[PATCH] D59188: [DAGCombine] Fold insert_subvector(bitcast(x),bitcast(y)) -> bitcast(insert_subvector(x,y))

[Nikita Popov via Phabricator via llvm-commits]

nikic added inline comments.


================
Comment at: lib/CodeGen/SelectionDAG/DAGCombiner.cpp:18353
+        unsigned Scale = N1SrcSVT.getSizeInBits() / EltSizeInBits;
+        if ((InsIdx % Scale) == 0) {
+          NewVT = EVT::getVectorVT(Ctx, N1SrcSVT, NumElts / Scale);
----------------
Can things like `insert_subvector x:v7i8, (bitcast y:v2i16 to v4i8), 2` occur? Do we need to check `NumElts % Scale == 0` here as well?


Repository:
  rL LLVM

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D59188/new/

https://reviews.llvm.org/D59188