[PATCH] D79454: [IR] `byval` arguments cause reads at call sites

[Eli Friedman via Phabricator via llvm-commits]

efriedma added a comment.

> Now that I wrote this I am concerned we already have this problem without byval

If you do an inbounds GEP on a deallocated pointer, I think the "inbounds"-ness should apply to what the allocation looked like during its lifetime.  (The only alternative is that gep effectively reads memory, and that's obviously a lot more problematic.)

> perhaps we really want the optimizer to just remove the byval attribute?

In general, we want the optimizer to remove the byval attribute; it's terrible for optimization and code generation.  But removing it changes the calling convention, so we can only do it in limited circumstances.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D79454/new/

https://reviews.llvm.org/D79454