[PATCH] D114580: [AArch64][SVE] Mark fixed-type FP extending/truncating loads/stores as custom

[Bradley Smith via Phabricator via llvm-commits]

bsmith added inline comments.


================
Comment at: llvm/lib/Target/AArch64/AArch64ISelLowering.cpp:1545
+  if (VT.isFloatingPoint()) {
+    MVT InnerVT = VT.changeVectorElementType(MVT::i8);
+    while (InnerVT != VT) {
----------------
paulwalker-arm wrote:
> This looks weird to me, shouldn't InnerVT be floating point? I guess the reason this works is because the i8 case is essentially bogus and you end up with the necessary floating point types from the second iteration.  Either way I think this wants to be `MVT::f16`.
I meant to change this but clearly forgot!


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D114580/new/

https://reviews.llvm.org/D114580