[DAGCombiner] Add more rules to fold shuffles.

This patch adds two new rules to the DAGCombiner:
 1.  shuffle (shuffle A, Undef, M0), B, M1 -> shuffle A, B, M2
 2.  shuffle (shuffle A, Undef, M0), A, M1 -> shuffle A, Undef, M2

We only do this if the combined shuffle is legal for the target.

Example:
;;
define <4 x float> @test(<4 x float> %a, <4 x float> %b) {
  %1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32><i32 6, i32 0, i32 1, i32 7>
  %2 = shufflevector <4 x float> %1, <4 x float> %b, <4 x i32><i32 1, i32 2, i32 4, i32 5>
  ret <4 x i32> %2
}
;;

(using llc -mcpu=corei7 -march=x86-64)
Before, the x86 backend generated:
  pshufd $120, %xmm0, %xmm0
  shufps $-108, %xmm0, %xmm1
  movaps %xmm1, %xmm0

Now the x86 backend generates:
  movsd %xmm1, %xmm0

llvm-svn: 213069
diff --git a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
index 9d91b3e..28031bb 100644
--- a/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
@@ -10780,11 +10780,13 @@
   }
 
   // Try to fold according to rules:
-  //   shuffle(shuffle A, B, M0), B, M1) -> shuffle(A, B, M2)
-  //   shuffle(shuffle A, B, M0), A, M1) -> shuffle(A, B, M2)
+  //   shuffle(shuffle(A, B, M0), B, M1) -> shuffle(A, B, M2)
+  //   shuffle(shuffle(A, B, M0), A, M1) -> shuffle(A, B, M2)
+  //   shuffle(shuffle(A, Undef, M0), B, M1) -> shuffle(A, B, M2)
+  //   shuffle(shuffle(A, Undef, M0), A, M1) -> shuffle(A, Undef, M2)
   // Don't try to fold shuffles with illegal type.
   if (N0.getOpcode() == ISD::VECTOR_SHUFFLE && Level < AfterLegalizeDAG &&
-      TLI.isTypeLegal(VT)) {
+      N1.getOpcode() != ISD::UNDEF && TLI.isTypeLegal(VT)) {
     ShuffleVectorSDNode *OtherSV = cast<ShuffleVectorSDNode>(N0);
 
     // The incoming shuffle must be of the same type as the result of the
@@ -10795,7 +10797,8 @@
     SDValue SV0 = OtherSV->getOperand(0);
     SDValue SV1 = OtherSV->getOperand(1);
     bool HasSameOp0 = N1 == SV0;
-    if (!HasSameOp0 && N1 != SV1)
+    bool IsSV1Undef = SV1.getOpcode() == ISD::UNDEF;
+    if (!HasSameOp0 && !IsSV1Undef && N1 != SV1)
       // Early exit.
       return SDValue();
 
@@ -10810,17 +10813,24 @@
         continue;
       }
 
-      if (Idx < (int)NumElts)
+      if (Idx < (int)NumElts) {
         Idx = OtherSV->getMaskElt(Idx);
-      else
+        if (IsSV1Undef && Idx >= (int) NumElts)
+          Idx = -1;  // Propagate Undef.
+      } else
         Idx = HasSameOp0 ? Idx - NumElts : Idx;
 
       Mask.push_back(Idx);
     }
 
     // Avoid introducing shuffles with illegal mask.
-    if (TLI.isShuffleMaskLegal(Mask, VT))
+    if (TLI.isShuffleMaskLegal(Mask, VT)) {
+      if (IsSV1Undef)
+        //   shuffle(shuffle(A, Undef, M0), B, M1) -> shuffle(A, B, M2)
+        //   shuffle(shuffle(A, Undef, M0), A, M1) -> shuffle(A, Undef, M2)
+        return DAG.getVectorShuffle(VT, SDLoc(N), SV0, N1, &Mask[0]);
       return DAG.getVectorShuffle(VT, SDLoc(N), SV0, SV1, &Mask[0]);
+    }
   }
 
   return SDValue();