http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52839
--- Comment #10 from Alan Modra <amodra at gmail dot com> 2012-04-04 14:20:57 UTC --- I caught the 64-bit failure in the act. It's dying on the gcc_assert in unwind-dw2.c:_Unwind_SetSpColumn, with the value read from dwarf_reg_size_table[1] being zero. The implication here is that __gthread_once, ie. pthread_once, does not provide proper locking for the power7 weakly ordered memory model. It's too late here tonight to go delving into libpthread to prove my case, but it certainly seems like the 64-bit failure is not a gcc problem.
