From 84fef1f55d6be054fa8a42c1561bcc97d0dd0f11 Mon Sep 17 00:00:00 2001
From: Bill Schmidt <wschmidt@linux.vnet.ibm.com>
Date: Mon, 4 Aug 2014 23:21:01 +0000
Subject: [PATCH] [PPC64LE] Fix wrong IR for vec_sld and vec_vsldoi

My original LE implementation of the vsldoi instruction, with its
altivec.h interfaces vec_sld and vec_vsldoi, produces incorrect
shufflevector operations in the LLVM IR.  Correct code is generated
because the back end handles the incorrect shufflevector in a
consistent manner.

This patch and a companion patch for Clang correct this problem by
removing the fixup from altivec.h and the corresponding fixup from the
PowerPC back end.  Several test cases are also modified to reflect the
now-correct LLVM IR.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@214800 91177308-0d34-0410-b5e6-96231b3b80d8
---
 lib/Target/PowerPC/PPCISelLowering.cpp | 42 ++++++++------------------
 test/CodeGen/PowerPC/vec_shuffle_le.ll |  4 +--
 2 files changed, 14 insertions(+), 32 deletions(-)

diff --git a/lib/Target/PowerPC/PPCISelLowering.cpp b/lib/Target/PowerPC/PPCISelLowering.cpp
index 4028fedb6c5..8ddfd776726 100644
--- a/lib/Target/PowerPC/PPCISelLowering.cpp
+++ b/lib/Target/PowerPC/PPCISelLowering.cpp
@@ -1009,38 +1009,20 @@ int PPC::isVSLDOIShuffleMask(SDNode *N, bool isUnary, SelectionDAG &DAG) {
   unsigned ShiftAmt = SVOp->getMaskElt(i);
   if (ShiftAmt < i) return -1;
 
-  if (DAG.getTarget().getSubtargetImpl()->getDataLayout()->isLittleEndian()) {
-
-    ShiftAmt += i;
-
-    if (!isUnary) {
-      // Check the rest of the elements to see if they are consecutive.
-      for (++i; i != 16; ++i)
-        if (!isConstantOrUndef(SVOp->getMaskElt(i), ShiftAmt - i))
-          return -1;
-    } else {
-      // Check the rest of the elements to see if they are consecutive.
-      for (++i; i != 16; ++i)
-        if (!isConstantOrUndef(SVOp->getMaskElt(i), (ShiftAmt - i) & 15))
-          return -1;
-    }
-
-  } else {  // Big Endian
+  ShiftAmt -= i;
 
-    ShiftAmt -= i;
-
-    if (!isUnary) {
-      // Check the rest of the elements to see if they are consecutive.
-      for (++i; i != 16; ++i)
-        if (!isConstantOrUndef(SVOp->getMaskElt(i), ShiftAmt+i))
-          return -1;
-    } else {
-      // Check the rest of the elements to see if they are consecutive.
-      for (++i; i != 16; ++i)
-        if (!isConstantOrUndef(SVOp->getMaskElt(i), (ShiftAmt+i) & 15))
-          return -1;
-    }
+  if (!isUnary) {
+    // Check the rest of the elements to see if they are consecutive.
+    for (++i; i != 16; ++i)
+      if (!isConstantOrUndef(SVOp->getMaskElt(i), ShiftAmt+i))
+        return -1;
+  } else {
+    // Check the rest of the elements to see if they are consecutive.
+    for (++i; i != 16; ++i)
+      if (!isConstantOrUndef(SVOp->getMaskElt(i), (ShiftAmt+i) & 15))
+        return -1;
   }
+
   return ShiftAmt;
 }
 
diff --git a/test/CodeGen/PowerPC/vec_shuffle_le.ll b/test/CodeGen/PowerPC/vec_shuffle_le.ll
index 4fd41a7830d..efebece09f4 100644
--- a/test/CodeGen/PowerPC/vec_shuffle_le.ll
+++ b/test/CodeGen/PowerPC/vec_shuffle_le.ll
@@ -189,7 +189,7 @@ entry:
 ; CHECK: VSLDOI_xy:
         %tmp = load <16 x i8>* %A
         %tmp2 = load <16 x i8>* %B
-        %tmp3 = shufflevector <16 x i8> %tmp, <16 x i8> %tmp2, <16 x i32> <i32 19, i32 18, i32 17, i32 16, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4>
+        %tmp3 = shufflevector <16 x i8> %tmp, <16 x i8> %tmp2, <16 x i32> <i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27>
 ; CHECK: vsldoi
         store <16 x i8> %tmp3, <16 x i8>* %A
         ret void
@@ -199,7 +199,7 @@ define void @VSLDOI_xx(<16 x i8>* %A) {
 entry:
 ; CHECK: VSLDOI_xx:
         %tmp = load <16 x i8>* %A
-        %tmp2 = shufflevector <16 x i8> %tmp, <16 x i8> %tmp, <16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4>
+        %tmp2 = shufflevector <16 x i8> %tmp, <16 x i8> %tmp, <16 x i32> <i32 12, i32 13, i32 14, i32 15, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11>
 ; CHECK: vsldoi
         store <16 x i8> %tmp2, <16 x i8>* %A
         ret void
-- 
2.34.1