[PATCH] D100088: [DAGCombiner] Fold step_vector with add/mul/shl

[David Sherwood via Phabricator via llvm-commits]

david-arm added a comment.

The folds here look sensible to me! I just have a minor comment about the add fold.



================
Comment at: llvm/lib/CodeGen/SelectionDAG/DAGCombiner.cpp:2523
+    EVT SVT = N1.getOperand(0).getValueType();
+    SDValue NewStep = DAG.getConstant(SV0 + SV1, DL, SVT);
+    SDValue SV = DAG.getStepVector(DL, VT, NewStep);
----------------
Is it worth adding an assert here like this?

  assert(N1.getOperand(0).getValueType() == N0.getOperand(1)->getOperand(0).getValueType());



Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D100088/new/

https://reviews.llvm.org/D100088