Binutils causes abnormally large, non-working binaries since 2.43

[Bastiaan Timmer]

On 07-03-2025 12:25, Jan Beulich wrote:
> That commit touching objcopy.c only (as to what's relevant for the writing
> of object files), there must be a use of objcopy or strip somewhere. The
> build logs you provided, however, don't have any indication thereof. Gcc
> would use objcopy when passed -gsplit-dwarf, but that doesn't even look to
> be potentially applicable here (iirc split debug info is only an option
> with ELF).
>
> Further I tried both Cygwin and MinGW objcopy on a random (small) object,
> without passing any --section-alignment option of course. That didn't
> change the object at all. If I pass --section-alignment, I can see the
> provided value taking effect. I wonder whether that's really intended for
> COFF objects; it looks to me as if it's solely intended for COFF/PE
> binaries.
>
> Would you mind trying with the simple patch below? If that makes a
> difference, we'd then still need to understand how objcopy (or strip)
> comes into play here in the first place. Depending on where it's used and
> how, this or another change may be warranted.
>
> Jan
>
> --- a/binutils/objcopy.c
> +++ b/binutils/objcopy.c
> @@ -4301,7 +4301,8 @@ setup_section (bfd *ibfd, sec_ptr isecti
>       alignment = p->alignment;
>     else if (pe_section_alignment != (bfd_vma) -1
>   	   && bfd_get_flavour (ibfd) == bfd_target_coff_flavour
> -	   && bfd_get_flavour (obfd) == bfd_target_coff_flavour)
> +	   && bfd_get_flavour (obfd) == bfd_target_coff_flavour
> +	   && bfd_pei_p (obfd))
>       {
>         alignment = power_of_two (pe_section_alignment);
>   

Thank you for taking the time to look into this further.

Looking at the PKGBUILD (a small configuration file, that is used on 
Arch Linux to build packages), I see the packaging-step runs:

      find "${pkgdir}/usr/${_arch}" -name '*.exe' -exec ${_arch}-strip {} \;
      find "${pkgdir}/usr/${_arch}" -name '*.dll' -exec ${_arch}-strip 
--strip-unneeded {} \;
      find "${pkgdir}/usr/${_arch}" -name '*.a' -o -name '*.dll' | xargs 
${_arch}-strip -g

Where `${_arch}-strip' would expand to 'x86_64-w64-mingw32-strip' (or 
similar for i686). Would that explain the apparent use of objcopy? For 
reference, the PKGBUILD can be found here: 
https://aur.archlinux.org/cgit/aur.git/plain/PKGBUILD?h=mingw-w64-openssl

I have built two versions of mingw-w64-openssl, one with the upstream 
2.44 version of binutils (I had only tried up until 2.43 before), and 
one with version 2.44 with your patch applied. I am happy to report your 
tiny patch also appears to solve the issues. The size of the openssl 
libraries is back to normal when compiled with 2.44_patched, and 
binaries linked to these openssl libraries successfully run on Windows.

I have opened a bug about this issue about two weeks ago (here: 
https://sourceware.org/bugzilla/show_bug.cgi?id=32732). I assume this 
info is useful there, maybe you want to chime in? Or, of course, I'd be 
happy to report these latest findings myself if you prefer.

Thanks again!

Bas