http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57389
--- Comment #4 from chrbr at gcc dot gnu.org --- Created attachment 30207 --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=30207&action=edit Patch to avoid assertion This patches restores to the previous state that don't check regno parameters produced by dwarf_register_span. necessary only for the powerpc, I don't really understand the reason why it produces a parallel with both a hard_reg and a dbx.
