[PATCH] D93376: [LangRef] Clarify the semantics of lifetime intrinsics

[Juneyoung Lee via Phabricator via llvm-commits]

aqjune added a comment.

Maybe the root of the problem was that LangRef was too obscure about what the intrinsic meant. :/
If there is a well-known use case that does not involve alloca, then this will seriously break the case, so the semantics of lifetime.start/end should consider it from the start.
But we haven't found one, and I think the dereferenceable_or_null on malloc is actually against the interpretation (maybe I should have found it before writing this patch and share it to reduce discussion cost, sorry).
So I think it is okay to specify it as working on alloca only first, and if there is any really problematic case, extend the lifetime's semantics as discussed in this thread.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D93376/new/

https://reviews.llvm.org/D93376