diff --git a/clang/lib/CodeGen/ItaniumCXXABI.cpp b/clang/lib/CodeGen/ItaniumCXXABI.cpp
--- a/clang/lib/CodeGen/ItaniumCXXABI.cpp
+++ b/clang/lib/CodeGen/ItaniumCXXABI.cpp
@@ -2472,7 +2472,10 @@
     CGF.EmitNounwindRuntimeCall(getGuardReleaseFn(CGM, guardPtrTy),
                                 guardAddr.getPointer());
   } else {
-    Builder.CreateStore(llvm::ConstantInt::get(guardTy, 1), guardAddr);
+    // Store 1 into the first byte of the guard variable after initialization is
+    // complete.
+    Builder.CreateStore(llvm::ConstantInt::get(CGM.Int8Ty, 1),
+                        Builder.CreateElementBitCast(guardAddr, CGM.Int8Ty));
   }
 
   CGF.EmitBlock(EndBlock);
diff --git a/clang/test/CodeGenCXX/aix-static-init-temp-spec-and-inline-var.cpp b/clang/test/CodeGenCXX/aix-static-init-temp-spec-and-inline-var.cpp
--- a/clang/test/CodeGenCXX/aix-static-init-temp-spec-and-inline-var.cpp
+++ b/clang/test/CodeGenCXX/aix-static-init-temp-spec-and-inline-var.cpp
@@ -44,6 +44,9 @@
 A<int> A<int>::instance = bar();
 } // namespace test2
 
+// CHECK: @_ZGVN5test12t2E = linkonce_odr global i64 0, align 8
+// CHECK: @_ZGVN5test21AIvE8instanceE = weak_odr global i64 0, align 8
+// CHECK: @_ZGVN5test12t1IiEE = linkonce_odr global i64 0, align 8
 // CHECK: @llvm.global_ctors = appending global [4 x { i32, void ()*, i8* }] [{ i32, void ()*, i8* } { i32 65535, void ()* @__cxx_global_var_init.1, i8* null }, { i32, void ()*, i8* } { i32 65535, void ()* @__cxx_global_var_init.2, i8* null }, { i32, void ()*, i8* } { i32 65535, void ()* @__cxx_global_var_init.4, i8* null }, { i32, void ()*, i8* } { i32 65535, void ()* @_GLOBAL__sub_I__, i8* null }]
 // CHECK: @llvm.global_dtors = appending global [4 x { i32, void ()*, i8* }] [{ i32, void ()*, i8* } { i32 65535, void ()* @__finalize__ZN5test12t2E, i8* null }, { i32, void ()*, i8* } { i32 65535, void ()* @__finalize__ZN5test21AIvE8instanceE, i8* null }, { i32, void ()*, i8* } { i32 65535, void ()* @__finalize__ZN5test12t1IiEE, i8* null }, { i32, void ()*, i8* } { i32 65535, void ()* @_GLOBAL__D_a, i8* null }]
 
@@ -77,7 +80,7 @@
 
 // CHECK: define internal void @__cxx_global_var_init.1() [[ATTR:#[0-9]+]] {
 // CHECK: entry:
-// CHECK:   %0 = load atomic i8, i8* bitcast (i64* @_ZGVN5test12t2E to i8*) acquire
+// CHECK:   %0 = load atomic i8, i8* bitcast (i64* @_ZGVN5test12t2E to i8*) acquire, align 8
 // CHECK:   %guard.uninitialized = icmp eq i8 %0, 0
 // CHECK:   br i1 %guard.uninitialized, label %init.check, label %init.end
 
@@ -119,14 +122,14 @@
 
 // CHECK: define internal void @__cxx_global_var_init.2() [[ATTR:#[0-9]+]] {
 // CHECK: entry:
-// CHECK:   %0 = load i8, i8* bitcast (i64* @_ZGVN5test21AIvE8instanceE to i8*)
+// CHECK:   %0 = load i8, i8* bitcast (i64* @_ZGVN5test21AIvE8instanceE to i8*), align 8
 // CHECK:   %guard.uninitialized = icmp eq i8 %0, 0
 // CHECK:   br i1 %guard.uninitialized, label %init.check, label %init.end
 
