[llvm] 68aeaab - [X86] Don't count the chain uses when forming broadcast loads in lowerBuildVectorAsBroadcast.

[Craig Topper via llvm-commits]

Author: Craig Topper
Date: 2020-03-03T08:41:31-08:00
New Revision: 68aeaab888094fa192bf9257ffd5929b93778a70

URL: https://github.com/llvm/llvm-project/commit/68aeaab888094fa192bf9257ffd5929b93778a70
DIFF: https://github.com/llvm/llvm-project/commit/68aeaab888094fa192bf9257ffd5929b93778a70.diff

LOG: [X86] Don't count the chain uses when forming broadcast loads in lowerBuildVectorAsBroadcast.

The build_vector needs to be the only user of the data, but the
chain will likely have another use. So we can't make sure the
build_vector is the only user of the node.

Added: 
    

Modified: 
    llvm/lib/Target/X86/X86ISelLowering.cpp
    llvm/test/CodeGen/X86/pr35443.ll

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Target/X86/X86ISelLowering.cpp b/llvm/lib/Target/X86/X86ISelLowering.cpp
index 4c3fd0d75e78..cef768e95dda 100644
--- a/llvm/lib/Target/X86/X86ISelLowering.cpp
+++ b/llvm/lib/Target/X86/X86ISelLowering.cpp
@@ -8621,10 +8621,12 @@ static SDValue lowerBuildVectorAsBroadcast(BuildVectorSDNode *BVOp,
 
   bool ConstSplatVal =
       (Ld.getOpcode() == ISD::Constant || Ld.getOpcode() == ISD::ConstantFP);
+  bool IsLoad = ISD::isNormalLoad(Ld.getNode());
 
   // Make sure that all of the users of a non-constant load are from the
   // BUILD_VECTOR node.
-  if (!ConstSplatVal && !BVOp->isOnlyUserOf(Ld.getNode()))
+  // FIXME: Is the use count needed for non-constant, non-load case?
+  if (!ConstSplatVal && !IsLoad && !BVOp->isOnlyUserOf(Ld.getNode()))
     return SDValue();
 
   unsigned ScalarSize = Ld.getValueSizeInBits();
@@ -8674,8 +8676,6 @@ static SDValue lowerBuildVectorAsBroadcast(BuildVectorSDNode *BVOp,
     }
   }
 
-  bool IsLoad = ISD::isNormalLoad(Ld.getNode());
-
   // Handle AVX2 in-register broadcasts.
   if (!IsLoad && Subtarget.hasInt256() &&
       (ScalarSize == 32 || (IsGE256 && ScalarSize == 64)))
@@ -8685,6 +8685,10 @@ static SDValue lowerBuildVectorAsBroadcast(BuildVectorSDNode *BVOp,
   if (!IsLoad)
     return SDValue();
 
+  // Make sure the non-chain result is only used by this build vector.
+  if (!Ld->hasNUsesOfValue(NumElts - NumUndefElts, 0))
+    return SDValue();
+
   if (ScalarSize == 32 || (IsGE256 && ScalarSize == 64) ||
       (Subtarget.hasVLX() && ScalarSize == 64))
     return DAG.getNode(X86ISD::VBROADCAST, dl, VT, Ld);

diff  --git a/llvm/test/CodeGen/X86/pr35443.ll b/llvm/test/CodeGen/X86/pr35443.ll
index a42a14bdbcab..01001b021bec 100644
--- a/llvm/test/CodeGen/X86/pr35443.ll
+++ b/llvm/test/CodeGen/X86/pr35443.ll
@@ -8,8 +8,7 @@
 define void @pr35443() {
 ; CHECK-LABEL: pr35443:
 ; CHECK:       # %bb.0: # %entry
-; CHECK-NEXT:    movzbl ac+{{.*}}(%rip), %eax
-; CHECK-NEXT:    vmovd %eax, %xmm0
+; CHECK-NEXT:    vpbroadcastb ac+{{.*}}(%rip), %xmm0
 ; CHECK-NEXT:    vpmovzxbq {{.*#+}} ymm0 = xmm0[0],zero,zero,zero,zero,zero,zero,zero,xmm0[1],zero,zero,zero,zero,zero,zero,zero,xmm0[2],zero,zero,zero,zero,zero,zero,zero,xmm0[3],zero,zero,zero,zero,zero,zero,zero
 ; CHECK-NEXT:    vpxor %xmm1, %xmm1, %xmm1
 ; CHECK-NEXT:    vpsubq %ymm0, %ymm1, %ymm0