t OBJC_CLASS_$_Sub
+// CHECK: extern "C" __declspec(dllexport) struct _class_t OBJC_CLASS_$_NSString;
+// CHECK: extern "C" __declspec(dllexport) struct _class_t OBJC_METACLASS_$_NSString
+// CHECK: extern "C" __declspec(dllexport) struct _class_t OBJC_CLASS_$_NSString
diff --git a/test/Sema/address-packed.c b/test/Sema/address-packed.c
index 2799e19c48f1..b0519bacd758 100644
--- a/test/Sema/address-packed.c
+++ b/test/Sema/address-packed.c
@@ -329,3 +329,12 @@ void g13(void) {
   uint32_t *p32;
   p32 = &a[0].x; // no-warning
 }
+
+struct Invalid0 {
+  void *x;
+  struct fwd f; // expected-error {{incomplete type}} expected-note {{forward declaration}}
+} __attribute__((packed));
+
+void *g14(struct Invalid0 *ivl) {
+  return &(ivl->x);
+}
diff --git a/test/Sema/attr-availability.c b/test/Sema/attr-availability.c
index e105037c8ed0..1b8cbd256ce9 100644
--- a/test/Sema/attr-availability.c
+++ b/test/Sema/attr-availability.c
@@ -21,6 +21,9 @@ ATSFontGetPostScriptName(int flags) __attribute__((availability(macosx,introduce
 extern void
 PartiallyAvailable() __attribute__((availability(macosx,introduced=10.8)));
 
+#ifdef WARN_PARTIAL
+// expected-note@+2 2 {{marked partial here}}
+#endif
 enum __attribute__((availability(macosx,introduced=10.8))) PartialEnum {
   kPartialEnumConstant,
 };
@@ -35,11 +38,19 @@ void test_10095131() {
   PartiallyAvailable();
 }
 
+#ifdef WARN_PARTIAL
+// FIXME: This note should point to the declaration with the availability
+// attribute.
+// expected-note@+2 {{marked partial here}}
+#endif
 extern void PartiallyAvailable() ;
 void with_redeclaration() {
-  PartiallyAvailable();  // Don't warn.
-
-  // enums should never warn.
+#ifdef WARN_PARTIAL
+  // expected-warning@+4 {{'PartiallyAvailable' is only available on macOS 10.8 or newer}} expected-note@+4 {{__builtin_available}}
+  // expected-warning@+4 {{'PartialEnum' is only available on macOS 10.8 or newer}} expected-note@+4 {{__builtin_available}}
+  // expected-warning@+3 {{'kPartialEnumConstant' is only available on macOS 10.8 or newer}} expected-note@+3 {{__builtin_available}}
+#endif
+  PartiallyAvailable();
   enum PartialEnum p = kPartialEnumConstant;
 }
 
@@ -86,13 +97,13 @@ enum Original {
   OriginalUnavailable __attribute__((availability(macosx, unavailable))) // expected-note + {{'OriginalUnavailable' has been explicitly marked unavailable here}}
 };
 
-enum AllDeprecated {
-  AllDeprecatedCase, // expected-note + {{'AllDeprecatedCase' has been explicitly marked deprecated here}}
+enum AllDeprecated { // expected-note + {{'AllDeprecated' has been explicitly marked deprecated here}}
+  AllDeprecatedCase,
   AllDeprecatedUnavailable __attribute__((availability(macosx, unavailable))) // expected-note + {{'AllDeprecatedUnavailable' has been explicitly marked unavailable here}}
 } __attribute__((availability(macosx, deprecated=10.2)));
 
-enum AllUnavailable {
-  AllUnavailableCase, // expected-note + {{'AllUnavailableCase' has been explicitly marked unavailable here}}
+enum AllUnavailable { // expected-note + {{'AllUnavailable' has been explicitly marked unavailable here}}
+  AllUnavailableCase,
 } __attribute__((availability(macosx, unavailable)));
 
 enum User {
diff --git a/test/Sema/attr-deprecated.c b/test/Sema/attr-deprecated.c
index 89c06860fb4e..9eeef5aed57d 100644
--- a/test/Sema/attr-deprecated.c
+++ b/test/Sema/attr-deprecated.c
@@ -104,9 +104,9 @@ foo_dep test17, // expected-warning {{'foo_dep' is deprecated}}
         test19;
 
 // rdar://problem/8518751
-enum __attribute__((deprecated)) Test20 { // expected-note {{'Test20' has been explicitly marked deprecated here}}
+enum __attribute__((deprecated)) Test20 { // expected-note 2 {{'Test20' has been explicitly marked deprecated here}}
   test20_a __attribute__((deprecated)), // expected-note {{'test20_a' has been explicitly marked deprecated here}}
-  test20_b // expected-note {{'test20_b' has been explicitly marked deprecated here}}
+  test20_b
 };
 void test20() {
   enum Test20 f; // expected-warning {{'Test20' is deprecated}}
diff --git a/test/Sema/attr-unavailable-message.c b/test/Sema/attr-unavailable-message.c
index 415cb2f079a2..70d5947086f9 100644
--- a/test/Sema/attr-unavailable-message.c
+++ b/test/Sema/attr-unavailable-message.c
@@ -36,13 +36,13 @@ void unavail(void) {
 
 // rdar://10201690
 enum foo {
-    a = 1, // expected-note {{'a' has been explicitly marked deprecated here}}
+    a = 1,
     b __attribute__((deprecated())) = 2, // expected-note {{'b' has been explicitly marked deprecated here}}
     c = 3
-}__attribute__((deprecated()));
+}__attribute__((deprecated())); // expected-note {{'foo' has been explicitly marked deprecated here}}
 
-enum fee { // expected-note {{'fee' has been explicitly marked unavailable here}}
-    r = 1, // expected-note {{'r' has been explicitly marked unavailable here}}
+enum fee { // expected-note 2 {{'fee' has been explicitly marked unavailable here}}
+    r = 1,
     s = 2,
     t = 3
 }__attribute__((unavailable()));
diff --git a/test/Sema/loop-control.c b/test/Sema/loop-control.c
index 6c33e8437bd0..1fc35d10218e 100644
--- a/test/Sema/loop-control.c
+++ b/test/Sema/loop-control.c
@@ -119,3 +119,51 @@ void pr8880_23(int x, int y) {
     for ( ; ({ ++y; break; y;}); ++y) {} // expected-warning{{'break' is bound to loop, GCC binds it to switch}}
   }
 }
+
+void pr32648_1(int x, int y) {
+  switch(x) {
+  case 1:
+    for ( ; ({ ++y; switch (y) { case 0: break; } y;}); ++y) {} // no warning
+  }
+}
+
+void pr32648_2(int x, int y) {
+  while(x) {
+    for ( ; ({ ++y; switch (y) { case 0: continue; } y;}); ++y) {}  // expected-warning {{'continue' is bound to current loop, GCC binds it to the enclosing loop}}
+  }
+}
+
+void pr32648_3(int x, int y) {
+  switch(x) {
+  case 1:
+    for ( ; ({ ++y; for (; y; y++) { break; } y;}); ++y) {} // no warning
+  }
+}
+
+void pr32648_4(int x, int y) {
+  switch(x) {
+  case 1:
+    for ( ; ({ ++y; for (({ break; }); y; y++) { } y;}); ++y) {} // expected-warning{{'break' is bound to loop, GCC binds it to switch}}
+  }
+}
+
+void pr32648_5(int x, int y) {
+  switch(x) {
+  case 1:
+    for ( ; ({ ++y; while (({ break; y; })) {} y;}); ++y) {} // expected-warning{{'break' is bound to current loop, GCC binds it to the enclosing loop}}
+  }
+}
+
+void pr32648_6(int x, int y) {
+  switch(x) {
+  case 1:
+    for ( ; ({ ++y; do {} while (({ break; y; })); y;}); ++y) {} // expected-warning{{'break' is bound to current loop, GCC binds it to the enclosing loop}}
+  }
+}
+
+void pr32648_7(int x, int y) {
+  switch(x) {
+  case 1:
+    for ( ; ({ ++y; do { break; } while (y); y;}); ++y) {} // no warning
+  }
+}
diff --git a/test/Sema/warn-cast-qual.c b/test/Sema/warn-cast-qual.c
index dc11f5717ebe..a682cad75ed6 100644
--- a/test/Sema/warn-cast-qual.c
+++ b/test/Sema/warn-cast-qual.c
@@ -1,4 +1,5 @@
 // RUN: %clang_cc1 -triple x86_64-unknown-unknown -fsyntax-only -Wcast-qual -verify %s
+// RUN: %clang_cc1 -triple x86_64-unknown-unknown -x c++ -fsyntax-only -Wcast-qual -verify %s
 
