[PATCH] D68964: cmake/modules/CheckAtomic.cmake: catch false positives in RISC-V

[JF Bastien via Phabricator via llvm-commits]

jfb added a comment.

In D68964#1711052 <https://reviews.llvm.org/D68964#1711052>, @jyknight wrote:

> This is bogus -- the RISCV GCC implementation is broken. It should not be mixing libatomic function calls with direct atomics.


Yeah, but if it doesn't work then we should work around it. Using a single atomic load in the test, or using multiple atomic instructions which are representative of what LLVM actually uses, is fine by me. The test is still simple, it's still brittle and silly, but then it lets them build so that's fine.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D68964/new/

https://reviews.llvm.org/D68964