[PATCH] D51906: [WebAssembly] SIMD shifts

[Thomas Lively via Phabricator via llvm-commits]

tlively added a comment.

In https://reviews.llvm.org/D51906#1235616, @aheejin wrote:

> Does this work when the first argument is not in a splat pattern as well? If so, maybe add a test case for that?


No, our shift is weird because it takes only a single shift amount to apply uniformly to all lanes rather than a vector of shift amounts. Currently LLVM shifts by non-splat vectors will fail in isel, but that's ok for now.


Repository:
  rL LLVM

https://reviews.llvm.org/D51906