DAG: Avoid forming shufflevector from a single extract_vector_elt (#122672)

This avoids regressions in a future AMDGPU commit. Previously we
would have a build_vector (extract_vector_elt x), undef with free
access to the elements bloated into a shuffle of one element + undef,
which has much worse combine support than the extract.

Alternatively could check aggressivelyPreferBuildVectorSources, but
I'm not sure it's really different than isExtractVecEltCheap.

Author: arsenm

llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp

@@ -23807,6 +23807,13 @@ SDValue DAGCombiner::reduceBuildVecToShuffle(SDNode *N) {
   SmallVector<SDValue, 8> VecIn;
   VecIn.push_back(SDValue());
 
+  // If we have a single extract_element with a constant index, track the index
+  // value.
+  unsigned OneConstExtractIndex = ~0u;
+
+  // Count the number of extract_vector_elt sources (i.e. non-constant or undef)
+  unsigned NumExtracts = 0;
+
   for (unsigned i = 0; i != NumElems; ++i) {
     SDValue Op = N->getOperand(i);
 
@@ -23824,23 +23831,28 @@ SDValue DAGCombiner::reduceBuildVecToShuffle(SDNode *N) {
 
     // Not an undef or zero. If the input is something other than an
     // EXTRACT_VECTOR_ELT with an in-range constant index, bail out.
-    if (Op.getOpcode() != ISD::EXTRACT_VECTOR_ELT ||
-        !isa<ConstantSDNode>(Op.getOperand(1)))
+    if (Op.getOpcode() != ISD::EXTRACT_VECTOR_ELT)
       return SDValue();
-    SDValue ExtractedFromVec = Op.getOperand(0);
 
+    SDValue ExtractedFromVec = Op.getOperand(0);
     if (ExtractedFromVec.getValueType().isScalableVector())
       return SDValue();
+    auto *ExtractIdx = dyn_cast<ConstantSDNode>(Op.getOperand(1));
+    if (!ExtractIdx)
+      return SDValue();
 
-    const APInt &ExtractIdx = Op.getConstantOperandAPInt(1);
-    if (ExtractIdx.uge(ExtractedFromVec.getValueType().getVectorNumElements()))
+    if (ExtractIdx->getAsAPIntVal().uge(
+            ExtractedFromVec.getValueType().getVectorNumElements()))
       return SDValue();
 
     // All inputs must have the same element type as the output.
     if (VT.getVectorElementType() !=
         ExtractedFromVec.getValueType().getVectorElementType())
       return SDValue();
 
+    OneConstExtractIndex = ExtractIdx->getZExtValue();
+    ++NumExtracts;
+
     // Have we seen this input vector before?
     // The vectors are expected to be tiny (usually 1 or 2 elements), so using
     // a map back from SDValues to numbers isn't worth it.
@@ -23863,6 +23875,20 @@ SDValue DAGCombiner::reduceBuildVecToShuffle(SDNode *N) {
   // VecIn accordingly.
   bool DidSplitVec = false;
   if (VecIn.size() == 2) {
+    // If we only found a single constant indexed extract_vector_elt feeding the
+    // build_vector, do not produce a more complicated shuffle if the extract is
+    // cheap with other constant/undef elements. Skip broadcast patterns with
+    // multiple uses in the build_vector.
+
+    // TODO: This should be more aggressive about skipping the shuffle
+    // formation, particularly if VecIn[1].hasOneUse(), and regardless of the
+    // index.
+    if (NumExtracts == 1 &&
+        TLI.isOperationLegalOrCustom(ISD::EXTRACT_VECTOR_ELT, VT) &&
+        TLI.isTypeLegal(VT.getVectorElementType()) &&
+        TLI.isExtractVecEltCheap(VT, OneConstExtractIndex))
+      return SDValue();
+
     unsigned MaxIndex = 0;
     unsigned NearestPow2 = 0;
     SDValue Vec = VecIn.back();

llvm/test/CodeGen/AMDGPU/insert_vector_dynelt.ll

@@ -452,11 +452,11 @@ define amdgpu_kernel void @byte8_inselt(ptr addrspace(1) %out, <8 x i8> %vec, i3
 ; GCN-NEXT:    s_and_b32 s6, s4, 0x1010101
 ; GCN-NEXT:    s_andn2_b64 s[2:3], s[2:3], s[4:5]
 ; GCN-NEXT:    s_or_b64 s[2:3], s[6:7], s[2:3]
-; GCN-NEXT:    v_mov_b32_e32 v3, s1
-; GCN-NEXT:    v_mov_b32_e32 v0, s2
-; GCN-NEXT:    v_mov_b32_e32 v1, s3
-; GCN-NEXT:    v_mov_b32_e32 v2, s0
-; GCN-NEXT:    flat_store_dwordx2 v[2:3], v[0:1]
+; GCN-NEXT:    v_mov_b32_e32 v0, s0
+; GCN-NEXT:    v_mov_b32_e32 v2, s2
+; GCN-NEXT:    v_mov_b32_e32 v1, s1
+; GCN-NEXT:    v_mov_b32_e32 v3, s3
+; GCN-NEXT:    flat_store_dwordx2 v[0:1], v[2:3]
 ; GCN-NEXT:    s_endpgm
 entry:
   %v = insertelement <8 x i8> %vec, i8 1, i32 %sel