 #include <stdint.h>
 
@@ -26,4 +27,34 @@ void foo() {
 
   const char **charptrptrc;
   char **charptrptr = (char **)charptrptrc; // expected-warning {{cast from 'const char *' to 'char *' drops const qualifier}}
+
+  const char *constcharptr;
+  char *charptr = (char *)constcharptr; // expected-warning {{cast from 'const char *' to 'char *' drops const qualifier}}
+  const char *constcharptr2 = (char *)constcharptr; // expected-warning {{cast from 'const char *' to 'char *' drops const qualifier}}
+  const char *charptr2 = (char *)charptr; // no warning
+}
+
+void bar_0() {
+  struct C {
+    const int a;
+    int b;
+  };
+
+  const struct C S = {0, 0};
+
+  *(int *)(&S.a) = 0; // expected-warning {{cast from 'const int *' to 'int *' drops const qualifier}}
+  *(int *)(&S.b) = 0; // expected-warning {{cast from 'const int *' to 'int *' drops const qualifier}}
+}
+
+void bar_1() {
+  struct C {
+    const int a;
+    int b;
+  };
+
+  struct C S = {0, 0};
+  S.b = 0; // no warning
+
+  *(int *)(&S.a) = 0; // expected-warning {{cast from 'const int *' to 'int *' drops const qualifier}}
+  *(int *)(&S.b) = 0; // no warning
 }
diff --git a/test/Sema/warn-documentation.cpp b/test/Sema/warn-documentation.cpp
index 0c92b2aa029f..ccf374ccd060 100644
--- a/test/Sema/warn-documentation.cpp
+++ b/test/Sema/warn-documentation.cpp
@@ -1186,7 +1186,7 @@ class Predicate
 /// @brief A C++ wrapper class for providing threaded access to a value
 /// of type T.
 ///
-/// A template specilization class.
+/// A template specialization class.
 //----------------------------------------------------------------------
 template<> class Predicate<int, char>
 {
diff --git a/test/SemaCXX/amdgpu-sizeof-alignof.cpp b/test/SemaCXX/amdgpu-sizeof-alignof.cpp
new file mode 100644
index 000000000000..97da71a1a3e2
--- /dev/null
+++ b/test/SemaCXX/amdgpu-sizeof-alignof.cpp
@@ -0,0 +1,47 @@
+// RUN: %clang_cc1 -triple amdgcn---amdgiz -std=c++11 -fsyntax-only -verify %s
+// expected-no-diagnostics
+typedef __SIZE_TYPE__ size_t;
+typedef __PTRDIFF_TYPE__ ptrdiff_t;
+typedef __INTPTR_TYPE__ intptr_t;
+typedef __UINTPTR_TYPE__ uintptr_t;
+typedef __attribute__((address_space(1))) void *global_ptr_t;
+typedef __attribute__((address_space(2))) void *constant_ptr_t;
+typedef __attribute__((address_space(3))) void *local_ptr_t;
+typedef __attribute__((address_space(5))) void *private_ptr_t;
+
+void test() {
+  static_assert(sizeof(size_t) == 8, "bad size");
+  static_assert(alignof(size_t) == 8, "bad alignment");
+  static_assert(sizeof(intptr_t) == 8, "bad size");
+  static_assert(alignof(intptr_t) == 8, "bad alignment");
+  static_assert(sizeof(uintptr_t) == 8, "bad size");
+  static_assert(alignof(uintptr_t) == 8, "bad alignment");
+  static_assert(sizeof(ptrdiff_t) == 8, "bad size");
+  static_assert(alignof(ptrdiff_t) == 8, "bad alignment");
+
+  static_assert(sizeof(char) == 1, "bad size");
+  static_assert(alignof(char) == 1, "bad alignment");
+  static_assert(sizeof(short) == 2, "bad size");
+  static_assert(alignof(short) == 2, "bad alignment");
+  static_assert(sizeof(int) == 4, "bad size");
+  static_assert(alignof(int) == 4, "bad alignment");
+  static_assert(sizeof(long) == 8, "bad size");
+  static_assert(alignof(long) == 8, "bad alignment");
+  static_assert(sizeof(long long) == 8, "bad size");
+  static_assert(alignof(long long) == 8, "bad alignment");
+  static_assert(sizeof(float) == 4, "bad size");
+  static_assert(alignof(float) == 4, "bad alignment");
+  static_assert(sizeof(double) == 8, "bad size");
+  static_assert(alignof(double) == 8, "bad alignment");
+
+  static_assert(sizeof(void*) == 8, "bad size");
+  static_assert(alignof(void*) == 8, "bad alignment");
+  static_assert(sizeof(global_ptr_t) == 8, "bad size");
+  static_assert(alignof(global_ptr_t) == 8, "bad alignment");
+  static_assert(sizeof(constant_ptr_t) == 8, "bad size");
+  static_assert(alignof(constant_ptr_t) == 8, "bad alignment");
+  static_assert(sizeof(local_ptr_t) == 4, "bad size");
+  static_assert(alignof(local_ptr_t) == 4, "bad alignment");
+  static_assert(sizeof(private_ptr_t) == 4, "bad size");
+  static_assert(alignof(private_ptr_t) == 4, "bad alignment");
+}
diff --git a/test/SemaCXX/attr-deprecated.cpp b/test/SemaCXX/attr-deprecated.cpp
index 1680c5c6760d..5ba55f0c23b5 100644
--- a/test/SemaCXX/attr-deprecated.cpp
+++ b/test/SemaCXX/attr-deprecated.cpp
@@ -199,8 +199,8 @@ namespace test5 {
 
 // rdar://problem/8518751
 namespace test6 {
-  enum __attribute__((deprecated)) A { // expected-note {{'A' has been explicitly marked deprecated here}}
-    a0 // expected-note {{'a0' has been explicitly marked deprecated here}}
+  enum __attribute__((deprecated)) A { // expected-note 2 {{'A' has been explicitly marked deprecated here}}
+    a0
   };
   void testA() {
     A x; // expected-warning {{'A' is deprecated}}
@@ -218,8 +218,8 @@ namespace test6 {
   }
 
   template <class T> struct C {
-    enum __attribute__((deprecated)) Enum { // expected-note {{'Enum' has been explicitly marked deprecated here}}
-      c0 // expected-note {{'c0' has been explicitly marked deprecated here}}
+    enum __attribute__((deprecated)) Enum { // expected-note 2 {{'Enum' has been explicitly marked deprecated here}}
+      c0
     };
   };
   void testC() {
diff --git a/test/SemaCXX/coroutines.cpp b/test/SemaCXX/coroutines.cpp
index 4eedc2162f07..6394829f356a 100644
--- a/test/SemaCXX/coroutines.cpp
+++ b/test/SemaCXX/coroutines.cpp
@@ -22,8 +22,24 @@ void no_coroutine_traits() {
 
 namespace std {
 namespace experimental {
-template <typename... T>
-struct coroutine_traits; // expected-note {{declared here}}
+
+template <class... Args>
+struct void_t_imp {
+  using type = void;
+};
+template <class... Args>
+using void_t = typename void_t_imp<Args...>::type;
+
+template <class T, class = void>
+struct traits_sfinae_base {};
+
+template <class T>
+struct traits_sfinae_base<T, void_t<typename T::promise_type>> {
+  using promise_type = typename T::promise_type;
+};
+
+template <class Ret, class... Args>
+struct coroutine_traits : public traits_sfinae_base<Ret> {};
 }}  // namespace std::experimental
 
 template<typename Promise> struct coro {};
@@ -50,8 +66,9 @@ struct suspend_never {
   void await_resume() {}
 };
 
-void no_specialization() {
-  co_await a; // expected-error {{implicit instantiation of undefined template 'std::experimental::coroutine_traits<void>'}}
+struct DummyVoidTag {};
+DummyVoidTag no_specialization() { // expected-error {{this function cannot be a coroutine: 'std::experimental::coroutine_traits<DummyVoidTag>' has no member named 'promise_type'}}
+  co_await a;
 }
 
