Index: lib/Transforms/InstCombine/InstCombineCalls.cpp
===================================================================
--- lib/Transforms/InstCombine/InstCombineCalls.cpp
+++ lib/Transforms/InstCombine/InstCombineCalls.cpp
@@ -576,6 +576,24 @@
     break;
   }
 
+  case Intrinsic::x86_sse4a_insertqi: {
+    // insertqi x, y, 64, 0 can just copy y's lower bits and leave the top
+    // ones undef
+    if (ConstantInt *Width = dyn_cast<ConstantInt>(II->getArgOperand(2)))
+      if (ConstantInt *Start = dyn_cast<ConstantInt>(II->getArgOperand(3)))
+        if (Width->equalsInt(64) && Start->isZero()) {
+          Value *Vec = II->getArgOperand(1);
+          Value *Undef = UndefValue::get(Vec->getType());
+          const uint32_t Mask[] = { 0, 2 };
+          return ReplaceInstUsesWith(
+              CI,
+              Builder->CreateShuffleVector(
+                  Vec, Undef, ConstantDataVector::get(
+                                  II->getContext(), ArrayRef<uint32_t>(Mask))));
+        }
+    break;
+  }
+
   case Intrinsic::ppc_altivec_vperm:
     // Turn vperm(V1,V2,mask) -> shuffle(V1,V2,mask) if mask is a constant.
     if (Constant *Mask = dyn_cast<Constant>(II->getArgOperand(2))) {
Index: test/Transforms/InstCombine/2014-04-11-SSE4a-insertqi-64-bits.ll
===================================================================
--- /dev/null
+++ test/Transforms/InstCombine/2014-04-11-SSE4a-insertqi-64-bits.ll
@@ -0,0 +1,11 @@
+; RUN: opt < %s -instcombine -S | FileCheck %s
+; CHECK-NOT: insertqi
+
+define <2 x i64> @test(<2 x i64> %v, <2 x i64> %i) #0 {
+  %1 = tail call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %v, <2 x i64> %i, i8 64, i8 0)
+  ret <2 x i64> %1
+}
+
+; CHECK: declare <2 x i64> @llvm.x86.sse4a.insertqi
+declare <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64>, <2 x i64>, i8, i8) nounwind
+