Index: lib/IR/IRBuilder.cpp
===================================================================
--- lib/IR/IRBuilder.cpp
+++ lib/IR/IRBuilder.cpp
@@ -107,6 +107,7 @@
   Function *TheFn = Intrinsic::getDeclaration(M, Intrinsic::memset, Tys);
 
   CallInst *CI = createCallHelper(TheFn, Ops, this);
+  CI->setTailCall(true);
 
   if (Align > 0)
     cast<MemSetInst>(CI)->setDestAlignment(Align);
@@ -138,6 +139,7 @@
       M, Intrinsic::memset_element_unordered_atomic, Tys);
 
   CallInst *CI = createCallHelper(TheFn, Ops, this);
+  CI->setTailCall(true);
 
   cast<AtomicMemSetInst>(CI)->setDestAlignment(Align);
 
@@ -169,6 +171,7 @@
   Function *TheFn = Intrinsic::getDeclaration(M, Intrinsic::memcpy, Tys);
 
   CallInst *CI = createCallHelper(TheFn, Ops, this);
+  CI->setTailCall(true);
 
   auto* MCI = cast<MemCpyInst>(CI);
   if (DstAlign > 0)
@@ -211,6 +214,7 @@
       M, Intrinsic::memcpy_element_unordered_atomic, Tys);
 
   CallInst *CI = createCallHelper(TheFn, Ops, this);
+  CI->setTailCall(true);
 
   // Set the alignment of the pointer args.
   auto *AMCI = cast<AtomicMemCpyInst>(CI);
@@ -249,6 +253,7 @@
   Function *TheFn = Intrinsic::getDeclaration(M, Intrinsic::memmove, Tys);
 
   CallInst *CI = createCallHelper(TheFn, Ops, this);
+  CI->setTailCall(true);
 
   auto *MMI = cast<MemMoveInst>(CI);
   if (DstAlign > 0)
@@ -287,6 +292,7 @@
       M, Intrinsic::memmove_element_unordered_atomic, Tys);
 
   CallInst *CI = createCallHelper(TheFn, Ops, this);
+  CI->setTailCall(true);
 
   // Set the alignment of the pointer args.
   CI->addParamAttr(0, Attribute::getWithAlignment(CI->getContext(), DstAlign));
Index: test/CodeGen/X86/memset-nonzero.ll
===================================================================
--- test/CodeGen/X86/memset-nonzero.ll
+++ test/CodeGen/X86/memset-nonzero.ll
@@ -142,14 +142,9 @@
 define void @memset_256_nonzero_bytes(i8* %x) {
 ; SSE-LABEL: memset_256_nonzero_bytes:
 ; SSE:       # %bb.0:
-; SSE-NEXT:    pushq %rax
-; SSE-NEXT:    .cfi_def_cfa_offset 16
 ; SSE-NEXT:    movl $256, %edx # imm = 0x100
 ; SSE-NEXT:    movl $42, %esi
-; SSE-NEXT:    callq memset
-; SSE-NEXT:    popq %rax
-; SSE-NEXT:    .cfi_def_cfa_offset 8
-; SSE-NEXT:    retq
+; SSE-NEXT:    jmp memset
 ;
 ; SSE2FAST-LABEL: memset_256_nonzero_bytes:
 ; SSE2FAST:       # %bb.0:
Index: test/Transforms/MemCpyOpt/enable-tail-call.ll
===================================================================
--- /dev/null
+++ test/Transforms/MemCpyOpt/enable-tail-call.ll
@@ -0,0 +1,40 @@
+; RUN: opt -memcpyopt -S < %s | FileCheck %s
+
+target datalayout = "e-i64:64-f80:128-n8:16:32:64"
+target triple = "x86_64-unknown-linux-gnu"
+
+%S = type { i8*, i8, i32 }
+
+define void @copy(%S* %src, %S* %dst) {
+; CHECK-LABEL: copy
+; CHECK-NOT: load
+; CHECK: tail call void @llvm.memmove.p0i8.p0i8.i64
+; CHECK-NEXT: ret void
+  %1 = load %S, %S* %src
+  store %S %1, %S* %dst
+  ret void
+}
+
+define void @noaliassrc(%S* noalias %src, %S* %dst) {
+; CHECK-LABEL: noaliassrc
+; CHECK-NOT: load
+; CHECK: tail call void @llvm.memcpy.p0i8.p0i8.i64
+; CHECK-NEXT: ret void
+  %1 = load %S, %S* %src
+  store %S %1, %S* %dst
+  ret void
+}
+
+
+define void @destroysrc(%S* %src, %S* %dst) {
+; CHECK-LABEL: destroysrc
+; CHECK: load %S, %S* %src
+; CHECK: tail call void @llvm.memset.p0i8.i64
+; CHECK-NEXT: store %S %1, %S* %dst
+; CHECK-NEXT: ret void
+  %1 = load %S, %S* %src
+  store %S zeroinitializer, %S* %src
+  store %S %1, %S* %dst
+  ret void
+}
+