 template <typename... T>
@@ -905,3 +922,243 @@ void test_dependent_param(T t, U) {
   co_return 42;
 }
 template void test_dependent_param(NoCopy<0>, NoCopy<1>); // expected-note {{requested here}}
+
+namespace CoroHandleMemberFunctionTest {
+struct CoroMemberTag {};
+struct BadCoroMemberTag {};
+
+template <class T, class U>
+constexpr bool IsSameV = false;
+template <class T>
+constexpr bool IsSameV<T, T> = true;
+
+template <class T>
+struct TypeTest {
+  template <class U>
+  static constexpr bool IsSame = IsSameV<T, U>;
+
+  template <class... Args>
+  static constexpr bool MatchesArgs = IsSameV<T,
+                                              std::experimental::coroutine_traits<CoroMemberTag, Args...>>;
+};
+
+template <class T>
+struct AwaitReturnsType {
+  bool await_ready() const;
+  void await_suspend(...) const;
+  T await_resume() const;
+};
+
+template <class... CoroTraitsArgs>
+struct CoroMemberPromise {
+  using TraitsT = std::experimental::coroutine_traits<CoroTraitsArgs...>;
+  using TypeTestT = TypeTest<TraitsT>;
+  using AwaitTestT = AwaitReturnsType<TypeTestT>;
+
+  CoroMemberTag get_return_object();
+  suspend_always initial_suspend();
+  suspend_always final_suspend();
+
+  AwaitTestT yield_value(int);
+
+  void return_void();
+  void unhandled_exception();
+};
+
+} // namespace CoroHandleMemberFunctionTest
+
+template <class... Args>
+struct ::std::experimental::coroutine_traits<CoroHandleMemberFunctionTest::CoroMemberTag, Args...> {
+  using promise_type = CoroHandleMemberFunctionTest::CoroMemberPromise<CoroHandleMemberFunctionTest::CoroMemberTag, Args...>;
+};
+
+namespace CoroHandleMemberFunctionTest {
+struct TestType {
+
+  CoroMemberTag test_qual() {
+    auto TC = co_yield 0;
+    static_assert(TC.MatchesArgs<TestType &>, "");
+    static_assert(!TC.MatchesArgs<TestType>, "");
+    static_assert(!TC.MatchesArgs<TestType *>, "");
+  }
+
+  CoroMemberTag test_sanity(int *) const {
+    auto TC = co_yield 0;
+    static_assert(TC.MatchesArgs<const TestType &>, ""); // expected-error {{static_assert failed}}
+    static_assert(TC.MatchesArgs<const TestType &>, ""); // expected-error {{static_assert failed}}
+    static_assert(TC.MatchesArgs<const TestType &, int *>, "");
+  }
+
+  CoroMemberTag test_qual(int *, const float &&, volatile void *volatile) const {
+    auto TC = co_yield 0;
+    static_assert(TC.MatchesArgs<const TestType &, int *, const float &&, volatile void *volatile>, "");
+  }
+
+  CoroMemberTag test_qual() const volatile {
+    auto TC = co_yield 0;
+    static_assert(TC.MatchesArgs<const volatile TestType &>, "");
+  }
+
+  CoroMemberTag test_ref_qual() & {
+    auto TC = co_yield 0;
+    static_assert(TC.MatchesArgs<TestType &>, "");
+  }
+  CoroMemberTag test_ref_qual() const & {
+    auto TC = co_yield 0;
+    static_assert(TC.MatchesArgs<TestType const &>, "");
+  }
+  CoroMemberTag test_ref_qual() && {
+    auto TC = co_yield 0;
+    static_assert(TC.MatchesArgs<TestType &&>, "");
+  }
+  CoroMemberTag test_ref_qual(const char *&) const volatile && {
+    auto TC = co_yield 0;
+    static_assert(TC.MatchesArgs<TestType const volatile &&, const char *&>, "");
+  }
+
+  CoroMemberTag test_args(int) {
+    auto TC = co_yield 0;
+    static_assert(TC.MatchesArgs<TestType &, int>, "");
+  }
+  CoroMemberTag test_args(int, long &, void *) const {
+    auto TC = co_yield 0;
+    static_assert(TC.MatchesArgs<TestType const &, int, long &, void *>, "");
+  }
+
+  template <class... Args>
+  CoroMemberTag test_member_template(Args...) const && {
+    auto TC = co_yield 0;
+    static_assert(TC.template MatchesArgs<TestType const &&, Args...>, "");
+  }
+
+  static CoroMemberTag test_static() {
+    auto TC = co_yield 0;
+    static_assert(TC.MatchesArgs<>, "");
+    static_assert(!TC.MatchesArgs<TestType>, "");
+    static_assert(!TC.MatchesArgs<TestType &>, "");
+    static_assert(!TC.MatchesArgs<TestType *>, "");
+  }
+
+  static CoroMemberTag test_static(volatile void *const, char &&) {
+    auto TC = co_yield 0;
+    static_assert(TC.MatchesArgs<volatile void *const, char &&>, "");
+  }
+
+  template <class Dummy>
+  static CoroMemberTag test_static_template(const char *volatile &, unsigned) {
+    auto TC = co_yield 0;
+    using TCT = decltype(TC);
+    static_assert(TCT::MatchesArgs<const char *volatile &, unsigned>, "");
+    static_assert(!TCT::MatchesArgs<TestType &, const char *volatile &, unsigned>, "");
+  }
+
+  BadCoroMemberTag test_diagnostics() {
+    // expected-error@-1 {{this function cannot be a coroutine: 'std::experimental::coroutine_traits<CoroHandleMemberFunctionTest::BadCoroMemberTag, CoroHandleMemberFunctionTest::TestType &>' has no member named 'promise_type'}}
+    co_return;
+  }
+  BadCoroMemberTag test_diagnostics(int) const && {
+    // expected-error@-1 {{this function cannot be a coroutine: 'std::experimental::coroutine_traits<CoroHandleMemberFunctionTest::BadCoroMemberTag, const CoroHandleMemberFunctionTest::TestType &&, int>' has no member named 'promise_type'}}
+    co_return;
+  }
+
+  static BadCoroMemberTag test_static_diagnostics(long *) {
+    // expected-error@-1 {{this function cannot be a coroutine: 'std::experimental::coroutine_traits<CoroHandleMemberFunctionTest::BadCoroMemberTag, long *>' has no member named 'promise_type'}}
+    co_return;
+  }
+};
+
+template CoroMemberTag TestType::test_member_template(long, const char *) const &&;
+template CoroMemberTag TestType::test_static_template<void>(const char *volatile &, unsigned);
+
+template <class... Args>
+struct DepTestType {
+
+  CoroMemberTag test_sanity(int *) const {
+    auto TC = co_yield 0;
+    static_assert(TC.template MatchesArgs<const DepTestType &>, ""); // expected-error {{static_assert failed}}
+    static_assert(TC.template MatchesArgs<>, ""); // expected-error {{static_assert failed}}
+    static_assert(TC.template MatchesArgs<const DepTestType &, int *>, "");
+  }
+
+  CoroMemberTag test_qual() {
+    auto TC = co_yield 0;
+    static_assert(TC.template MatchesArgs<DepTestType &>, "");
+    static_assert(!TC.template MatchesArgs<DepTestType>, "");
+    static_assert(!TC.template MatchesArgs<DepTestType *>, "");
+  }
+
+  CoroMemberTag test_qual(int *, const float &&, volatile void *volatile) const {
+    auto TC = co_yield 0;
+    static_assert(TC.template MatchesArgs<const DepTestType &, int *, const float &&, volatile void *volatile>, "");
+  }
+
+  CoroMemberTag test_qual() const volatile {
+    auto TC = co_yield 0;
+    static_assert(TC.template MatchesArgs<const volatile DepTestType &>, "");
+  }
+
+  CoroMemberTag test_ref_qual() & {
+    auto TC = co_yield 0;
+    static_assert(TC.template MatchesArgs<DepTestType &>, "");
+  }
+  CoroMemberTag test_ref_qual() const & {
+    auto TC = co_yield 0;
+    static_assert(TC.template MatchesArgs<DepTestType const &>, "");
+  }
+  CoroMemberTag test_ref_qual() && {
+    auto TC = co_yield 0;
+    static_assert(TC.template MatchesArgs<DepTestType &&>, "");
+  }
+  CoroMemberTag test_ref_qual(const char *&) const volatile && {
+    auto TC = co_yield 0;
+    static_assert(TC.template MatchesArgs<DepTestType const volatile &&, const char *&>, "");
+  }
+
+  CoroMemberTag test_args(int) {
+    auto TC = co_yield 0;
+    static_assert(TC.template MatchesArgs<DepTestType &, int>, "");
+  }
+  CoroMemberTag test_args(int, long &, void *) const {
+    auto TC = co_yield 0;
+    static_assert(TC.template MatchesArgs<DepTestType const &, int, long &, void *>, "");
+  }
+
+  template <class... UArgs>
+  CoroMemberTag test_member_template(UArgs...) const && {
+    auto TC = co_yield 0;
+    static_assert(TC.template MatchesArgs<DepTestType const &&, UArgs...>, "");
+  }
+
+  static CoroMemberTag test_static() {
+    auto TC = co_yield 0;
+    using TCT = decltype(TC);
+    static_assert(TCT::MatchesArgs<>, "");
+    static_assert(!TCT::MatchesArgs<DepTestType>, "");
+    static_assert(!TCT::MatchesArgs<DepTestType &>, "");
+    static_assert(!TCT::MatchesArgs<DepTestType *>, "");
+
+    // Ensure diagnostics are actually being generated here
+    static_assert(TCT::MatchesArgs<int>, ""); // expected-error {{static_assert failed}}
+  }
+
+  static CoroMemberTag test_static(volatile void *const, char &&) {
+    auto TC = co_yield 0;
+    using TCT = decltype(TC);
+    static_assert(TCT::MatchesArgs<volatile void *const, char &&>, "");
+  }
+
+  template <class Dummy>
+  static CoroMemberTag test_static_template(const char *volatile &, unsigned) {
+    auto TC = co_yield 0;
+    using TCT = decltype(TC);
+    static_assert(TCT::MatchesArgs<const char *volatile &, unsigned>, "");
+    static_assert(!TCT::MatchesArgs<DepTestType &, const char *volatile &, unsigned>, "");
+  }
+};
+
+template struct DepTestType<int>; // expected-note {{requested here}}
+template CoroMemberTag DepTestType<int>::test_member_template(long, const char *) const &&;
+
+template CoroMemberTag DepTestType<int>::test_static_template<void>(const char *volatile &, unsigned);
+
+} // namespace CoroHandleMemberFunctionTest
diff --git a/test/SemaCXX/dllimport-memptr.cpp b/test/SemaCXX/dllimport-memptr.cpp
new file mode 100644
index 000000000000..35bf5b8610b5
--- /dev/null
+++ b/test/SemaCXX/dllimport-memptr.cpp
@@ -0,0 +1,7 @@
+// RUN: %clang_cc1 -triple x86_64-windows-msvc -fms-extensions -verify -std=c++11 %s
+
+// expected-no-diagnostics
+
+struct __declspec(dllimport) Foo { int get_a(); };
+template <int (Foo::*Getter)()> struct HasValue { };
+HasValue<&Foo::get_a> hv;
diff --git a/test/SemaCXX/modules-ts.cppm b/test/SemaCXX/modules-ts.cppm
index 29122ec7dab5..2ea4958bffda 100644
--- a/test/SemaCXX/modules-ts.cppm
+++ b/test/SemaCXX/modules-ts.cppm
@@ -13,7 +13,13 @@ export module foo;
 // expected-note@modules-ts.cppm:* {{loaded from}}
 #endif
 
