[TargetLowering] SimplifyDemandedBits - call SimplifyMultipleUseDemandedBits for ISD::TRUNCATE

llvm-svn: 368553

2 files changed, 13 insertions, 10 deletions

diff --git a/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp b/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
index 411d146e2ab..9027bea0390 100644
--- a/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp
@@ -1700,6 +1700,11 @@ bool TargetLowering::SimplifyDemandedBits(
       return true;
     Known = Known.trunc(BitWidth);
 
+    // Attempt to avoid multi-use ops if we don't need anything from them.
+    if (SDValue NewSrc = SimplifyMultipleUseDemandedBits(
+            Src, TruncMask, DemandedElts, TLO.DAG, Depth + 1))
+      return TLO.CombineTo(Op, TLO.DAG.getNode(ISD::TRUNCATE, dl, VT, NewSrc));
+
     // If the input is only used by this truncate, see if we can shrink it based
     // on the known demanded bits.
     if (Src.getNode()->hasOneUse()) {
diff --git a/llvm/test/CodeGen/X86/bitcast-vector-bool.ll b/llvm/test/CodeGen/X86/bitcast-vector-bool.ll
index 07ae428e24b..57aa53199af 100644
--- a/llvm/test/CodeGen/X86/bitcast-vector-bool.ll
+++ b/llvm/test/CodeGen/X86/bitcast-vector-bool.ll
@@ -376,12 +376,11 @@ define i16 @bitcast_v32i8_to_v2i16(<32 x i8> %a0) nounwind {
 ;
 ; AVX1-LABEL: bitcast_v32i8_to_v2i16:
 ; AVX1:       # %bb.0:
-; AVX1-NEXT:    vpmovmskb %xmm0, %eax
-; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm0
 ; AVX1-NEXT:    vpmovmskb %xmm0, %ecx
-; AVX1-NEXT:    shll $16, %ecx
-; AVX1-NEXT:    orl %eax, %ecx
-; AVX1-NEXT:    vmovd %ecx, %xmm0
+; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm0
+; AVX1-NEXT:    vpmovmskb %xmm0, %eax
+; AVX1-NEXT:    shll $16, %eax
+; AVX1-NEXT:    vmovd %eax, %xmm0
 ; AVX1-NEXT:    vpextrw $1, %xmm0, %eax
 ; AVX1-NEXT:    addl %ecx, %eax
 ; AVX1-NEXT:    # kill: def $ax killed $ax killed $eax
@@ -632,13 +631,12 @@ define i16 @bitcast_v32i16_to_v2i16(<32 x i16> %a0) nounwind {
 ; AVX1:       # %bb.0:
 ; AVX1-NEXT:    vextractf128 $1, %ymm0, %xmm2
 ; AVX1-NEXT:    vpacksswb %xmm2, %xmm0, %xmm0
-; AVX1-NEXT:    vpmovmskb %xmm0, %eax
+; AVX1-NEXT:    vpmovmskb %xmm0, %ecx
 ; AVX1-NEXT:    vextractf128 $1, %ymm1, %xmm0
 ; AVX1-NEXT:    vpacksswb %xmm0, %xmm1, %xmm0
-; AVX1-NEXT:    vpmovmskb %xmm0, %ecx
-; AVX1-NEXT:    shll $16, %ecx
-; AVX1-NEXT:    orl %eax, %ecx
-; AVX1-NEXT:    vmovd %ecx, %xmm0
+; AVX1-NEXT:    vpmovmskb %xmm0, %eax
+; AVX1-NEXT:    shll $16, %eax
+; AVX1-NEXT:    vmovd %eax, %xmm0
 ; AVX1-NEXT:    vpextrw $1, %xmm0, %eax
 ; AVX1-NEXT:    addl %ecx, %eax
 ; AVX1-NEXT:    # kill: def $ax killed $ax killed $eax