https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93599
--- Comment #7 from Thomas Koenig <tkoenig at gcc dot gnu.org> --- Hi Jerry, this is without Janne's patch (which, as far as I see, concerns handling I/O when the threads have already started). Regarding testing: I could not get the original test case to fail on gcc135, a POWER9 machine.