-static int m; // ok, internal linkage, so no redefinition error
+static int m;
+#if TEST == 2 // FIXME: 'm' has internal linkage, so there should be no error here
+// expected-error@-2 {{redefinition of '}}
+// expected-note@-3 {{unguarded header; consider using #ifdef guards or #pragma once}}
+// FIXME: We should drop the "header from" in this diagnostic.
+// expected-note-re@modules-ts.cppm:1 {{'{{.*}}modules-ts.cppm' included multiple times, additional include site in header from module 'foo'}}
+#endif
 int n;
 #if TEST >= 2
 // expected-error@-2 {{redefinition of '}}
diff --git a/test/SemaCXX/warn-cast-qual.cpp b/test/SemaCXX/warn-cast-qual.cpp
new file mode 100644
index 000000000000..d1f0cc73a63d
--- /dev/null
+++ b/test/SemaCXX/warn-cast-qual.cpp
@@ -0,0 +1,140 @@
+// RUN: %clang_cc1 -triple x86_64-unknown-unknown -fsyntax-only -Wcast-qual -verify %s
+
+#include <stdint.h>
+
+// do *NOT* warn on const_cast<>()
+// use clang-tidy's cppcoreguidelines-pro-type-const-cast for that.
+void foo_ptr() {
+  const char *const ptr = 0;
+  char *t0 = const_cast<char *>(ptr); // no warning
+
+  volatile char *ptr2 = 0;
+  char *t1 = const_cast<char *>(ptr2); // no warning
+
+  const volatile char *ptr3 = 0;
+  char *t2 = const_cast<char *>(ptr3); // no warning
+}
+
+void cstr() {
+  void* p0 = (void*)(const void*)"txt"; // expected-warning {{cast from 'const void *' to 'void *' drops const qualifier}}
+  void* p1 = (void*)"txt"; // FIXME
+  char* p2 = (char*)"txt"; // expected-warning {{cast from 'const char *' to 'char *' drops const qualifier}}
+}
+
+void foo_0() {
+  const int a = 0;
+
+  const int &a0 = a;              // no warning
+  const int &a1 = (const int &)a; // no warning
+
+  int &a2 = (int &)a;                      // expected-warning {{cast from 'const int' to 'int &' drops const qualifier}}
+  const int &a3 = (int &)a;                // expected-warning {{cast from 'const int' to 'int &' drops const qualifier}}
+  int &a4 = (int &)((const int &)a);       // expected-warning {{cast from 'const int' to 'int &' drops const qualifier}}
+  int &a5 = (int &)((int &)a);             // expected-warning {{cast from 'const int' to 'int &' drops const qualifier}}
+  const int &a6 = (int &)((int &)a);       // expected-warning {{cast from 'const int' to 'int &' drops const qualifier}}
+  const int &a7 = (int &)((const int &)a); // expected-warning {{cast from 'const int' to 'int &' drops const qualifier}}
+  const int &a8 = (const int &)((int &)a); // expected-warning {{cast from 'const int' to 'int &' drops const qualifier}}
+}
+
+void foo_1() {
+  volatile int a = 0;
+
+  volatile int &a0 = a;                 // no warning
+  volatile int &a1 = (volatile int &)a; // no warning
+
+  int &a2 = (int &)a;                            // expected-warning {{cast from 'volatile int' to 'int &' drops volatile qualifier}}
+  volatile int &a3 = (int &)a;                   // expected-warning {{cast from 'volatile int' to 'int &' drops volatile qualifier}}
+  int &a4 = (int &)((volatile int &)a);          // expected-warning {{cast from 'volatile int' to 'int &' drops volatile qualifier}}
+  int &a5 = (int &)((int &)a);                   // expected-warning {{cast from 'volatile int' to 'int &' drops volatile qualifier}}
+  volatile int &a6 = (int &)((int &)a);          // expected-warning {{cast from 'volatile int' to 'int &' drops volatile qualifier}}
+  volatile int &a7 = (int &)((volatile int &)a); // expected-warning {{cast from 'volatile int' to 'int &' drops volatile qualifier}}
+  volatile int &a8 = (volatile int &)((int &)a); // expected-warning {{cast from 'volatile int' to 'int &' drops volatile qualifier}}
+}
+
+void foo_2() {
+  const volatile int a = 0;
+
+  const volatile int &a0 = a;                       // no warning
+  const volatile int &a1 = (const volatile int &)a; // no warning
+
+  int &a2 = (int &)a;                                        // expected-warning {{cast from 'const volatile int' to 'int &' drops const and volatile qualifiers}}
+  const volatile int &a3 = (int &)a;                         // expected-warning {{cast from 'const volatile int' to 'int &' drops const and volatile qualifiers}}
+  int &a4 = (int &)((const volatile int &)a);                // expected-warning {{cast from 'const volatile int' to 'int &' drops const and volatile qualifiers}}
+  int &a5 = (int &)((int &)a);                               // expected-warning {{cast from 'const volatile int' to 'int &' drops const and volatile qualifiers}}
+  const volatile int &a6 = (int &)((int &)a);                // expected-warning {{cast from 'const volatile int' to 'int &' drops const and volatile qualifiers}}
+  const volatile int &a7 = (int &)((const volatile int &)a); // expected-warning {{cast from 'const volatile int' to 'int &' drops const and volatile qualifiers}}
+  const volatile int &a8 = (const volatile int &)((int &)a); // expected-warning {{cast from 'const volatile int' to 'int &' drops const and volatile qualifiers}}
+}
+
+void bar_0() {
+  const int *_a = 0;
+  const int **a = &_a;
+
+  int **a0 = (int **)((const int **)a); // expected-warning {{cast from 'const int *' to 'int *' drops const qualifier}}
+  int **a1 = (int **)((int **)a);       // expected-warning {{cast from 'const int *' to 'int *' drops const qualifier}}
+
+  // const int **a2 = (int **)((int **)a);
+  // const int **a3 = (int **)((const int **)a);
+
+  const int **a4 = (const int **)((int **)a);        // expected-warning {{cast from 'const int *' to 'int *' drops const qualifier}} expected-warning {{cast from 'int **' to 'const int **' must have all intermediate pointers const qualified to be safe}}
+  const int **a5 = (const int **)((const int **)a); // no warning
+}
+
+void bar_1() {
+  const int *_a = 0;
+  const int *&a = _a;
+
+  int *&a0 = (int *&)((const int *&)a); // expected-warning {{cast from 'const int *' to 'int *' drops const qualifier}}
+  int *&a1 = (int *&)((int *&)a);       // expected-warning {{cast from 'const int *' to 'int *' drops const qualifier}}
+
+  // const int *&a2 = (int *&)((int *&)a);
+  // const int *&a3 = (int *&)((const int *&)a);
+
+  const int *&a4 = (const int *&)((int *&)a);        // expected-warning {{cast from 'const int *' to 'int *' drops const qualifier}} expected-warning {{cast from 'int *' to 'const int *&' must have all intermediate pointers const qualified to be safe}}
+  const int *&a5 = (const int *&)((const int *&)a); // no warning
+}
+
+void baz_0() {
+  struct C {
+    void A() {}
+    void B() const {}
+  };
+
+  const C S;
+  S.B();
+
+  ((C &)S).B(); // expected-warning {{cast from 'const C' to 'C &' drops const qualifier}}
+  ((C &)S).A(); // expected-warning {{cast from 'const C' to 'C &' drops const qualifier}}
+
+  ((C *)&S)->B(); // expected-warning {{cast from 'const C *' to 'C *' drops const qualifier}}
+  ((C *)&S)->A(); // expected-warning {{cast from 'const C *' to 'C *' drops const qualifier}}
+}
+
+void baz_1() {
+  struct C {
+    const int a;
+    int b;
+
+    C() : a(0) {}
+  };
+
+  {
+    C S;
+    S.b = 0;
+
+    (int &)(S.a) = 0; // expected-warning {{cast from 'const int' to 'int &' drops const qualifier}}
+    (int &)(S.b) = 0; // no warning
+
+    *(int *)(&S.a) = 0; // expected-warning {{cast from 'const int *' to 'int *' drops const qualifier}}
+    *(int *)(&S.b) = 0; // no warning
+  }
+  {
+    const C S;
+
+    (int &)(S.a) = 0; // expected-warning {{cast from 'const int' to 'int &' drops const qualifier}}
+    (int &)(S.b) = 0; // expected-warning {{cast from 'const int' to 'int &' drops const qualifier}}
+
+    *(int *)(&S.a) = 0; // expected-warning {{cast from 'const int *' to 'int *' drops const qualifier}}
+    *(int *)(&S.b) = 0; // expected-warning {{cast from 'const int *' to 'int *' drops const qualifier}}
+  }
+}
diff --git a/test/SemaCXX/warn-loop-analysis.cpp b/test/SemaCXX/warn-loop-analysis.cpp
index 25ec7a7862ed..2934003848a2 100644
--- a/test/SemaCXX/warn-loop-analysis.cpp
+++ b/test/SemaCXX/warn-loop-analysis.cpp
@@ -202,6 +202,12 @@ void test7() {
     if (true) continue;
     i--;
   }
+
+  // But do warn if the continue is in a nested loop.
+  for (;;i--) { // expected-note{{decremented here}}
+    for (int j = 0; j < 10; ++j) continue;
+    i--; // expected-warning{{decremented both}}
+  }
 }
 
 struct iterator {
@@ -259,6 +265,12 @@ void test8() {
     if (true) continue;
     i--;
   }
+
+  // But do warn if the continue is in a nested loop.
+  for (;;i--) { // expected-note{{decremented here}}
+    for (int j = 0; j < 10; ++j) continue;
+    i--; // expected-warning{{decremented both}}
+  }
 }
 
 int f(int);
diff --git a/test/SemaCXX/warn-throw-out-noexcept-func.cpp b/test/SemaCXX/warn-throw-out-noexcept-func.cpp
index fc2919a1e327..67ffbd93940a 100644
--- a/test/SemaCXX/warn-throw-out-noexcept-func.cpp
+++ b/test/SemaCXX/warn-throw-out-noexcept-func.cpp
@@ -2,8 +2,8 @@
 struct A_ShouldDiag {
   ~A_ShouldDiag(); // implicitly noexcept(true)
 };
-A_ShouldDiag::~A_ShouldDiag() { // expected-note {{destructor or deallocator has a (possibly implicit) non-throwing excepton specification}}
-  throw 1; // expected-warning {{has a non-throwing exception specification but can still throw, resulting in unexpected program termination}}
+A_ShouldDiag::~A_ShouldDiag() { // expected-note {{destructor has a implicit non-throwing exception specification}}
+  throw 1; // expected-warning {{has a non-throwing exception specification but can still throw}}
 }
 struct B_ShouldDiag {
   int i;
@@ -11,7 +11,7 @@ struct B_ShouldDiag {
 };
 struct R_ShouldDiag : A_ShouldDiag {
   B_ShouldDiag b;
-  ~R_ShouldDiag() {     // expected-note  {{destructor or deallocator has a}}
+  ~R_ShouldDiag() { // expected-note  {{destructor has a implicit non-throwing exception specification}}
     throw 1; // expected-warning {{has a non-throwing exception specification but}}
   }
 };
@@ -30,18 +30,18 @@ struct N_ShouldDiag {
   ~N_ShouldDiag(); //implicitly noexcept(true)
 };
 
