[PATCH v10 16/28] OAv2 merge: create a section for object attributes

[Jan Beulich]

On 20.11.2025 18:59, Matthieu Longo wrote:
> bfd copies object attributes to the output BFD only if the input BFD —
> used to store the merge result— contains an object attribute section.
> bfd_linear_find_first_with_obj_attrs() returns an input BFD that may or
> may not already include such a section.  If no object attributes section
> is found, the merge process must create one.
> 
> This patch implements create_object_attributes_section(), which constructs
> the missing section with the appropriate flags, type, and alignment.  The
> implementation is adapted from _bfd_elf_link_create_gnu_property_sec() in
> bfd/elf-properties.c.  The code preserves the different alignment
> requirements for 32-bit and 64-bit ELF formats, although the rationale for
> this distinction remains unclear in the original source and patch history.
> ---
>  bfd/elf-attrs.c | 33 +++++++++++++++++++++++++++++----
>  1 file changed, 29 insertions(+), 4 deletions(-)
> 
> diff --git a/bfd/elf-attrs.c b/bfd/elf-attrs.c
> index c34d4b21daf..f432412e45e 100644
> --- a/bfd/elf-attrs.c
> +++ b/bfd/elf-attrs.c
> @@ -812,10 +812,35 @@ static asection *
>  create_object_attributes_section (struct bfd_link_info *info,
>  				  bfd *abfd)
>  {
> -  (void) info;
> -  (void) abfd;
> -  /* TO IMPLEMENT */
> -  return NULL;
> +  asection *sec;
> +  const char *sec_name = get_elf_backend_data (abfd)->obj_attrs_section;
> +  sec = bfd_make_section_with_flags (abfd,
> +				     sec_name,
> +				     (SEC_READONLY
> +				      | SEC_HAS_CONTENTS
> +				      | SEC_DATA));
> +  if (sec == NULL)
> +    info->callbacks->fatal (_("%P: failed to create %s section\n"), sec_name);
> +
> +  /* It is unclear why the format requires a different alignment for
> +     32- vs 64-bit ELF.  This code was copy-pasted from
> +     _bfd_elf_link_create_gnu_property_sec() in bfd/elf-properties.c.  The
> +     original patch which introduced it, does not provide any reason and no
> +     discussion during the patch review focussed on this part of the code.
> +     https://inbox.sourceware.org/binutils/20210624132411.1993105-1-hjl.tools
> +     @gmail.com/  */
> +  unsigned align
> +    = (get_elf_backend_data (info->output_bfd)->s->elfclass == ELFCLASS64
> +       ? 3
> +       : 2);

H.J., can you please shed some light on the situation? I would much prefer if
the anomaly there wouldn't be duplicated here.

Thanks, Jan

> +  if (!bfd_set_section_alignment (sec, align))
> +    info->callbacks->fatal (_("%pA: failed to align section\n"), sec);
> +
> +  elf_section_type (sec) = get_elf_backend_data (abfd)->obj_attrs_section_type;
> +
> +  bfd_set_section_size (sec, bfd_elf_obj_attr_size (abfd));
> +
> +  return sec;
>  }
>  
>  /* Translate GNU properties that have object attributes v2 equivalents.  */