This is the mail archive of the
mailing list for the glibc project.
Re: [PATCH] Add ARM relocation constants to elf/elf.h
"Joseph S. Myers" <firstname.lastname@example.org> writes:
> On Wed, 23 Jan 2013, Petr Machata wrote:
>> Roland McGrath <email@example.com> writes:
>> > We strongly prefer to have a short comment on every reloc type #define.
>> I don't think I have enough expertise to concisely explain what every
>> relocation does. I can at least put in e.g. what instructions each
>> static relocation applies to, and add snippets from the standard where
>> available. Joseph, what is your preference? FWIW, few relocation
>> values in ARM block have any comments at all.
> I'm not clear there is a great deal of use in having all these relocations
> (that are irrelevant to ld.so) in elf.h; what are the expected users,
> given that binutils has its own relocation definitions? (Cf. how AArch64
> only lists a limited subset of relocations in elf.h, not all the static
> relocations that ld.so doesn't use.)
My impression always was that elf.h is essentially a C language
rendition of various Elf standards (sans algorithms).
What prompted me to post this patch in particular was that GCL folks
seem to be in a need of those defines . I'm not in contact with them
personally, this issue was brought to my attention via Fedora ARM list.
I offered composing a patch as it seemed an obvious omission.
> But if they are added, I think including comments is appropriate.
OK, I'll prepare a patch with comments.