[PATCH] D84959: [NewPM][LVI] Abandon LVI after CVP

Arthur Eubanks via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Thu Jul 30 14:05:18 PDT 2020


aeubanks added a comment.

In D84959#2185767 <https://reviews.llvm.org/D84959#2185767>, @aeubanks wrote:

> Hmm I wonder if we can just delete the "Invalidating all non-preserved analyses for" log statement, since "Invalidating analysis: FooAnalysis" does get printed when an analysis is actually invalidated. Or maybe at least only print it only when any analyses are actually invalidated (as you said). I'll take a quick look.
>
> Also, looks like thinlto-distributed-newpm.ll is failing.

Sorry, I think I misunderstood the issue. LazyValueAnalysis is actually being invalidated, right? So "Invalidating all non-preserved analyses for" would get printed regardless if we went with the option of only printing it when an analysis is actually invalidated.
What about removing all "Invalidating all non-preserved analyses for" and checking that there's no "Invalidating analysis: " aside from LazyValueAnalysis? That's probably worse than this solution...


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D84959/new/

https://reviews.llvm.org/D84959



More information about the llvm-commits mailing list