https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78597
--- Comment #1 from joseph at codesourcery dot com <joseph at codesourcery dot com> --- If these fail but weren't previously run then it's not a regression, but an indication of a wrong-code bug in the float128 support for powerpc (for which testing was previously very limited - even now, the _FloatN / _FloatNx tests are fairly basic).