X86: Handle poison in buildFromShuffleMostly

[arsenm]

No description provided.

[arsenm]

• X86: Handle poison in buildFromShuffleMostly #168218 👈 (View in Graphite)
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[llvmbot]

@llvm/pr-subscribers-backend-x86

Author: Matt Arsenault (arsenm)

Changes

---

Full diff: https://github.com/llvm/llvm-project/pull/168218.diff

2 Files Affected:

• (modified) llvm/lib/Target/X86/X86ISelLowering.cpp (+2-2)
• (modified) llvm/test/CodeGen/X86/vselect.ll (+3-3)

diff --git a/llvm/lib/Target/X86/X86ISelLowering.cpp b/llvm/lib/Target/X86/X86ISelLowering.cpp
index a2f505b9cff8a..66f5802a67465 100644
--- a/llvm/lib/Target/X86/X86ISelLowering.cpp
+++ b/llvm/lib/Target/X86/X86ISelLowering.cpp
@@ -7959,7 +7959,7 @@ static SDValue buildFromShuffleMostly(SDValue Op, const SDLoc &DL,
   for (unsigned i = 0; i != NumElems; ++i) {
     unsigned Opc = Op.getOperand(i).getOpcode();
 
-    if (Opc == ISD::UNDEF)
+    if (Opc == ISD::POISON || Opc == ISD::UNDEF)
       continue;
 
     if (Opc != ISD::EXTRACT_VECTOR_ELT) {
@@ -8002,7 +8002,7 @@ static SDValue buildFromShuffleMostly(SDValue Op, const SDLoc &DL,
   if (!VecIn1.getNode())
     return SDValue();
 
-  VecIn2 = VecIn2.getNode() ? VecIn2 : DAG.getUNDEF(VT);
+  VecIn2 = VecIn2.getNode() ? VecIn2 : DAG.getPOISON(VT);
   SDValue NV = DAG.getVectorShuffle(VT, DL, VecIn1, VecIn2, Mask);
 
   for (unsigned Idx : InsertIndices)
diff --git a/llvm/test/CodeGen/X86/vselect.ll b/llvm/test/CodeGen/X86/vselect.ll
index 0fe3edab4ac38..f70145d6b21c2 100644
--- a/llvm/test/CodeGen/X86/vselect.ll
+++ b/llvm/test/CodeGen/X86/vselect.ll
@@ -641,10 +641,10 @@ define <2 x i32> @simplify_select(i32 %x, <2 x i1> %z) {
 ; SSE2-NEXT:    movd %edi, %xmm1
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm1[1,0,1,1]
 ; SSE2-NEXT:    por %xmm1, %xmm2
-; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm2[1,1,1,1]
-; SSE2-NEXT:    punpckldq {{.*#+}} xmm3 = xmm3[0],xmm1[0],xmm3[1],xmm1[1]
+; SSE2-NEXT:    punpcklqdq {{.*#+}} xmm1 = xmm1[0,0]
+; SSE2-NEXT:    shufps {{.*#+}} xmm1 = xmm1[2,0],xmm2[1,1]
 ; SSE2-NEXT:    pand %xmm0, %xmm2
-; SSE2-NEXT:    pandn %xmm3, %xmm0
+; SSE2-NEXT:    pandn %xmm1, %xmm0
 ; SSE2-NEXT:    por %xmm2, %xmm0
 ; SSE2-NEXT:    retq
 ;