[PATCH] D115497: [Inline] Disable deferred inlining

[Wenlei He via Phabricator via llvm-commits]

wenlei added a comment.

Are you using newpm with PGO? FWIW, we've also ran into bloated inlining multiple times after switching to npm a while back, but most of that turned out to be due to a combination of npm's better use of BFI for inline and some kind of profiling issue leading to unusual flat profile which makes inliner super aggressive (flat profile has a relative lower hot threshold so more callsites are considered hot).

We haven't found problem with inline deferral so far, though we also haven't measured its perf benefit..


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D115497/new/

https://reviews.llvm.org/D115497