-N_ShouldDiag::~N_ShouldDiag() {  // expected-note  {{destructor or deallocator has a}}
+N_ShouldDiag::~N_ShouldDiag() { // expected-note  {{destructor has a implicit non-throwing exception specification}}
   throw 1; // expected-warning {{has a non-throwing exception specification but}}
 }
 struct X_ShouldDiag {
   B_ShouldDiag b;
-  ~X_ShouldDiag() noexcept { // expected-note  {{destructor or deallocator has a}}
-    throw 1;      // expected-warning {{has a non-throwing exception specification but}}
+  ~X_ShouldDiag() noexcept { // expected-note  {{destructor has a non-throwing exception}}
+    throw 1; // expected-warning {{has a non-throwing exception specification but}}
   }
 };
 struct Y_ShouldDiag : A_ShouldDiag {
-  ~Y_ShouldDiag() noexcept(true) { // expected-note  {{destructor or deallocator has a}}
-    throw 1;            // expected-warning {{has a non-throwing exception specification but}}
+  ~Y_ShouldDiag() noexcept(true) { // expected-note  {{destructor has a non-throwing exception specification}}
+    throw 1; // expected-warning {{has a non-throwing exception specification but}}
   }
 };
 struct C_ShouldNotDiag {
@@ -54,7 +54,7 @@ struct D_ShouldNotDiag {
     throw 1;
   }
 };
-struct E_ShouldNotDiag  {
+struct E_ShouldNotDiag {
   C_ShouldNotDiag c;
   ~E_ShouldNotDiag(); //implicitly noexcept(false)
 };
@@ -68,7 +68,7 @@ class A1_ShouldDiag {
   T b;
 
 public:
-  ~A1_ShouldDiag() {    // expected-note  {{destructor or deallocator has a}}
+  ~A1_ShouldDiag() { // expected-note  {{destructor has a implicit non-throwing exception specification}}
     throw 1; // expected-warning {{has a non-throwing exception specification but}}
   }
 };
@@ -81,19 +81,19 @@ template <typename T>
 struct R1_ShouldDiag : A1_ShouldDiag<T> //expected-note {{in instantiation of member function}}
 {
   B1_ShouldDiag<T> b;
-  ~R1_ShouldDiag() {    // expected-note  {{destructor or deallocator has a}}
+  ~R1_ShouldDiag() { // expected-note  {{destructor has a implicit non-throwing exception specification}}
     throw 1; // expected-warning {{has a non-throwing exception specification but}}
   }
 };
 template <typename T>
 struct S1_ShouldDiag : A1_ShouldDiag<T> {
   B1_ShouldDiag<T> b;
-  ~S1_ShouldDiag() noexcept { // expected-note  {{destructor or deallocator has a}}
-    throw 1;       // expected-warning {{has a non-throwing exception specification but}}
+  ~S1_ShouldDiag() noexcept { // expected-note  {{destructor has a non-throwing exception specification}}
+    throw 1; // expected-warning {{has a non-throwing exception specification but}}
   }
 };
-void operator delete(void *ptr) noexcept { // expected-note  {{destructor or deallocator has a}}
-  throw 1;                                 // expected-warning {{has a non-throwing exception specification but}}
+void operator delete(void *ptr) noexcept { // expected-note  {{deallocator has a non-throwing exception specification}}
+  throw 1; // expected-warning {{has a non-throwing exception specification but}}
 }
 struct except_fun {
   static const bool i = false;
@@ -109,18 +109,18 @@ struct dependent_warn {
 };
 template <typename T>
 struct dependent_warn_noexcept {
-  ~dependent_warn_noexcept() noexcept(T::i) { // expected-note  {{destructor or deallocator has a}}
-    throw 1;                                  // expected-warning {{has a non-throwing exception specification but}}
+  ~dependent_warn_noexcept() noexcept(T::i) { // expected-note  {{destructor has a non-throwing exception specification}}
+    throw 1; // expected-warning {{has a non-throwing exception specification but}}
   }
 };
 template <typename T>
 struct dependent_warn_both {
-  ~dependent_warn_both() noexcept(T::i) { // expected-note  {{destructor or deallocator has a}}
-    throw 1;                              // expected-warning {{has a non-throwing exception specification but}}
+  ~dependent_warn_both() noexcept(T::i) { // expected-note  {{destructor has a non-throwing exception specification}}
+    throw 1; // expected-warning {{has a non-throwing exception specification but}}
   }
 };
-void foo() noexcept { //expected-note {{non-throwing function declare here}}
-  throw 1;            // expected-warning {{has a non-throwing exception specification but}}
+void foo() noexcept { //expected-note {{function declared non-throwing here}}
+  throw 1; // expected-warning {{has a non-throwing exception specification but}}
 }
 struct Throws {
   ~Throws() noexcept(false);
@@ -128,14 +128,14 @@ struct Throws {
 
 struct ShouldDiagnose {
   Throws T;
-  ~ShouldDiagnose() noexcept { //expected-note {{destructor or deallocator has a}}
+  ~ShouldDiagnose() noexcept { //expected-note {{destructor has a non-throwing exception specification}}
     throw; // expected-warning {{has a non-throwing exception specification but}}
   }
 };
 struct ShouldNotDiagnose {
   Throws T;
-  ~ShouldNotDiagnose() { 
-    throw; 
+  ~ShouldNotDiagnose() {
+    throw;
   }
 };
 
@@ -158,21 +158,21 @@ void g_ShouldNotDiag() noexcept {
   }
 }
 
-void h_ShouldDiag() noexcept { //expected-note {{non-throwing function declare here}}
+void h_ShouldDiag() noexcept { //expected-note {{function declared non-throwing here}}
   try {
     throw 12; // expected-warning {{has a non-throwing exception specification but}}
   } catch (const char *) {
   }
 }
 
-void i_ShouldDiag() noexcept { //expected-note {{non-throwing function declare here}}
+void i_ShouldDiag() noexcept { //expected-note {{function declared non-throwing here}}
   try {
     throw 12;
   } catch (int) {
     throw; // expected-warning {{has a non-throwing exception specification but}}
   }
 }
-void j_ShouldDiag() noexcept { //expected-note {{non-throwing function declare here}}
+void j_ShouldDiag() noexcept { //expected-note {{function declared non-throwing here}}
   try {
     throw 12;
   } catch (int) {
@@ -180,7 +180,7 @@ void j_ShouldDiag() noexcept { //expected-note {{non-throwing function declare h
   }
 }
 
-void k_ShouldDiag() noexcept { //expected-note {{non-throwing function declare here}}
+void k_ShouldDiag() noexcept { //expected-note {{function declared non-throwing here}}
   try {
     throw 12;
   } catch (...) {
@@ -188,7 +188,7 @@ void k_ShouldDiag() noexcept { //expected-note {{non-throwing function declare h
   }
 }
 
-void loo_ShouldDiag(int i) noexcept { //expected-note {{non-throwing function declare here}}
+void loo_ShouldDiag(int i) noexcept { //expected-note {{function declared non-throwing here}}
   if (i)
     try {
       throw 12;
@@ -203,13 +203,13 @@ void loo1_ShouldNotDiag() noexcept {
     throw 12;
 }
 
-void loo2_ShouldDiag() noexcept { //expected-note {{non-throwing function declare here}}
+void loo2_ShouldDiag() noexcept { //expected-note {{function declared non-throwing here}}
   if (1)
     throw 12; // expected-warning {{has a non-throwing exception specification but}}
 }
 struct S {};
 
-void l_ShouldDiag() noexcept { //expected-note {{non-throwing function declare here}}
+void l_ShouldDiag() noexcept { //expected-note {{function declared non-throwing here}}
   try {
     throw S{}; //expected-warning {{has a non-throwing exception specification but}}
   } catch (S *s) {
@@ -222,7 +222,6 @@ void m_ShouldNotDiag() noexcept {
     throw s;
   } catch (S s) {
   }
-
 }
 void n_ShouldNotDiag() noexcept {
   try {
@@ -231,7 +230,7 @@ void n_ShouldNotDiag() noexcept {
   } catch (const S &s) {
   }
 }
-void o_ShouldDiag() noexcept { //expected-note {{non-throwing function declare here}}
+void o_ShouldDiag() noexcept { //expected-note {{function declared non-throwing here}}
   try {
     throw; //expected-warning {{has a non-throwing exception specification but}}
   } catch (...) {
@@ -239,7 +238,7 @@ void o_ShouldDiag() noexcept { //expected-note {{non-throwing function declare h
 }
 
 #define NOEXCEPT noexcept
-void with_macro() NOEXCEPT { //expected-note {{non-throwing function declare here}}
+void with_macro() NOEXCEPT { //expected-note {{function declared non-throwing here}}
   throw 1; // expected-warning {{has a non-throwing exception specification but}}
 }
 
@@ -248,8 +247,8 @@ void with_try_block() try {
 } catch (...) {
 }
 
-void with_try_block1() noexcept try { //expected-note {{non-throwing function declare here}}
-  throw 2;  // expected-warning {{has a non-throwing exception specification but}}
+void with_try_block1() noexcept try { //expected-note {{function declared non-throwing here}}
+  throw 2; // expected-warning {{has a non-throwing exception specification but}}
 } catch (char *) {
 }
 
