[Bug binutils/19938] objcopy breaks sh_link and sh_info of SHT_SUNW_LDYNSYM and SHT_SUNW_symsort sections

[thopre01 at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=19938

--- Comment #22 from Thomas Preud'homme <thopre01 at gcc dot gnu.org> ---
(In reply to Nick Clifton from comment #21)
> Created attachment 9222 [details]
> extended patch
> 
> Right, here is another revision of the patch.  In this version:
> 
>   * The EI_ABIVERSION field of the input file is copied to the output file.
> 
>   * The section header copying code now makes two passes over the input
>     sections, trying to find a match between an input section and an output
>     section.  The first pass looks for a direct mapping between the sections,
>     based upon the pointers stored in the structures.  (Thanks to Thomas for
>     suggesting this).  The second pass tries to match up sections based upon
>     their characteristics (address, type, flags, etc).  If both passes fail
>     the backend copying function is given one last chance to set the fields,
>     but with a NULL matching input section.
> 
>   * Updates readelf to report Solaris specific program header types.
> 
>   * General code tidying and cleanup.
> 
> Ali, Rainer. Thomas - please could you try out this version and let me know
> if
> you are happy with it.
> 
> One thing that this patch does not do is zero the p_paddr fields of the
> program
> headers for Solaris targets.  I did look at doing this, but it would a lot of
> hacking on very complicated, fragile code in the BFGD library.  Something
> that
> I think would be very likely to introduce new bugs.
> 
> Cheers
>   Nick

Hi Nick,

As for the previous version, the patch does fix the problem we have at hand.
I'm glad my suggestion was useful to you.

I skimmed over your patch as I was curious to learn about objcopy and I
realized that it is not working exactly as I thought. I had in mind that
setup_section () in binutils/objcopy.c would create a record input section
number -> output section number for each section it copy and that we could use
this to skip find_link altogether. Seeing how old is binutils I guess this must
be impossible for a reason I'm missing.

Thank you for such a quick fix.

Best regards,

Thomas

-- 
You are receiving this mail because:
You are on the CC list for the bug.