Index: lib/Basic/Targets/X86.cpp
===================================================================
--- lib/Basic/Targets/X86.cpp
+++ lib/Basic/Targets/X86.cpp
@@ -1081,6 +1081,9 @@
   if (HasMWAITX)
     Builder.defineMacro("__MWAITX__");
 
+  if (HasMOVBE)
+    Builder.defineMacro("__MOVBE__");
+
   switch (XOPLevel) {
   case XOP:
     Builder.defineMacro("__XOP__");
Index: lib/Headers/immintrin.h
===================================================================
--- lib/Headers/immintrin.h
+++ lib/Headers/immintrin.h
@@ -306,6 +306,58 @@
 #endif
 #endif /* __FSGSBASE__ */
 
+#if !defined(_MSC_VER) || __has_feature(modules) || defined(__MOVBE__)
+static __inline__ short __attribute__((__always_inline__, __nodebug__, __target__("movbe")))
+_loadbe_i16(void const * __P) {
+  struct __loadu_i16 {
+    short __v;
+  } __attribute__((__packed__, __may_alias__));
+  return __builtin_bswap16(((struct __loadu_i16*)__P)->__v);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__, __target__("movbe")))
+_storebe_i16(void * __P, short __D) {
+  struct __storeu_i16 {
+    short __v;
+  } __attribute__((__packed__, __may_alias__));
+  ((struct __storeu_i16*)__P)->__v = __builtin_bswap16(__D);
+}
+
+static __inline__ int __attribute__((__always_inline__, __nodebug__, __target__("movbe")))
+_loadbe_i32(void const * __P) {
+  struct __loadu_i32 {
+    int __v;
+  } __attribute__((__packed__, __may_alias__));
+  return __builtin_bswap32(((struct __loadu_i32*)__P)->__v);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__, __target__("movbe")))
+_storebe_i32(void * __P, int __D) {
+  struct __storeu_i32 {
+    int __v;
+  } __attribute__((__packed__, __may_alias__));
+  ((struct __storeu_i32*)__P)->__v = __builtin_bswap32(__D);
+}
+
+#ifdef __x86_64__
+static __inline__ long long __attribute__((__always_inline__, __nodebug__, __target__("movbe")))
+_loadbe_i64(void const * __P) {
+  struct __loadu_i64 {
+    long long __v;
+  } __attribute__((__packed__, __may_alias__));
+  return __builtin_bswap64(((struct __loadu_i64*)__P)->__v);
+}
+
+static __inline__ void __attribute__((__always_inline__, __nodebug__, __target__("movbe")))
+_storebe_i64(void * __P, long long __D) {
+  struct __storeu_i64 {
+    long long __v;
+  } __attribute__((__packed__, __may_alias__));
+  ((struct __storeu_i64*)__P)->__v = __builtin_bswap64(__D);
+}
+#endif
+#endif /* __MOVEBE */
+
 #if !defined(_MSC_VER) || __has_feature(modules) || defined(__RTM__)
 #include <rtmintrin.h>
 #include <xtestintrin.h>
Index: test/CodeGen/movbe-builtins.c
===================================================================
--- /dev/null
+++ test/CodeGen/movbe-builtins.c
@@ -0,0 +1,49 @@
+// RUN: %clang_cc1 -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +movbe -emit-llvm -o - | FileCheck %s --check-prefixes=CHECK,CHECK-X64
+// RUN: %clang_cc1 -ffreestanding %s -triple=i686-apple-darwin -target-feature +movbe -emit-llvm -o - | FileCheck %s
+
+
+#include <immintrin.h>
+
+short test_loadbe_i16(const short *P) {
+  // CHECK-LABEL: @test_loadbe_i16
+  // CHECK: [[LOAD:%.*]] = load i16, i16* %{{.*}}, align 1
+  // CHECK: call i16 @llvm.bswap.i16(i16 [[LOAD]])
+  return _loadbe_i16(P);
+}
+
+void test_storebe_i16(short *P, short D) {
+  // CHECK-LABEL: @test_storebe_i16
+  // CHECK: [[DATA:%.*]] = call i16 @llvm.bswap.i16(i16 %{{.*}})
+  // CHECK: store i16 [[DATA]], i16* %{{.*}}, align 1
+  _storebe_i16(P, D);
+}
+
+int test_loadbe_i32(const int *P) {
+  // CHECK-LABEL: @test_loadbe_i32
+  // CHECK: [[LOAD:%.*]] = load i32, i32* %{{.*}}, align 1
+  // CHECK: call i32 @llvm.bswap.i32(i32 [[LOAD]])
+  return _loadbe_i32(P);
+}
+
+void test_storebe_i32(int *P, int D) {
+  // CHECK-LABEL: @test_storebe_i32
+  // CHECK: [[DATA:%.*]] = call i32 @llvm.bswap.i32(i32 %{{.*}})
+  // CHECK: store i32 [[DATA]], i32* %{{.*}}, align 1
+  _storebe_i32(P, D);
+}
+
+#ifdef __x86_64__
+long long test_loadbe_i64(const long long *P) {
+  // CHECK-X64-LABEL: @test_loadbe_i64
+  // CHECK-X64: [[LOAD:%.*]] = load i64, i64* %{{.*}}, align 1
+  // CHECK-X64: call i64 @llvm.bswap.i64(i64 [[LOAD]])
+  return _loadbe_i64(P);
+}
+
+void test_storebe_i64(long long *P, long long D) {
+  // CHECK-X64-LABEL: @test_storebe_i64
+  // CHECK-X64: [[DATA:%.*]] = call i64 @llvm.bswap.i64(i64 %{{.*}})
+  // CHECK-X64: store i64 [[DATA]], i64* %{{.*}}, align 1
+  _storebe_i64(P, D);
+}
+#endif