From 55e03c1992776334e04330e1afc1a652e92f88dc Mon Sep 17 00:00:00 2001
From: Benjamin Kramer <benny.kra@googlemail.com>
Date: Sun, 27 Apr 2014 11:41:06 +0000
Subject: [PATCH] SelectionDAG: Aggressively fold shuffles of constant splats.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207352 91177308-0d34-0410-b5e6-96231b3b80d8
---
 lib/CodeGen/SelectionDAG/SelectionDAG.cpp | 5 +++++
 test/CodeGen/X86/vector-idiv.ll           | 8 +++++---
 2 files changed, 10 insertions(+), 3 deletions(-)

diff --git a/lib/CodeGen/SelectionDAG/SelectionDAG.cpp b/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
index 3f68d49da68..46d0fcac410 100644
--- a/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
+++ b/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
@@ -1441,6 +1441,11 @@ SDValue SelectionDAG::getVectorShuffle(EVT VT, SDLoc dl, SDValue N1,
   if (Identity && NElts)
     return N1;
 
+  // Shuffling a constant splat doesn't change the result.
+  if (N2Undef && N1.getOpcode() == ISD::BUILD_VECTOR)
+    if (cast<BuildVectorSDNode>(N1)->getConstantSplatValue())
+      return N1;
+
   FoldingSetNodeID ID;
   SDValue Ops[2] = { N1, N2 };
   AddNodeIDNode(ID, ISD::VECTOR_SHUFFLE, getVTList(VT), Ops, 2);
diff --git a/test/CodeGen/X86/vector-idiv.ll b/test/CodeGen/X86/vector-idiv.ll
index 3b300f74061..981c317157f 100644
--- a/test/CodeGen/X86/vector-idiv.ll
+++ b/test/CodeGen/X86/vector-idiv.ll
@@ -115,6 +115,7 @@ define <4 x i32> @test8(<4 x i32> %a) {
 ; SSE-LABEL: test8:
 ; SSE: pmuldq
 ; SSE: pshufd	$57
+; SSE-NOT: pshufd	$57
 ; SSE: pmuldq
 ; SSE: shufps	$-35
 ; SSE: pshufd	$-40
@@ -126,6 +127,7 @@ define <4 x i32> @test8(<4 x i32> %a) {
 ; AVX-LABEL: test8:
 ; AVX: vpmuldq
 ; AVX: vpshufd	$57
+; AVX-NOT: vpshufd	$57
 ; AVX: vpmuldq
 ; AVX: vshufps	$-35
 ; AVX: vpshufd	$-40
@@ -140,7 +142,7 @@ define <8 x i32> @test9(<8 x i32> %a) {
   ret <8 x i32> %div
 
 ; AVX-LABEL: test9:
-; AVX: vpermd
+; AVX: vpbroadcastd
 ; AVX: vpmuldq
 ; AVX: vshufps	$-35
 ; AVX: vpmuldq
@@ -157,7 +159,7 @@ define <8 x i32> @test10(<8 x i32> %a) {
   ret <8 x i32> %rem
 
 ; AVX-LABEL: test10:
-; AVX: vpermd
+; AVX: vpbroadcastd
 ; AVX: vpmuludq
 ; AVX: vshufps	$-35
 ; AVX: vpmuludq
@@ -174,7 +176,7 @@ define <8 x i32> @test11(<8 x i32> %a) {
   ret <8 x i32> %rem
 
 ; AVX-LABEL: test11:
-; AVX: vpermd
+; AVX: vpbroadcastd
 ; AVX: vpmuldq
 ; AVX: vshufps	$-35
 ; AVX: vpmuldq