[PATCH] D72330: [ARM][MVE] Enable masked gathers from base + vector of offsets

[Anna Welker via Phabricator via llvm-commits]

anwel marked 4 inline comments as done.
anwel added inline comments.


================
Comment at: llvm/lib/Target/ARM/MVEGatherScatterLowering.cpp:181
+    // Ensure all the other indices are 0.
+    for (unsigned i = 1; i < FinalIndex; ++i) {
+      auto *C = dyn_cast<Constant>(GEP->getOperand(i));
----------------
dmgreen wrote:
> This code comes from the ISel gather lowering? I think that for the moment we can just check that the number of GEP operands is 2 (Base + offset) and remove this loop. I'm not exactly sure when it would be useful. We don't seem to have any tests for it, and can add it back in later if they come up in the future. (Otherwise the types don't sound like they would be correct, and the getOperand(1) should be getOperand(FinalIndex)?)
As as of now I cannot present a test that would show this is actually relevant I agree.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D72330/new/

https://reviews.llvm.org/D72330