[PATCH] D80208: [CodeGen] Add support for extracting elements of scalable vectors

[Eli Friedman via Phabricator via llvm-commits]

efriedma added inline comments.


================
Comment at: llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp:5374
     // expanding copies of large vectors from registers.
-    if (N2C &&
-        N1.getOpcode() == ISD::CONCAT_VECTORS &&
-        N1.getNumOperands() > 0) {
+    if (N2C && N1.getOperand(0).getValueType().isFixedLengthVector() &&
+        N1.getOpcode() == ISD::CONCAT_VECTORS && N1.getNumOperands() > 0) {
----------------
Please note why this doesn't work for scalable vectors.


================
Comment at: llvm/lib/CodeGen/SelectionDAG/SelectionDAG.cpp:5428
     if (N1.getOpcode() == ISD::EXTRACT_SUBVECTOR &&
+        N1.getValueType().isFixedLengthVector() &&
         N1.getValueType().getVectorNumElements() == 1) {
----------------
Please note why this doesn't work for scalable vectors.


================
Comment at: llvm/lib/Target/AArch64/AArch64SVEInstrInfo.td:1875
+                                        (INDEX_II_B 0, 1),
+                                        (DUP_ZR_B (i32 (EXTRACT_SUBREG GPR64:$index, sub_32)))),
+                         ZPR:$vec)>;
----------------
Maybe better to use `lastb(whilels(0, idx), vec)` or something like that?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D80208/new/

https://reviews.llvm.org/D80208