Re: x86-64 large data sections updated

[Jan Hubicka <jh at suse dot cz>]

> On Wed, Jun 15, 2005 at 12:13:15AM +0200, Jan Hubicka wrote:
> > *************** get_program_header_size (bfd *abfd)
> > *** 4593,4598 ****
> > --- 4593,4607 ----
> >         segs += 2;
> >       }
> >   
> > +   if (bfd_get_section_by_name (abfd, ".ldata") != NULL
> > +       || bfd_get_section_by_name (abfd, ".lrodata") != NULL
> > +       || bfd_get_section_by_name (abfd, ".lbss") != NULL)
> 
> This shouldn't be in elf.c.  And you can need 2 segments, not
> just one.
> 
> > + elf64_x86_64_elf_section_from_bfd_section (bfd *abfd ATTRIBUTE_UNUSED,
> > + 					asection *sec, int *retval)
> > + {
> > +   if (strcmp (bfd_get_section_name (abfd, sec), ".lcommon") == 0
> > +       || strcmp (bfd_get_section_name (abfd, sec), ".lbss") == 0)
> > +     {
> > +       *retval = SHN_X86_64_LCOMMON;
> 
> You never ever want this.  COMMON does not mean the same thing as bss.

Hmm, actually this is only place where SHN_X86_64_LCOMMON is introduced.
If I set the section as it is constructed from tc-i386.c I still get
GLOBAL visibility out of resulting file and I am not quite sure where
the SHN_X86_64_LCOMMON should be introduced (I originally attempted to
set it directly in gas when creating the symbol but that thing don't
seems to be used).
Perhaps I am constructing the .lbss section incorrectly but I am not
quite sure what is proper way here...

Honza
> 
> 
> r~