https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122172
--- Comment #9 from Jonathan Wakely <redi at gcc dot gnu.org> --- (In reply to Hans-Peter Nilsson from comment #7) > Thanks for the quick response. I prefer something like the above, not > liking silencing warnings. I'll test this and post a patch, unless you've > already done this of course... I haven't. > ...or is some other stuff needed too? > (Still testing the above patch.) I think the patches in comments 2 and 3 should fix the test regressions. We might also want to revert the changes to atomicity_dir so that the custom cris/atomicity.h is used again, but that's independent. I think it's "only" a code-quality issue, not related to the 1571 test regressions you saw. But the regressions might have helped us notice the code-quality issue.