@@ -272,20 +271,20 @@ void goodPointer() noexcept {
     throw &d;
   } catch (B *) {}
 }
-void badPlain() noexcept { // expected-note {{non-throwing function declare here}}
+void badPlain() noexcept { //expected-note {{function declared non-throwing here}}
   try {
-    throw B(); // expected-warning {{'badPlain' has a non-throwing exception specification but can still throw, resulting in unexpected program termination}}
+    throw B(); // expected-warning {{'badPlain' has a non-throwing exception specification but can still throw}}
   } catch (D) {}
 }
-void badReference() noexcept { // expected-note {{non-throwing function declare here}}
+void badReference() noexcept { //expected-note {{function declared non-throwing here}}
   try {
-    throw B(); // expected-warning {{'badReference' has a non-throwing exception specification but can still throw, resulting in unexpected program termination}}
+    throw B(); // expected-warning {{'badReference' has a non-throwing exception specification but can still throw}}
   } catch (D &) {}
 }
-void badPointer() noexcept { // expected-note {{non-throwing function declare here}}
+void badPointer() noexcept { //expected-note {{function declared non-throwing here}}
   B b;
   try {
-    throw &b; // expected-warning {{'badPointer' has a non-throwing exception specification but can still throw, resulting in unexpected program termination}}
+    throw &b; // expected-warning {{'badPointer' has a non-throwing exception specification but can still throw}}
   } catch (D *) {}
 }
 }
diff --git a/test/SemaObjC/attr-availability.m b/test/SemaObjC/attr-availability.m
index 02b7c5c81491..1245ac7409b8 100644
--- a/test/SemaObjC/attr-availability.m
+++ b/test/SemaObjC/attr-availability.m
@@ -13,7 +13,7 @@
 @interface A <P>
 - (void)method __attribute__((availability(macosx,introduced=10.1,deprecated=10.2))); // expected-note {{'method' has been explicitly marked deprecated here}}
 #if defined(WARN_PARTIAL)
-  // expected-note@+2 {{'partialMethod' has been explicitly marked partial here}}
+  // expected-note@+2 2 {{'partialMethod' has been explicitly marked partial here}}
 #endif
 - (void)partialMethod __attribute__((availability(macosx,introduced=10.8)));
 
@@ -66,7 +66,10 @@ void f(A *a, B *b) {
 @end
 
 void f_after_redecl(A *a, B *b) {
-  [a partialMethod]; // no warning
+#ifdef WARN_PARTIAL
+  // expected-warning@+2{{'partialMethod' is only available on macOS 10.8 or newer}} expected-note@+2 {{@available}}
+#endif
+  [a partialMethod];
   [b partialMethod]; // no warning
   [a partial_proto_method]; // no warning
   [b partial_proto_method]; // no warning
@@ -133,6 +136,10 @@ id NSNibOwner, topNibObjects;
 @end
 
 @interface PartialI <PartialProt>
+#ifdef WARN_PARTIAL
+// expected-note@+3{{marked partial here}}
+// expected-note@+3{{marked partial here}}
+#endif
 - (void)partialMethod __attribute__((availability(macosx,introduced=10.8)));
 + (void)partialMethod __attribute__((availability(macosx,introduced=10.8)));
 @end
@@ -160,14 +167,20 @@ id NSNibOwner, topNibObjects;
 @end
 
 void partialfun(PartialI* a) {
-  [a partialMethod]; // no warning
+#ifdef WARN_PARTIAL
+  // expected-warning@+2 {{'partialMethod' is only available on macOS 10.8 or newer}} expected-note@+2{{@available}}
+#endif
+  [a partialMethod];
   [a ipartialMethod1]; // no warning
 #if defined(WARN_PARTIAL)
   // expected-warning@+2 {{'ipartialMethod2' is only available on macOS 10.8 or newer}} expected-note@+2 {{enclose 'ipartialMethod2' in an @available check to silence this warning}}
 #endif
   [a ipartialMethod2];
   [a ppartialMethod]; // no warning
-  [PartialI partialMethod]; // no warning
+#ifdef WARN_PARTIAL
+  // expected-warning@+2 {{'partialMethod' is only available on macOS 10.8 or newer}} expected-note@+2 {{@available}}
+#endif
+  [PartialI partialMethod];
   [PartialI ipartialMethod1]; // no warning
 #if defined(WARN_PARTIAL)
   // expected-warning@+2 {{'ipartialMethod2' is only available on macOS 10.8 or newer}} expected-note@+2 {{enclose 'ipartialMethod2' in an @available check to silence this warning}}
@@ -177,20 +190,23 @@ void partialfun(PartialI* a) {
 }
 
 #if defined(WARN_PARTIAL)
-  // expected-note@+2 {{'PartialI2' has been explicitly marked partial here}}
+  // expected-note@+2 2 {{'PartialI2' has been explicitly marked partial here}}
 #endif
 __attribute__((availability(macosx, introduced = 10.8))) @interface PartialI2
 @end
 
 #if defined(WARN_PARTIAL)
-  // expected-warning@+2 {{'PartialI2' is partial: introduced in macOS 10.8}} expected-note@+2 {{explicitly redeclare 'PartialI2' to silence this warning}}
+// expected-warning@+2 {{'PartialI2' is partial: introduced in macOS 10.8}} expected-note@+2 {{annotate 'partialinter1' with an availability attribute to silence}}
 #endif
 void partialinter1(PartialI2* p) {
 }
 
 @class PartialI2;
 
-void partialinter2(PartialI2* p) { // no warning
+#ifdef WARN_PARTIAL
+// expected-warning@+2 {{'PartialI2' is partial: introduced in macOS 10.8}} expected-note@+2 {{annotate 'partialinter2' with an availability attribute to silence}}
+#endif
+void partialinter2(PartialI2* p) {
 }
 
 
diff --git a/test/SemaObjC/default-synthesize-3.m b/test/SemaObjC/default-synthesize-3.m
index fe2b35f61523..9a05408aa060 100644
--- a/test/SemaObjC/default-synthesize-3.m
+++ b/test/SemaObjC/default-synthesize-3.m
@@ -173,13 +173,13 @@ typedef NSObject<Fooing> FooObject;
 @end
 
 @implementation Okay // expected-warning {{auto property synthesis will not synthesize property 'muahahaha' declared in protocol 'Fooing'}} expected-warning {{auto property synthesis will not synthesize property 'hoho' declared in protocol 'SubFooling'}}
-@end
+@end // expected-note 2 {{add a '@synthesize' directive}}
 
 @interface Fail : FooObject
 @end
 
 @implementation Fail // expected-warning {{auto property synthesis will not synthesize property 'muahahaha' declared in protocol 'Fooing'}} expected-warning {{auto property synthesis will not synthesize property 'hoho' declared in protocol 'SubFooling'}}
-@end
+@end // expected-note 2 {{add a '@synthesize' directive}}
 
 // rdar://16089191
 @class NSURL;
diff --git a/test/SemaObjC/default-synthesize.m b/test/SemaObjC/default-synthesize.m
index 3f0ae0261daf..61ce9317c519 100644
--- a/test/SemaObjC/default-synthesize.m
+++ b/test/SemaObjC/default-synthesize.m
@@ -137,7 +137,7 @@
 @end
  
 @implementation MyClass // expected-warning {{auto property synthesis will not synthesize property 'requiredString' declared in protocol 'MyProtocol'}}
-@end
+@end // expected-note {{add a '@synthesize' directive}}
 
 // rdar://18152478
 @protocol NSObject @end
diff --git a/test/SemaObjC/forward-protocol-incomplete-impl-warn.m b/test/SemaObjC/forward-protocol-incomplete-impl-warn.m
index c235e32316a9..583bb4dd891d 100644
--- a/test/SemaObjC/forward-protocol-incomplete-impl-warn.m
+++ b/test/SemaObjC/forward-protocol-incomplete-impl-warn.m
@@ -17,4 +17,4 @@
 
 @implementation IBImageCatalogDocument // expected-warning {{auto property synthesis will not synthesize property 'Prop' declared in protocol 'DVTInvalidation'}} \
 				       // expected-warning {{method 'invalidate' in protocol 'DVTInvalidation' not implemented}}
-@end
+@end // expected-note {{add a '@synthesize' directive}}
diff --git a/test/SemaObjC/objc-container-subscripting-1.m b/test/SemaObjC/objc-container-subscripting-1.m
index a58a7c3bda81..b5ed5a68dd84 100644
--- a/test/SemaObjC/objc-container-subscripting-1.m
+++ b/test/SemaObjC/objc-container-subscripting-1.m
@@ -16,8 +16,7 @@ id oldObject = array[10]; // expected-warning {{instance method '-objectAtIndexe
 array[10] = 0; // expected-warning {{instance method '-setObject:atIndexedSubscript:' not found (return type defaults to 'id')}}
 
 id<P> p_array;
-oldObject = p_array[10]; // expected-warning {{instance method '-objectAtIndexedSubscript:' not found (return type defaults to 'id')}}
+oldObject = p_array[10]; // expected-error {{expected method to read array element not found on object of type 'id<P>'}}
 
-p_array[10] = 0; // expected-warning {{instance method '-setObject:atIndexedSubscript:' not found (return type defaults to 'id')}}
+p_array[10] = 0; // expected-error {{expected method to write array element not found on object of type 'id<P>'}}
 }
-
diff --git a/test/SemaObjC/objc-container-subscripting-2.m b/test/SemaObjC/objc-container-subscripting-2.m
index 62320fcebb77..6e7ee7c40e58 100644
--- a/test/SemaObjC/objc-container-subscripting-2.m
+++ b/test/SemaObjC/objc-container-subscripting-2.m
@@ -28,3 +28,22 @@ void test_unused() {
   dict[array]; // expected-warning {{container access result unused - container access should not be used for side effects}}
 }
 
