[PATCH] D106164: [LV] Don't assume isScalarAfterVectorization if one of the uses needs widening.

[Sander de Smalen via Phabricator via llvm-commits]

sdesmalen added inline comments.


================
Comment at: llvm/test/Transforms/LoopVectorize/AArch64/sve-widen-gep.ll:4
+
+target triple = "aarch64-unknown-linux-gnu"
+
----------------
fhahn wrote:
> Do we still need this test? The issue is not  SVE/AArch64 specific and it should already be covered by the general Test in Pointer-induction.ll.
I'm a bit cautious about someone just regenerating the pointer-induction.ll test and reviewers not realising this is a big functional regression, because for scalable vectors it actually causes something to break. This test would guard that.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D106164/new/

https://reviews.llvm.org/D106164