[Bug target/67172] [5/6 regression] i686-w64-mingw32 dwarf2 bootstrap fails with undefined reference to __EH_FRAME_BEGIN__

["jakub at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67172

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org

--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
If cygwin or mingw or what wants to ensure EH_FRAME_SECTION_NAME is not set to
the default, then perhaps allow it to be set to NULL too, and change the few
uses of that macro, like:
#ifdef EH_FRAME_SECTION_NAME
      eh_frame_section = get_section (EH_FRAME_SECTION_NAME, flags, NULL);
#else
      eh_frame_section = ((flags == SECTION_WRITE)
                          ? data_section : readonly_data_section);
#endif /* EH_FRAME_SECTION_NAME */
to
#ifdef EH_FRAME_SECTION_NAME
      if (EH_FRAME_SECTION_NAME != NULL)
        eh_frame_section = get_section (EH_FRAME_SECTION_NAME, flags, NULL);
      else
#endif
        eh_frame_section = ((flags == SECTION_WRITE)
                          ? data_section : readonly_data_section);
and
#ifdef EH_FRAME_SECTION_NAME
      builtin_define_with_value ("__LIBGCC_EH_FRAME_SECTION_NAME__",
                                 EH_FRAME_SECTION_NAME, 1);
#endif
to
#ifdef EH_FRAME_SECTION_NAME
      if (EH_FRAME_SECTION_NAME != NULL)
        builtin_define_with_value ("__LIBGCC_EH_FRAME_SECTION_NAME__",
                                   EH_FRAME_SECTION_NAME, 1);
#endif
and then just #define EH_FRAME_SECTION_NAME NULL for the ports that need that
(and add big comment why).  You'd need to adjust also the tm.texi documentation
to explain that NULL means use the data sections as if named sections weren't
supported.