From: Matt Arsenault <Matthew.Arsenault@amd.com>
Date: Tue, 2 Sep 2014 18:33:51 +0000 (+0000)
Subject: Fix comment and unnecessary check for FP build_vectors.
X-Git-Url: http://demsky.eecs.uci.edu/git/?a=commitdiff_plain;h=1c7650f67cf8399b79af28ee42a45fffd5ffe77d;p=oota-llvm.git

Fix comment and unnecessary check for FP build_vectors.

This was copy-paste from the integer version, but
FP build_vectors don't truncate.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@216928 91177308-0d34-0410-b5e6-96231b3b80d8
---

diff --git a/lib/CodeGen/SelectionDAG/DAGCombiner.cpp b/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
index 1186683aeba..e556e74980c 100644
--- a/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
+++ b/lib/CodeGen/SelectionDAG/DAGCombiner.cpp
@@ -711,11 +711,7 @@ static ConstantFPSDNode *isConstOrConstSplatFP(SDValue N) {
     BitVector UndefElements;
     ConstantFPSDNode *CN = BV->getConstantFPSplatNode(&UndefElements);
 
-    // BuildVectors can truncate their operands. Ignore that case here.
-    // FIXME: We blindly ignore splats which include undef which is overly
-    // pessimistic.
-    if (CN && UndefElements.none() &&
-        CN->getValueType(0) == N.getValueType().getScalarType())
+    if (CN && UndefElements.none())
       return CN;
   }