 // CHECK: init.check:
 // CHECK:   call void @_ZN5test21AIvEC1Ev(%"struct.test2::A"* {{[^,]*}} @_ZN5test21AIvE8instanceE)
 // CHECK:   %1 = call i32 @atexit(void ()* @__dtor__ZN5test21AIvE8instanceE)
-// CHECK:   store i64 1, i64* @_ZGVN5test21AIvE8instanceE
+// CHECK:   store i8 1, i8* bitcast (i64* @_ZGVN5test21AIvE8instanceE to i8*), align 8
 // CHECK:   br label %init.end
 
 // CHECK: init.end:
@@ -182,7 +185,7 @@
 
 // CHECK: define internal void @__cxx_global_var_init.4() [[ATTR:#[0-9]+]] {
 // CHECK: entry:
-// CHECK:   %0 = load i8, i8* bitcast (i64* @_ZGVN5test12t1IiEE to i8*)
+// CHECK:   %0 = load i8, i8* bitcast (i64* @_ZGVN5test12t1IiEE to i8*), align 8
 // CHECK:   %guard.uninitialized = icmp eq i8 %0, 0
 // CHECK:   br i1 %guard.uninitialized, label %init.check, label %init.end
 
@@ -190,7 +193,7 @@
 // CHECK32: call void @_ZN5test15Test1C1Ei(%"struct.test1::Test1"* {{[^,]*}} @_ZN5test12t1IiEE, i32 2)
 // CHECK64: call void @_ZN5test15Test1C1Ei(%"struct.test1::Test1"* {{[^,]*}} @_ZN5test12t1IiEE, i32 signext 2)
 // CHECK:   %1 = call i32 @atexit(void ()* @__dtor__ZN5test12t1IiEE)
-// CHECK:   store i64 1, i64* @_ZGVN5test12t1IiEE
+// CHECK:   store i8 1, i8* bitcast (i64* @_ZGVN5test12t1IiEE to i8*), align 8
 // CHECK:   br label %init.end
 
 // CHECK: init.end:
diff --git a/clang/test/CodeGenCXX/aix-static-init.cpp b/clang/test/CodeGenCXX/aix-static-init.cpp
--- a/clang/test/CodeGenCXX/aix-static-init.cpp
+++ b/clang/test/CodeGenCXX/aix-static-init.cpp
@@ -38,6 +38,7 @@
   }
 } // namespace test4
 
+// CHECK: @_ZGVZN5test41fEvE11staticLocal = internal global i64 0, align 8
 // CHECK: @llvm.global_ctors = appending global [1 x { i32, void ()*, i8* }] [{ i32, void ()*, i8* } { i32 65535, void ()* @_GLOBAL__sub_I__, i8* null }]
 // CHECK: @llvm.global_dtors = appending global [1 x { i32, void ()*, i8* }] [{ i32, void ()*, i8* } { i32 65535, void ()* @_GLOBAL__D_a, i8* null }]
 
@@ -135,7 +136,7 @@
 
 // CHECK: define void @_ZN5test41fEv() [[ATTR:#[0-9]+]] {
 // CHECK: entry:
-// CHECK:   %0 = load atomic i8, i8* bitcast (i64* @_ZGVZN5test41fEvE11staticLocal to i8*) acquire
+// CHECK:   %0 = load atomic i8, i8* bitcast (i64* @_ZGVZN5test41fEvE11staticLocal to i8*) acquire, align 8
 // CHECK:   %guard.uninitialized = icmp eq i8 %0, 0
 // CHECK:   br i1 %guard.uninitialized, label %init.check, label %init.end
 
diff --git a/clang/test/CodeGenCXX/cxx11-thread-local.cpp b/clang/test/CodeGenCXX/cxx11-thread-local.cpp
--- a/clang/test/CodeGenCXX/cxx11-thread-local.cpp
+++ b/clang/test/CodeGenCXX/cxx11-thread-local.cpp
@@ -184,7 +184,7 @@
 // need init:
 // CHECK: call i32 @_Z1gv()
 // CHECK: store i32 %{{.*}}, i32* @_ZN1VIfE1mE, align 4
-// CHECK: store i64 1, i64* @_ZGVN1VIfE1mE
+// CHECK: store i8 1, i8* bitcast (i64* @_ZGVN1VIfE1mE to i8*)
 // CHECK: br label
 
