[PATCH] D88897: [AArch64] WIP for review D88742

[Vinay Madhusudan via Phabricator via llvm-commits]

mivnay added a comment.

In D88897#2314408 <https://reviews.llvm.org/D88897#2314408>, @dmgreen wrote:

> There's some code in performExtendCombine from the look of it, still looking at the intrinsics ID's.

oops! I thought it would still match the intrinsic and completely ignored it. Sorry! I have fixed it and added support for SABD node as well. Now, there is a new error because of the "FALLBACK-NOT: remark" checks in the same file.

  Selecting:
    %4:fpr64(<8 x s8>) = G_INTRINSIC intrinsic(@llvm.aarch64.neon.sabd), %2:fpr(<8 x s8>), %3:fpr(<8 x s8>)
  remark: <unknown>:0:0: cannot select: %4:fpr64(<8 x s8>) = G_INTRINSIC intrinsic(@llvm.aarch64.neon.sabd), %2:fpr(<8 x s8>), %3:fpr(<8 x s8>) (in function: sabdl8h)
  warning: Instruction selection used fallback path for sabdl8h


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D88897/new/

https://reviews.llvm.org/D88897