[PATCH] D129009: [LTO] Fix LTO for aliased IFuncs

[Schrodinger ZHU Yifan via Phabricator via llvm-commits]

SchrodingerZhu added a comment.

First of all, I am really sorry about the noise and the misbehavior here. I have to admit that some of the changes here are because I am still in the way of getting familiar with the patch system. I apologize for being careless in the progress.

> I don't know how you added reviewers or whether you got upset after I clicked "Request Changes".

I am very glad to recieve any review from you and any other member from the team! It is just that I saw you push some commits to the monorepo recently, so I thought that you were too busy to review the patch. Again, I am sorry for the confusion.

> I haven't taken a second deep look at this patch but the applyAlongResolverPath change looks too intrusive to me. One previous version seems more favorable to me.

Back to the patch itself, I made the changes because in the previous version, I assumed that, along the de-aliasing path to the `resolver` object, the immediate instance of `Constant*` need to be `GlobalObject` or `GlobalAlias`; but the `resolver` is stored as a `Constant *`, behind which, according to the implementation of `findBaseObject`, can also be instantiated as `Expr`.

I am looking forward to your suggestions, on the code and on how to follow the community conventions better!


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D129009/new/

https://reviews.llvm.org/D129009