[PATCH] D80964: [X86] Add an Unoptimized Load Value Injection (LVI) Load Hardening Pass

Zola Bridges via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Wed Jun 17 10:13:30 PDT 2020


zbrid added a comment.

In D80964#2097211 <https://reviews.llvm.org/D80964#2097211>, @sconstab wrote:

> From a security perspective, I think that this could become problematic. It would mean that at `-O0` I would get more security than I would at `-O[1-3]`. IMO optimization levels should not work like that.


I don't really think this is a concern. As long as the pass provides //at least// the same level of security as what users need from the LVI pass we can use SESES. The particular implementation used is hidden behind the abstraction of the compiler flag.

A related note is that it seems like the unoptimized LVI pass does not provide the exact same level of security as the graph LVI pass, so this concern is already an issue whether we use SESES or the LVI upoptimized pass.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D80964/new/

https://reviews.llvm.org/D80964





More information about the llvm-commits mailing list