https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122172
--- Comment #19 from Hans-Peter Nilsson <hp at gcc dot gnu.org> --- (In reply to Hans-Peter Nilsson from comment #17) > FWIW, a patch constructed from comments #2, #3 and the fixup in comment #16 > yields two regressions at r16-4259-g9242a89d088f44 for cris-elf compared to > r16-3809-g46028a2f40329b > g++.sum g++-dg-lto-devirt-2-01.exe > g++.sum g++.dg/ipa/devirt-2.C JFTR, those two regressions were "already" there at r16-4208-g8cd2db265e73c5. I'll be deal with those in separate PRs, though hopefully they're already fixed, or at least having PRs open, when this here PR is fixed.
