Index: llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
===================================================================
--- llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
+++ llvm/trunk/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
@@ -304,6 +304,7 @@
     SDValue visitCONCAT_VECTORS(SDNode *N);
     SDValue visitEXTRACT_SUBVECTOR(SDNode *N);
     SDValue visitVECTOR_SHUFFLE(SDNode *N);
+    SDValue visitSCALAR_TO_VECTOR(SDNode *N);
     SDValue visitINSERT_SUBVECTOR(SDNode *N);
     SDValue visitMLOAD(SDNode *N);
     SDValue visitMSTORE(SDNode *N);
@@ -1371,6 +1372,7 @@
   case ISD::CONCAT_VECTORS:     return visitCONCAT_VECTORS(N);
   case ISD::EXTRACT_SUBVECTOR:  return visitEXTRACT_SUBVECTOR(N);
   case ISD::VECTOR_SHUFFLE:     return visitVECTOR_SHUFFLE(N);
+  case ISD::SCALAR_TO_VECTOR:   return visitSCALAR_TO_VECTOR(N);
   case ISD::INSERT_SUBVECTOR:   return visitINSERT_SUBVECTOR(N);
   case ISD::MLOAD:              return visitMLOAD(N);
   case ISD::MSTORE:             return visitMSTORE(N);
@@ -12188,6 +12190,34 @@
   return SDValue();
 }
 
+SDValue DAGCombiner::visitSCALAR_TO_VECTOR(SDNode *N) {
+  SDValue InVal = N->getOperand(0);
+  EVT VT = N->getValueType(0);
+
+  // Replace a SCALAR_TO_VECTOR(EXTRACT_VECTOR_ELT(V,C0)) pattern
+  // with a VECTOR_SHUFFLE.
+  if (InVal.getOpcode() == ISD::EXTRACT_VECTOR_ELT) {
+    SDValue InVec = InVal->getOperand(0);
+    SDValue EltNo = InVal->getOperand(1);
+
+    // FIXME: We could support implicit truncation if the shuffle can be
+    // scaled to a smaller vector scalar type.
+    ConstantSDNode *C0 = dyn_cast<ConstantSDNode>(EltNo);
+    if (C0 && VT == InVec.getValueType() &&
+        VT.getScalarType() == InVal.getValueType()) {
+      SmallVector<int, 8> NewMask(VT.getVectorNumElements(), -1);
+      int Elt = C0->getZExtValue();
+      NewMask[0] = Elt;
+
+      if (TLI.isShuffleMaskLegal(NewMask, VT))
+        return DAG.getVectorShuffle(VT, SDLoc(N), InVec, DAG.getUNDEF(VT),
+                                    NewMask);
+    }
+  }
+
+  return SDValue();
+}
+
 SDValue DAGCombiner::visitINSERT_SUBVECTOR(SDNode *N) {
   SDValue N0 = N->getOperand(0);
   SDValue N2 = N->getOperand(2);
Index: llvm/trunk/test/CodeGen/X86/mmx-arg-passing-x86-64.ll
===================================================================
--- llvm/trunk/test/CodeGen/X86/mmx-arg-passing-x86-64.ll
+++ llvm/trunk/test/CodeGen/X86/mmx-arg-passing-x86-64.ll
@@ -25,11 +25,9 @@
 ; X86-64-NEXT:    movq %mm0, -{{[0-9]+}}(%rsp)
 ; X86-64-NEXT:    movdq2q %xmm0, %mm0
 ; X86-64-NEXT:    movq %mm0, -{{[0-9]+}}(%rsp)
-; X86-64-NEXT:    movq {{.*#+}} xmm0 = mem[0],zero
 ; X86-64-NEXT:    movq {{.*#+}} xmm1 = mem[0],zero
-; X86-64-NEXT:    paddb %xmm0, %xmm1
-; X86-64-NEXT:    movd %xmm1, %rax
-; X86-64-NEXT:    movd %rax, %xmm0
+; X86-64-NEXT:    movq {{.*#+}} xmm0 = mem[0],zero
+; X86-64-NEXT:    paddb %xmm1, %xmm0
 ; X86-64-NEXT:    movb $1, %al
 ; X86-64-NEXT:    jmp _pass_v8qi ## TAILCALL
   %v1a = bitcast x86_mmx %v1 to <8 x i8>
Index: llvm/trunk/test/CodeGen/X86/pr14161.ll
===================================================================
--- llvm/trunk/test/CodeGen/X86/pr14161.ll
+++ llvm/trunk/test/CodeGen/X86/pr14161.ll
@@ -26,9 +26,8 @@
 ; CHECK:       # BB#0: # %entry
 ; CHECK-NEXT:    movdqa (%rdi), %xmm0
 ; CHECK-NEXT:    pminud {{.*}}(%rip), %xmm0
-; CHECK-NEXT:    pextrd $1, %xmm0, %eax
-; CHECK-NEXT:    movd %eax, %xmm0
-; CHECK-NEXT:    pmovzxwq %xmm0, %xmm0
+; CHECK-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,2,3]
+; CHECK-NEXT:    pmovzxwq {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero
 ; CHECK-NEXT:    retq
 entry:
   %2 = load <4 x i32>, <4 x i32>* %0, align 16