Index: lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
===================================================================
--- lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
+++ lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
@@ -17,6 +17,7 @@
 #include "llvm/ADT/BitVector.h"
 #include "llvm/ADT/Optional.h"
 #include "llvm/ADT/SmallSet.h"
+#include "llvm/ADT/Triple.h"
 #include "llvm/Analysis/AliasAnalysis.h"
 #include "llvm/Analysis/BranchProbabilityInfo.h"
 #include "llvm/Analysis/ConstantFolding.h"
@@ -6989,10 +6990,26 @@
 /// migrated to using LowerCall, this hook should be integrated into SDISel.
 std::pair<SDValue, SDValue>
 TargetLowering::LowerCallTo(TargetLowering::CallLoweringInfo &CLI) const {
+  Triple Trip(CLI.DAG.getTarget().getTargetTriple());
+  int Win64 = Trip.isArch64Bit() && Trip.isOSMSVCRT();
+
   // Handle the incoming return values from the call.
   CLI.Ins.clear();
   SmallVector<EVT, 4> RetTys;
   ComputeValueVTs(*this, CLI.RetTy, RetTys);
+  bool ret_m128 = false; EVT RetVT_m128;
+  if (Win64 && RetTys.size() == 1) {
+    EVT VT = RetTys[0];
+    if ((VT.isFloatingPoint() || VT.isInteger()) && VT.getSizeInBits() == 128) {
+      // change the return type to something llvm can represent, that
+      // we know will end up in the xmm0 register. we will patch up the
+      // return type later in this function
+      RetTys[0] = static_cast<EVT>(MVT::v2i64);
+      CLI.IsTailCall = false;
+      ret_m128 = true;
+      RetVT_m128 = VT;
+    }
+  }
   for (unsigned I = 0, E = RetTys.size(); I != E; ++I) {
     EVT VT = RetTys[I];
     MVT RegisterVT = getRegisterType(CLI.RetTy->getContext(), VT);
@@ -7018,7 +7035,22 @@
   ArgListTy &Args = CLI.Args;
   for (unsigned i = 0, e = Args.size(); i != e; ++i) {
     SmallVector<EVT, 4> ValueVTs;
-    ComputeValueVTs(*this, Args[i].Ty, ValueVTs);
+    TargetLowering::ArgListEntry Entry = Args[i];
+    if (Win64 && Entry.Node.getValueType().getSizeInBits() > 64) {
+      SDValue StackPtr = CLI.DAG.CreateStackTemporary(Entry.Node.getValueType(), 16);
+      int SPFI = cast<FrameIndexSDNode>(StackPtr.getNode())->getIndex();
+      MachinePointerInfo PtrInfo = MachinePointerInfo::getFixedStack(SPFI);
+      // Emit a store to the stack slot.
+      CLI.Chain = CLI.DAG.getStore(CLI.Chain, CLI.DL, Entry.Node, StackPtr, PtrInfo,
+                                   false, false, 16);
+      Entry.Node = StackPtr;
+      Entry.Ty = PointerType::get(Entry.Ty,0);
+      Entry.isSExt = false;
+      Entry.isZExt = false;
+      Args[i] = Entry;
+      CLI.IsTailCall = false;
+    }
+    ComputeValueVTs(*this, Entry.Ty, ValueVTs);
     for (unsigned Value = 0, NumValues = ValueVTs.size();
          Value != NumValues; ++Value) {
       EVT VT = ValueVTs[Value];
@@ -7160,6 +7192,8 @@
   SDValue Res = CLI.DAG.getNode(ISD::MERGE_VALUES, CLI.DL,
                                 CLI.DAG.getVTList(&RetTys[0], RetTys.size()),
                             &ReturnValues[0], ReturnValues.size());
+  if (ret_m128)
+    Res = CLI.DAG.getNode(ISD::BITCAST, CLI.DL, RetVT_m128, Res);
   return std::make_pair(Res, CLI.Chain);
 }
 
Index: test/CodeGen/X86/mod128.ll
===================================================================
--- /dev/null
+++ test/CodeGen/X86/mod128.ll
@@ -0,0 +1,15 @@
+; RUN: llc < %s -mtriple=x86_64-linux | FileCheck %s -check-prefix=X86-64
+; RUN: llc < %s -mtriple=x86_64-win32 | FileCheck %s -check-prefix=WIN64
+
+define i128 @mod128(i128 %x) nounwind {
+    ; X86-64: movl  $3, %edx
+    ; X86-64: xorl  %ecx, %ecx
+    ; X86-64: callq __modti3
+
+    ; WIN64-NOT: movl       $3, %r8d
+    ; WIN64-NOT: xorl       %r9d, %r9d
+    ; WIN64: callq   __modti3
+
+    %1 = srem i128 %x, 3
+    ret i128 %1
+}