https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68730
--- Comment #12 from Ilya Enkovich <ienkovich at gcc dot gnu.org> --- (In reply to Jakub Jelinek from comment #11) > Reverting r232500 and r232413 results in the miscompilation again. So if > those changes are optimizations only we have a latent issue. Those changes fix performance regression and were not supposed to resolve any stability issue.