+void testQualifiedId(id<P> qualifiedId) {
+  id object = qualifiedId[10];   // expected-error {{expected method to read array element not found on object of type 'id<P>'}}
+  qualifiedId[10] = qualifiedId; // expected-error {{expected method to write array element not found on object of type 'id<P>'}}
+}
+
+void testUnqualifiedId(id unqualId) {
+  id object = unqualId[10];
+  unqualId[10] = object;
+}
+
+@protocol Subscriptable
+- (id)objectAtIndexedSubscript:(size_t)index;
+- (void)setObject:(id)object atIndexedSubscript:(size_t)index;
+@end
+
+void testValidQualifiedId(id<Subscriptable> qualifiedId) {
+  id object = qualifiedId[10];
+  qualifiedId[10] = object;
+}
diff --git a/test/SemaObjC/unguarded-availability-new.m b/test/SemaObjC/unguarded-availability-new.m
index 33baedebc279..9c44b164d541 100644
--- a/test/SemaObjC/unguarded-availability-new.m
+++ b/test/SemaObjC/unguarded-availability-new.m
@@ -96,16 +96,16 @@ typedef int AVAILABLE_NEXT new_int;
 FUNC_AVAILABLE new_int x;
 #ifndef NO_WARNING
 #ifdef MAC
-  // expected-warning@-3 {{'new_int' is partial: introduced in macOS 10.14}} expected-note@-3 {{explicitly redeclare 'new_int' to silence this warning}}
+  // expected-warning@-3 {{'new_int' is partial: introduced in macOS 10.14}} expected-note@-3 {{annotate 'x' with an availability attribute to silence}}
 #endif
 #ifdef IOS
-  // expected-warning@-6 {{'new_int' is partial: introduced in iOS 12}} expected-note@-6 {{explicitly redeclare 'new_int' to silence this warning}}
+  // expected-warning@-6 {{'new_int' is partial: introduced in iOS 12}} expected-note@-6 {{annotate 'x' with an availability attribute to silence}}
 #endif
 #ifdef TVOS
-  // expected-warning@-9 {{'new_int' is partial: introduced in tvOS 13}} expected-note@-9 {{explicitly redeclare 'new_int' to silence this warning}}
+  // expected-warning@-9 {{'new_int' is partial: introduced in tvOS 13}} expected-note@-9 {{annotate 'x' with an availability attribute to silence}}
 #endif
 #ifdef WATCHOS
-  // expected-warning@-12 {{'new_int' is partial: introduced in watchOS 5}} expected-note@-12 {{explicitly redeclare 'new_int' to silence this warning}}
+  // expected-warning@-12 {{'new_int' is partial: introduced in watchOS 5}} expected-note@-12 {{annotate 'x' with an availability attribute to silence}}
 #endif
 #endif
 
diff --git a/test/SemaObjC/unguarded-availability.m b/test/SemaObjC/unguarded-availability.m
index 5febee038e3c..139b79158afe 100644
--- a/test/SemaObjC/unguarded-availability.m
+++ b/test/SemaObjC/unguarded-availability.m
@@ -5,6 +5,8 @@
 #define AVAILABLE_10_11 __attribute__((availability(macos, introduced = 10.11)))
 #define AVAILABLE_10_12 __attribute__((availability(macos, introduced = 10.12)))
 
+typedef int AVAILABLE_10_12 new_int; // expected-note + {{marked partial here}}
+
 int func_10_11() AVAILABLE_10_11; // expected-note 4 {{'func_10_11' has been explicitly marked partial here}}
 
 #ifdef OBJCPP
@@ -70,9 +72,9 @@ void use_typedef() {
 }
 
 __attribute__((objc_root_class))
-AVAILABLE_10_11 @interface Class_10_11 {
+AVAILABLE_10_11 @interface Class_10_11 { // expected-note{{annotate 'Class_10_11' with an availability attribute to silence}}
   int_10_11 foo;
-  int_10_12 bar; // expected-warning {{'int_10_12' is partial: introduced in macOS 10.12}} expected-note{{redeclare}}
+  int_10_12 bar; // expected-warning {{'int_10_12' is partial: introduced in macOS 10.12}}
 }
 - (void)method1;
 - (void)method2;
@@ -125,7 +127,7 @@ void test_blocks() {
   };
 }
 
-void test_params(int_10_12 x); // expected-warning {{'int_10_12' is partial: introduced in macOS 10.12}} expected-note{{redeclare}}
+void test_params(int_10_12 x); // expected-warning {{'int_10_12' is partial: introduced in macOS 10.12}} expected-note{{annotate 'test_params' with an availability attribute to silence}}
 
 void test_params2(int_10_12 x) AVAILABLE_10_12; // no warn
 
@@ -234,3 +236,54 @@ void functionInFunction() {
 }
 
 #endif
+
+struct InStruct { // expected-note{{annotate 'InStruct' with an availability attribute to silence}}
+  new_int mem; // expected-warning{{'new_int' is partial}}
+
+  struct { new_int mem; } anon; // expected-warning{{'new_int' is partial}} expected-note{{annotate anonymous struct with an availability attribute}}
+};
+
+#ifdef OBJCPP
+static constexpr int AVAILABLE_10_12 SomeConstexprValue = 2; // expected-note{{marked partial here}}
+typedef enum { // expected-note{{annotate anonymous enum with an availability attribute}}
+  SomeValue = SomeConstexprValue // expected-warning{{'SomeConstexprValue' is partial}} 
+} SomeEnum;
+#endif
+
+@interface InInterface
+-(new_int)meth; // expected-warning{{'new_int' is partial}} expected-note{{annotate 'meth' with an availability attribute}}
+@end
+
+@interface Proper // expected-note{{annotate 'Proper' with an availability attribute}}
+@property (class) new_int x; // expected-warning{{'new_int' is partial}}
+@end
+
+void with_local_struct() {
+  struct local { // expected-note{{annotate 'local' with an availability attribute}}
+    new_int x; // expected-warning{{'new_int' is partial}}
+  };
+}
+
+// rdar://33156429:
+// Avoid the warning on protocol requirements.
+
+AVAILABLE_10_12
+@protocol NewProtocol // expected-note {{'NewProtocol' has been explicitly marked partial here}}
+@end
+
+@protocol ProtocolWithNewProtocolRequirement <NewProtocol> // expected-note {{annotate 'ProtocolWithNewProtocolRequirement' with an availability attribute to silence}}
+
+@property(copy) id<NewProtocol> prop; // expected-warning {{'NewProtocol' is partial: introduced in macOS 10.12}}
+
+@end
+
+@interface BaseClass
+@end
+
+@interface ClassWithNewProtocolRequirement : BaseClass <NewProtocol>
+
+@end
+
+@interface BaseClass (CategoryWithNewProtocolRequirement) <NewProtocol>
+
+@end
diff --git a/test/SemaOpenCL/cl20-device-side-enqueue.cl b/test/SemaOpenCL/cl20-device-side-enqueue.cl
index bafbe447ee28..3f6527afeadc 100644
--- a/test/SemaOpenCL/cl20-device-side-enqueue.cl
+++ b/test/SemaOpenCL/cl20-device-side-enqueue.cl
@@ -19,19 +19,19 @@ kernel void enqueue_kernel_tests() {
     return 0;
   });
 
-  enqueue_kernel(vptr, flags, ndrange, ^(void) { // expected-error{{illegal call to enqueue_kernel, expected 'queue_t' argument type}}
+  enqueue_kernel(vptr, flags, ndrange, ^(void) { // expected-error{{illegal call to 'enqueue_kernel', expected 'queue_t' argument type}}
     return 0;
   });
 
-  enqueue_kernel(default_queue, vptr, ndrange, ^(void) { // expected-error{{illegal call to enqueue_kernel, expected 'kernel_enqueue_flags_t' (i.e. uint) argument type}}
+  enqueue_kernel(default_queue, vptr, ndrange, ^(void) { // expected-error{{illegal call to 'enqueue_kernel', expected 'kernel_enqueue_flags_t' (i.e. uint) argument type}}
     return 0;
   });
 
-  enqueue_kernel(default_queue, flags, vptr, ^(void) { // expected-error{{illegal call to enqueue_kernel, expected 'ndrange_t' argument type}}
+  enqueue_kernel(default_queue, flags, vptr, ^(void) { // expected-error{{illegal call to 'enqueue_kernel', expected 'ndrange_t' argument type}}
     return 0;
   });
 
-  enqueue_kernel(default_queue, flags, ndrange, vptr); // expected-error{{illegal call to enqueue_kernel, expected block argument}}
+  enqueue_kernel(default_queue, flags, ndrange, vptr); // expected-error{{illegal call to 'enqueue_kernel', expected block argument}}
 
   enqueue_kernel(default_queue, flags, ndrange, ^(int i) { // expected-error{{blocks with parameters are not accepted in this prototype of enqueue_kernel call}}
     return 0;
@@ -46,21 +46,21 @@ kernel void enqueue_kernel_tests() {
                                                            return 0;
                                                          });
 
-  enqueue_kernel(default_queue, flags, ndrange, vptr, &event_wait_list, &evt, ^(void) { // expected-error{{illegal call to enqueue_kernel, expected integer argument type}}
+  enqueue_kernel(default_queue, flags, ndrange, vptr, &event_wait_list, &evt, ^(void) { // expected-error{{illegal call to 'enqueue_kernel', expected integer argument type}}
     return 0;
   });
 
-  enqueue_kernel(default_queue, flags, ndrange, 1, vptr, &evt, ^(void) // expected-error{{illegal call to enqueue_kernel, expected 'clk_event_t *' argument type}}
+  enqueue_kernel(default_queue, flags, ndrange, 1, vptr, &evt, ^(void) // expected-error{{illegal call to 'enqueue_kernel', expected 'clk_event_t *' argument type}}
                                                                {
                                                                  return 0;
                                                                });
 
-  enqueue_kernel(default_queue, flags, ndrange, 1, &event_wait_list, vptr, ^(void) // expected-error{{illegal call to enqueue_kernel, expected 'clk_event_t *' argument type}}
+  enqueue_kernel(default_queue, flags, ndrange, 1, &event_wait_list, vptr, ^(void) // expected-error{{illegal call to 'enqueue_kernel', expected 'clk_event_t *' argument type}}
                                                                            {
                                                                              return 0;
                                                                            });
 
-  enqueue_kernel(default_queue, flags, ndrange, 1, &event_wait_list, &evt, vptr); // expected-error{{illegal call to enqueue_kernel, expected block argument}}
+  enqueue_kernel(default_queue, flags, ndrange, 1, &event_wait_list, &evt, vptr); // expected-error{{illegal call to 'enqueue_kernel', expected block argument}}
 
   // Testing the third overload type
   enqueue_kernel(default_queue, flags, ndrange,
diff --git a/test/SemaOpenCL/images.cl b/test/SemaOpenCL/images.cl
index f963de4e1359..9d6092c8c916 100644
--- a/test/SemaOpenCL/images.cl
+++ b/test/SemaOpenCL/images.cl
@@ -1,9 +1,32 @@
-// RUN: %clang_cc1 %s -verify -pedantic -fsyntax-only
+// RUN: %clang_cc1 %s -cl-std=CL2.0 -verify -pedantic -fsyntax-only
 
