[PATCH] D81862: [gicombiner] Allow generated CombinerHelpers to have additional arguments

Daniel Sanders via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Mon Jun 15 15:29:48 PDT 2020


dsanders added a comment.

In D81862#2094055 <https://reviews.llvm.org/D81862#2094055>, @aditya_nandakumar wrote:

> Doesn't https://reviews.llvm.org/D81863 make this unnecessary?


No, at the moment some targets are still using this for their CombinerHelper. Even if we make CombinerHelper a standard one there's still potential for this to be used by non-static things like the results of compute known bits* or some other (possibly target specific) localized knowledge. Essentially, D81863 <https://reviews.llvm.org/D81863> covers static/global knowledge and this covers dynamic/local knowledge.

*The analysis pass itself would be static and but the results are dynamic so it really comes down to the preferred way to use that one.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D81862/new/

https://reviews.llvm.org/D81862





More information about the llvm-commits mailing list