[X86][SSE] Add PEXTR*(PINSR*(v, s, c), c) -> s combine.

We should probably extend this to cover bitcasts as well to help other cases in promote-vec3.ll.

llvm-svn: 367582

diff --git a/llvm/lib/Target/X86/X86ISelLowering.cpp b/llvm/lib/Target/X86/X86ISelLowering.cpp
index c2be03f..4b8987b 100644 (file)
--- a/llvm/lib/Target/X86/X86ISelLowering.cpp
+++ b/llvm/lib/Target/X86/X86ISelLowering.cpp
@@ -35907,15 +35907,26 @@ static SDValue combineExtractVectorElt(SDNode *N, SelectionDAG &DAG,
     }
   }
 
-  // TODO - Remove this once we can handle the implicit zero-extension of
-  // X86ISD::PEXTRW/X86ISD::PEXTRB in:
-  // XFormVExtractWithShuffleIntoLoad, combineHorizontalPredicateResult and
-  // combineBasicSADPattern.
   if (IsPextr) {
     const TargetLowering &TLI = DAG.getTargetLoweringInfo();
     if (TLI.SimplifyDemandedBits(
             SDValue(N, 0), APInt::getAllOnesValue(VT.getSizeInBits()), DCI))
       return SDValue(N, 0);
+
+    // PEXTR*(PINSR*(v, s, c), c) -> s (with implicit zext handling).
+    if ((InputVector.getOpcode() == X86ISD::PINSRB ||
+         InputVector.getOpcode() == X86ISD::PINSRW) &&
+        InputVector.getOperand(2) == EltIdx) {
+      assert(SrcVT == InputVector.getOperand(0).getValueType() &&
+             "Vector type mismatch");
+      SDValue Scl = InputVector.getOperand(1);
+      Scl = DAG.getNode(ISD::TRUNCATE, dl, SrcVT.getScalarType(), Scl);
+      return DAG.getZExtOrTrunc(Scl, dl, VT);
+    }
+
+    // TODO - Remove this once we can handle the implicit zero-extension of
+    // X86ISD::PEXTRW/X86ISD::PEXTRB in XFormVExtractWithShuffleIntoLoad,
+    // combineHorizontalPredicateResult and combineBasicSADPattern.
     return SDValue();
   }
 

diff --git a/llvm/test/CodeGen/X86/promote-vec3.ll b/llvm/test/CodeGen/X86/promote-vec3.ll
index b66570e..6c0a064 100644 (file)
--- a/llvm/test/CodeGen/X86/promote-vec3.ll
+++ b/llvm/test/CodeGen/X86/promote-vec3.ll
@@ -8,15 +8,11 @@
 define <3 x i16> @zext_i8(<3 x i8>) {
 ; SSE3-LABEL: zext_i8:
 ; SSE3:       # %bb.0:
+; SSE3-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
 ; SSE3-NEXT:    movzbl {{[0-9]+}}(%esp), %ecx
 ; SSE3-NEXT:    movzbl {{[0-9]+}}(%esp), %edx
-; SSE3-NEXT:    movzbl {{[0-9]+}}(%esp), %eax
 ; SSE3-NEXT:    movd %eax, %xmm0
 ; SSE3-NEXT:    pextrw $0, %xmm0, %eax
-; SSE3-NEXT:    pinsrw $1, %edx, %xmm0
-; SSE3-NEXT:    pextrw $1, %xmm0, %edx
-; SSE3-NEXT:    pinsrw $2, %ecx, %xmm0
-; SSE3-NEXT:    pextrw $2, %xmm0, %ecx
 ; SSE3-NEXT:    # kill: def $ax killed $ax killed $eax
 ; SSE3-NEXT:    # kill: def $dx killed $dx killed $edx
 ; SSE3-NEXT:    # kill: def $cx killed $cx killed $ecx