-void img2d_ro(__read_only image2d_t img) {} // expected-note{{passing argument to parameter 'img' here}} expected-note{{passing argument to parameter 'img' here}}
+void img2d_ro(read_only image2d_t); // expected-note 3{{passing argument to parameter here}}
+void img2d_wo(write_only image2d_t); // expected-note 2{{passing argument to parameter here}}
+void img2d_rw(read_write image2d_t); // expected-note 2{{passing argument to parameter here}}
+void img2d_default(image2d_t); // expected-note 2{{passing argument to parameter here}}
 
-void imgage_access_test(image2d_t img2dro, write_only image2d_t img2dwo, image3d_t img3dro) {
-  img2d_ro(img2dro);
-  img2d_ro(img2dwo); // expected-error{{passing '__write_only image2d_t' to parameter of incompatible type '__read_only image2d_t'}}
+void imgage_access_test(image2d_t img2dro, image3d_t img3dro) {
+  img2d_ro(img2dro); // read_only = read_only
   img2d_ro(img3dro); // expected-error{{passing '__read_only image3d_t' to parameter of incompatible type '__read_only image2d_t'}}
 }
+
+kernel void read_only_access_test(read_only image2d_t img) {
+  img2d_ro(img); // read_only = read_only
+  img2d_wo(img); // expected-error {{passing '__read_only image2d_t' to parameter of incompatible type '__write_only image2d_t'}}
+  img2d_rw(img); // expected-error {{passing '__read_only image2d_t' to parameter of incompatible type '__read_write image2d_t'}}
+  img2d_default(img); // read_only = read_only
+}
+
+kernel void write_only_access_test(write_only image2d_t img) {
+  img2d_ro(img); // expected-error {{passing '__write_only image2d_t' to parameter of incompatible type '__read_only image2d_t'}}
+  img2d_wo(img); // write_only = write_only
+  img2d_rw(img); // expected-error {{passing '__write_only image2d_t' to parameter of incompatible type '__read_write image2d_t'}}
+  img2d_default(img); // expected-error {{passing '__write_only image2d_t' to parameter of incompatible type '__read_only image2d_t'}}
+}
+
+kernel void read_write_access_test(read_write image2d_t img) {
+  img2d_ro(img);  // expected-error {{passing '__read_write image2d_t' to parameter of incompatible type '__read_only image2d_t'}}
+  img2d_wo(img); // expected-error {{passing '__read_write image2d_t' to parameter of incompatible type '__write_only image2d_t'}}
+  img2d_rw(img); //read_write = read_write
+  img2d_default(img); // expected-error {{passing '__read_write image2d_t' to parameter of incompatible type '__read_only image2d_t'}}
+}
diff --git a/test/SemaTemplate/constexpr-instantiate.cpp b/test/SemaTemplate/constexpr-instantiate.cpp
index dfb8a07d3b7d..31dbdb617a6a 100644
--- a/test/SemaTemplate/constexpr-instantiate.cpp
+++ b/test/SemaTemplate/constexpr-instantiate.cpp
@@ -191,7 +191,7 @@ namespace Unevaluated {
       static constexpr bool f() { return sizeof(T) < U::size; }
 
       template<typename U>
-      static typename enable_if<f<U>(), void>::type g() {} // expected-note {{disabled by 'enable_if'}}
+      static typename enable_if<f<U>(), void>::type g() {} // expected-note {{requirement 'f<Unevaluated::PR13423::U>()' was not satisfied}}
     };
 
     struct U { static constexpr int size = 2; };
diff --git a/test/SemaTemplate/overload-candidates.cpp b/test/SemaTemplate/overload-candidates.cpp
index c0abb5b17488..de998d74f9af 100644
--- a/test/SemaTemplate/overload-candidates.cpp
+++ b/test/SemaTemplate/overload-candidates.cpp
@@ -47,7 +47,7 @@ namespace boost {
   template<bool, typename = void> struct enable_if {};
   template<typename T> struct enable_if<true, T> { typedef T type; };
 }
-template<typename T> typename boost::enable_if<sizeof(T) == 4, int>::type if_size_4(); // expected-note{{candidate template ignored: disabled by 'enable_if' [with T = char]}}
+template<typename T> typename boost::enable_if<sizeof(T) == 4, int>::type if_size_4(); // expected-note{{candidate template ignored: requirement 'sizeof(char) == 4' was not satisfied [with T = char]}}
 int k = if_size_4<char>(); // expected-error{{no matching function}}
 
 namespace llvm {
@@ -61,7 +61,7 @@ void test_if_int() {
 }
 
 template<typename T> struct NonTemplateFunction {
-  typename boost::enable_if<sizeof(T) == 4, int>::type f(); // expected-error{{no type named 'type' in 'boost::enable_if<false, int>'; 'enable_if' cannot be used to disable this declaration}}
+  typename boost::enable_if<sizeof(T) == 4, int>::type f(); // expected-error{{failed requirement 'sizeof(char) == 4'; 'enable_if' cannot be used to disable this declaration}}
 };
 NonTemplateFunction<char> NTFC; // expected-note{{here}}
 
@@ -100,7 +100,7 @@ namespace PR15673 {
 #if __cplusplus <= 199711L
   // expected-warning@-2 {{default template arguments for a function template are a C++11 extension}}
 #endif
-  // expected-note@-4 {{candidate template ignored: disabled by 'enable_if' [with T = int]}}
+  // expected-note@+1 {{candidate template ignored: requirement 'a_trait<int>::value' was not satisfied [with T = int]}}
   void foo() {}
   void bar() { foo<int>(); } // expected-error {{no matching function for call to 'foo'}}
 
@@ -128,7 +128,7 @@ namespace PR15673 {
 #if __cplusplus <= 199711L
   // expected-warning@-2 {{alias declarations are a C++11 extension}}
 #endif
-  // expected-note@-4 {{candidate template ignored: disabled by 'enable_if' [with T = int]}}
+  // expected-note@+7 {{candidate template ignored: requirement 'some_trait<int>::value' was not satisfied [with T = int]}}
 
   template<typename T,
            typename Requires = unicorns<T> >
@@ -137,4 +137,30 @@ namespace PR15673 {
 #endif
   void wibble() {}
   void wobble() { wibble<int>(); } // expected-error {{no matching function for call to 'wibble'}}
+
+  template<typename T>
+  struct some_passing_trait : std::true_type {};
+
+#if __cplusplus <= 199711L
+  // expected-warning@+4 {{default template arguments for a function template are a C++11 extension}}
+  // expected-warning@+4 {{default template arguments for a function template are a C++11 extension}}
+#endif
+  template<typename T,
+           int n = 42,
+           typename std::enable_if<n == 43 || (some_passing_trait<T>::value && some_trait<T>::value), int>::type = 0>
+  void almost_rangesv3(); // expected-note{{candidate template ignored: requirement '42 == 43 || (some_passing_trait<int>::value && some_trait<int>::value)' was not satisfied}}
+  void test_almost_rangesv3() { almost_rangesv3<int>(); } // expected-error{{no matching function for call to 'almost_rangesv3'}}
+
+  #define CONCEPT_REQUIRES_(...)                                        \
+    int x = 42,                                                         \
+    typename std::enable_if<(x == 43) || (__VA_ARGS__)>::type = 0
+
+#if __cplusplus <= 199711L
+  // expected-warning@+4 {{default template arguments for a function template are a C++11 extension}}
+  // expected-warning@+3 {{default template arguments for a function template are a C++11 extension}}
+#endif
+  template<typename T,
+           CONCEPT_REQUIRES_(some_passing_trait<T>::value && some_trait<T>::value)>
+  void rangesv3(); // expected-note{{candidate template ignored: requirement 'some_trait<int>::value' was not satisfied [with T = int, x = 42]}}
+  void test_rangesv3() { rangesv3<int>(); } // expected-error{{no matching function for call to 'rangesv3'}}
 }
diff --git a/test/Unit/lit.cfg b/test/Unit/lit.cfg
index 39fad0099ceb..90eb2ac604a5 100644
--- a/test/Unit/lit.cfg
+++ b/test/Unit/lit.cfg
@@ -4,6 +4,7 @@
 
 import os
 import platform
+import subprocess
 
 import lit.formats
 import lit.util
@@ -65,8 +66,8 @@ if config.test_exec_root is None:
         lit_config.fatal('No site specific configuration available!')
 
     # Get the source and object roots.
-    llvm_src_root = lit.util.capture(['llvm-config', '--src-root']).strip()
-    llvm_obj_root = lit.util.capture(['llvm-config', '--obj-root']).strip()
+    llvm_src_root = subprocess.check_output(['llvm-config', '--src-root']).strip()
+    llvm_obj_root = subprocess.check_output(['llvm-config', '--obj-root']).strip()
     clang_src_root = os.path.join(llvm_src_root, "tools", "clang")
     clang_obj_root = os.path.join(llvm_obj_root, "tools", "clang")
 
diff --git a/test/lit.cfg b/test/lit.cfg
index e72eca6bd397..28026578299b 100644
--- a/test/lit.cfg
+++ b/test/lit.cfg
@@ -148,8 +148,8 @@ if config.test_exec_root is None:
         lit_config.fatal('No site specific configuration available!')
 
     # Get the source and object roots.
-    llvm_src_root = lit.util.capture(['llvm-config', '--src-root']).strip()
-    llvm_obj_root = lit.util.capture(['llvm-config', '--obj-root']).strip()
+    llvm_src_root = subprocess.check_output(['llvm-config', '--src-root']).strip()
+    llvm_obj_root = subprocess.check_output(['llvm-config', '--obj-root']).strip()
     clang_src_root = os.path.join(llvm_src_root, "tools", "clang")
     clang_obj_root = os.path.join(llvm_obj_root, "tools", "clang")