 // LINUX: define internal void @[[XF_M_INIT]]()
@@ -197,7 +197,7 @@
 // need init:
 // LINUX: call {{.*}}__cxa_thread_atexit
 // DARWIN: call {{.*}}_tlv_atexit
-// CHECK: store i64 1, i64* @_ZGVN1XIfE1mE
+// CHECK: store i8 1, i8* bitcast (i64* @_ZGVN1XIfE1mE to i8*)
 // CHECK: br label
 
 // LINUX: declare i32 @__cxa_thread_atexit(void (i8*)*, i8*, i8*)
@@ -285,7 +285,7 @@
 // need init:
 // CHECK: call i32 @_Z1gv()
 // CHECK: store i32 %{{.*}}, i32* @_ZN1VIiE1mE, align 4
-// CHECK: store i64 1, i64* @_ZGVN1VIiE1mE
+// CHECK: store i8 1, i8* bitcast (i64* @_ZGVN1VIiE1mE to i8*)
 // CHECK: br label
 
 // LINUX: define internal void @[[X_M_INIT]]()
@@ -298,7 +298,7 @@
 // need init:
 // LINUX: call {{.*}}__cxa_thread_atexit
 // DARWIN: call {{.*}}_tlv_atexit
-// CHECK: store i64 1, i64* @_ZGVN1XIiE1mE
+// CHECK: store i8 1, i8* bitcast (i64* @_ZGVN1XIiE1mE to i8*)
 // CHECK: br label
 
 // CHECK: define {{.*}}@[[GLOBAL_INIT:.*]]()
diff --git a/clang/test/CodeGenCXX/global-init.cpp b/clang/test/CodeGenCXX/global-init.cpp
--- a/clang/test/CodeGenCXX/global-init.cpp
+++ b/clang/test/CodeGenCXX/global-init.cpp
@@ -78,10 +78,10 @@
   extern int foo();
 
   // This needs an initialization function and guard variables.
-  // CHECK: load i8, i8* bitcast (i64* @_ZGVN5test41xE
+  // CHECK: load i8, i8* bitcast (i64* @_ZGVN5test41xE to i8*)
   // CHECK: [[CALL:%.*]] = call i32 @_ZN5test43fooEv
   // CHECK-NEXT: store i32 [[CALL]], i32* @_ZN5test41xE
-  // CHECK-NEXT: store i64 1, i64* @_ZGVN5test41xE
+  // CHECK-NEXT: store i8 1, i8* bitcast (i64* @_ZGVN5test41xE to i8*)
   __attribute__((weak)) int x = foo();
 }
 
diff --git a/clang/test/CodeGenCXX/static-data-member.cpp b/clang/test/CodeGenCXX/static-data-member.cpp
--- a/clang/test/CodeGenCXX/static-data-member.cpp
+++ b/clang/test/CodeGenCXX/static-data-member.cpp
@@ -72,7 +72,7 @@
   // CHECK-NEXT: br i1 [[UNINITIALIZED]]
   // CHECK:      [[TMP:%.*]] = call i32 @_ZN5test33fooEv()
   // CHECK-NEXT: store i32 [[TMP]], i32* @_ZN5test31AIiE1xE, align 4
-  // CHECK-NEXT: store i64 1, i64* @_ZGVN5test31AIiE1xE
+  // CHECK-NEXT: store i8 1, i8* bitcast (i64* @_ZGVN5test31AIiE1xE to i8*)
   // CHECK-NEXT: br label
   // CHECK:      ret void
 }