From f51e05ffbc28d184db78574f8477338093f358f8 Mon Sep 17 00:00:00 2001
From: Daniel Jasper <djasper@google.com>
Date: Fri, 9 Dec 2016 09:04:51 +0000
Subject: [PATCH] Revert "[SelectionDAG] Add knownbits support for
 EXTRACT_VECTOR_ELT opcodes"

This reverts commit r288916 as it is currently causing a crasher in
Halide. Reproducer on llvm.org/PR31323. While it might be that halide is
generating invalid IR, llc shouldn't crash.

llvm-svn: 289194
---
 llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp | 36 --------------------------
 llvm/test/CodeGen/X86/known-bits-vector.ll     | 12 +++++++--
 2 files changed, 10 insertions(+), 38 deletions(-)

diff --git a/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp b/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
index c652210..f276902 100644
--- a/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
@@ -2590,42 +2590,6 @@ void SelectionDAG::computeKnownBits(SDValue Op, APInt &KnownZero,
     }
     break;
   }
-  case ISD::INSERT_VECTOR_ELT: {
-    SDValue InVec = Op.getOperand(0);
-    SDValue InVal = Op.getOperand(1);
-    SDValue EltNo = Op.getOperand(2);
-
-    ConstantSDNode *CEltNo = dyn_cast<ConstantSDNode>(EltNo);
-    if (CEltNo && CEltNo->getAPIntValue().ult(NumElts)) {
-      // If we know the element index, split the demand between the
-      // source vector and the inserted element.
-      KnownZero = KnownOne = APInt::getAllOnesValue(BitWidth);
-      unsigned EltIdx = CEltNo->getZExtValue();
-
-      // If we demand the inserted element then add its common known bits.
-      if (DemandedElts[EltIdx]) {
-        computeKnownBits(InVal, KnownZero2, KnownOne2, Depth + 1);
-        KnownOne &= KnownOne2.zextOrTrunc(KnownOne.getBitWidth());
-        KnownZero &= KnownZero2.zextOrTrunc(KnownZero.getBitWidth());;
-      }
-
-      // If we demand the source vector then add its common known bits, ensuring
-      // that we don't demand the inserted element.
-      APInt VectorElts = DemandedElts & ~(APInt::getOneBitSet(NumElts, EltIdx));
-      if (!!VectorElts) {
-        computeKnownBits(InVec, KnownZero2, KnownOne2, VectorElts, Depth + 1);
-        KnownOne &= KnownOne2;
-        KnownZero &= KnownZero2;
-      }
-    } else {
-      // Unknown element index, so ignore DemandedElts and demand them all.
-      computeKnownBits(InVec, KnownZero, KnownOne, Depth + 1);
-      computeKnownBits(InVal, KnownZero2, KnownOne2, Depth + 1);
-      KnownOne &= KnownOne2.zextOrTrunc(KnownOne.getBitWidth());
-      KnownZero &= KnownZero2.zextOrTrunc(KnownZero.getBitWidth());;
-    }
-    break;
-  }
   case ISD::BSWAP: {
     computeKnownBits(Op.getOperand(0), KnownZero2, KnownOne2, DemandedElts,
                      Depth + 1);
diff --git a/llvm/test/CodeGen/X86/known-bits-vector.ll b/llvm/test/CodeGen/X86/known-bits-vector.ll
index 717c7c6..944b370 100644
--- a/llvm/test/CodeGen/X86/known-bits-vector.ll
+++ b/llvm/test/CodeGen/X86/known-bits-vector.ll
@@ -58,7 +58,11 @@ define <4 x float> @knownbits_insert_uitofp(<4 x i32> %a0, i16 %a1, i16 %a2) nou
 ; X32-NEXT:    vpinsrd $0, %eax, %xmm0, %xmm0
 ; X32-NEXT:    vpinsrd $2, %ecx, %xmm0, %xmm0
 ; X32-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,0,2,2]
-; X32-NEXT:    vcvtdq2ps %xmm0, %xmm0
+; X32-NEXT:    vpblendw {{.*#+}} xmm1 = xmm0[0],mem[1],xmm0[2],mem[3],xmm0[4],mem[5],xmm0[6],mem[7]
+; X32-NEXT:    vpsrld $16, %xmm0, %xmm0
+; X32-NEXT:    vpblendw {{.*#+}} xmm0 = xmm0[0],mem[1],xmm0[2],mem[3],xmm0[4],mem[5],xmm0[6],mem[7]
+; X32-NEXT:    vaddps {{\.LCPI.*}}, %xmm0, %xmm0
+; X32-NEXT:    vaddps %xmm0, %xmm1, %xmm0
 ; X32-NEXT:    retl
 ;
 ; X64-LABEL: knownbits_insert_uitofp:
@@ -68,7 +72,11 @@ define <4 x float> @knownbits_insert_uitofp(<4 x i32> %a0, i16 %a1, i16 %a2) nou
 ; X64-NEXT:    vpinsrd $0, %eax, %xmm0, %xmm0
 ; X64-NEXT:    vpinsrd $2, %ecx, %xmm0, %xmm0
 ; X64-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[0,0,2,2]
-; X64-NEXT:    vcvtdq2ps %xmm0, %xmm0
+; X64-NEXT:    vpblendw {{.*#+}} xmm1 = xmm0[0],mem[1],xmm0[2],mem[3],xmm0[4],mem[5],xmm0[6],mem[7]
+; X64-NEXT:    vpsrld $16, %xmm0, %xmm0
+; X64-NEXT:    vpblendw {{.*#+}} xmm0 = xmm0[0],mem[1],xmm0[2],mem[3],xmm0[4],mem[5],xmm0[6],mem[7]
+; X64-NEXT:    vaddps {{.*}}(%rip), %xmm0, %xmm0
+; X64-NEXT:    vaddps %xmm0, %xmm1, %xmm0
 ; X64-NEXT:    retq
   %1 = zext i16 %a1 to i32
   %2 = zext i16 %a2 to i32
-- 
2.7.4