From: Nadav Rotem <nadav.rotem@intel.com>
Date: Tue, 3 Jan 2012 22:19:42 +0000 (+0000)
Subject: Revert 147426 because it caused pr11696.
X-Git-Url: http://demsky.eecs.uci.edu/git/?a=commitdiff_plain;h=c2d064f0283504d7fe91369021bfd44ce5c8eb2a;p=oota-llvm.git

Revert 147426 because it caused pr11696.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147485 91177308-0d34-0410-b5e6-96231b3b80d8
---

diff --git a/lib/Target/X86/X86ISelLowering.cpp b/lib/Target/X86/X86ISelLowering.cpp
index 47b80d036e5..74ae8a85bda 100644
--- a/lib/Target/X86/X86ISelLowering.cpp
+++ b/lib/Target/X86/X86ISelLowering.cpp
@@ -13139,24 +13139,6 @@ static SDValue PerformSELECTCombine(SDNode *N, SelectionDAG &DAG,
       }
   }
 
-  // The VSELECT instruction is lowered to SSE blend instructions. In many cases
-  // the mask is sign-extended to fill the entire lane. However, we only care
-  // for the highest bit. Convert sign_extend to srl because it is cheaper.
-  // (vselect(sign_extend(x))) ->  vselect(srl(x))
-  if (N->getOpcode() == ISD::VSELECT &&
-      Cond.getOpcode() == ISD::SIGN_EXTEND_INREG && Cond.hasOneUse()) {
-    EVT CondVT = Cond.getValueType();
-    EVT SExtTy = cast<VTSDNode>(Cond.getOperand(1))->getVT();
-    unsigned BitsDiff = CondVT.getScalarType().getSizeInBits() -
-                        SExtTy.getScalarType().getSizeInBits();
-
-    EVT ShiftType = EVT::getVectorVT(*DAG.getContext(),
-                MVT::i32, CondVT.getVectorNumElements());
-    SDValue SHL = DAG.getNode(ISD::SHL, DL, CondVT, Cond.getOperand(0),
-                DAG.getConstant(BitsDiff, ShiftType));
-    return DAG.getNode(ISD::VSELECT, DL, VT, SHL, LHS, RHS);
-  }
-
   return SDValue();
 }
 
diff --git a/test/CodeGen/X86/2011-12-28-vselecti8.ll b/test/CodeGen/X86/2011-12-28-vselecti8.ll
index fc1b83b2224..dbc122ac6e4 100644
--- a/test/CodeGen/X86/2011-12-28-vselecti8.ll
+++ b/test/CodeGen/X86/2011-12-28-vselecti8.ll
@@ -5,10 +5,8 @@ target triple = "x86_64-apple-darwin11.2.0"
 
 ; CHECK: @foo8
 ; CHECK: psll
-; CHECK-NOT: sra
-; CHECK: pandn
-; CHECK: pand
-; CHECK: or
+; CHECK: psraw
+; CHECK: pblendvb
 ; CHECK: ret
 define void @foo8(float* nocapture %RET) nounwind {
 allocas:
diff --git a/test/CodeGen/X86/sext-blend.ll b/test/CodeGen/X86/sext-blend.ll
deleted file mode 100644
index b1f9573f307..00000000000
--- a/test/CodeGen/X86/sext-blend.ll
+++ /dev/null
@@ -1,15 +0,0 @@
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=corei7 -promote-elements -mattr=+sse41 | FileCheck %s
-
-; CHECK: foo
-define <4 x double> @foo(<4 x double> %x, <4 x double> %y) {
-  ; CHECK: cmpnlepd
-  ; CHECK: psllq
-  ; CHECK-NEXT: blendvpd
-  ; CHECK: psllq
-  ; CHECK-NEXT: blendvpd
-  ; CHECK: ret
-  %min_is_x = fcmp ult <4 x double> %x, %y
-  %min = select <4 x i1> %min_is_x, <4 x double> %x, <4 x double> %y
-  ret <4 x double> %min
-}
-
diff --git a/test/CodeGen/X86/sse2-blend.ll b/test/CodeGen/X86/sse2-blend.ll
index c6602d3a3b4..2f4317bf294 100644
--- a/test/CodeGen/X86/sse2-blend.ll
+++ b/test/CodeGen/X86/sse2-blend.ll
@@ -28,10 +28,10 @@ define void@vsel_i32(<4 x i32>* %v1, <4 x i32>* %v2) {
 
 ; Without forcing instructions, fall back to the preferred PS domain.
 ; CHECK: vsel_i64
-; CHECK: pxor
-; CHECK: and
-; CHECK: andn
-; CHECK: or
+; CHECK: xorps
+; CHECK: andps
+; CHECK: andnps
+; CHECK: orps
 ; CHECK: ret
 
 define void@vsel_i64(<4 x i64>* %v1, <4 x i64>* %v2) {
@@ -44,10 +44,10 @@ define void@vsel_i64(<4 x i64>* %v1, <4 x i64>* %v2) {
 
 ; Without forcing instructions, fall back to the preferred PS domain.
 ; CHECK: vsel_double
-; CHECK: xor
-; CHECK: and
-; CHECK: andn
-; CHECK: or
+; CHECK: xorps
+; CHECK: andps
+; CHECK: andnps
+; CHECK: orps
 ; CHECK: ret
 
 define void@vsel_double(<4 x double>* %v1, <4 x double>* %v2) {
diff --git a/test/CodeGen/X86/sse41-blend.ll b/test/CodeGen/X86/sse41-blend.ll
index 0a71dd0d04d..78604a0e963 100644
--- a/test/CodeGen/X86/sse41-blend.ll
+++ b/test/CodeGen/X86/sse41-blend.ll
@@ -36,7 +36,6 @@ define <4 x i32> @vsel_i32(<4 x i32> %v1, <4 x i32> %v2) {
 
 
 ;CHECK: vsel_double
-;CHECK-NOT: sra
 ;CHECK: blendvpd
 ;CHECK: ret
 define <4 x double> @vsel_double(<4 x double> %v1, <4 x double> %v2) {
@@ -55,7 +54,6 @@ define <4 x i64> @vsel_i64(<4 x i64> %v1, <4 x i64> %v2) {
 
 
 ;CHECK: vsel_i8
-;CHECK-NOT: sra
 ;CHECK: pblendvb
 ;CHECK: ret
 define <16 x i8> @vsel_i8(<16 x i8> %v1, <16 x i8> %v2) {
@@ -67,7 +65,6 @@ define <16 x i8> @vsel_i8(<16 x i8> %v1, <16 x i8> %v2) {
 ; CHECK: A
 define <2 x double> @A(<2 x double> %x, <2 x double> %y) {
   ; CHECK: cmplepd
-  ; CHECK-NOT: sra
   ; CHECK: blendvpd
   %max_is_x = fcmp oge <2 x double> %x, %y
   %max = select <2 x i1> %max_is_x, <2 x double> %x, <2 x double> %y
@@ -77,7 +74,6 @@ define <2 x double> @A(<2 x double> %x, <2 x double> %y) {
 ; CHECK: B
 define <2 x double> @B(<2 x double> %x, <2 x double> %y) {
   ; CHECK: cmpnlepd
-  ; CHECK-NOT: sra
   ; CHECK: blendvpd
   %min_is_x = fcmp ult <2 x double> %x, %y
   %min = select <2 x i1> %min_is_x, <2 x double> %x, <2 x double> %y