This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: [PATCH] allow easier overriding of ELF_DYNAMIC_INTERPRETER
- From: Hans-Peter Nilsson <hp at bitrange dot com>
- To: Ian Lance Taylor <ian at airs dot com>
- Cc: David O'Brien <obrien at FreeBSD dot org>, Nick Clifton <nickc at cambridge dot redhat dot com>, GNU Binutils mailing list <binutils at sources dot redhat dot com>
- Date: Sun, 17 Feb 2002 12:17:11 -0500 (EST)
- Subject: Re: [PATCH] allow easier overriding of ELF_DYNAMIC_INTERPRETER
On 17 Feb 2002, Ian Lance Taylor wrote:
> Hans-Peter Nilsson <hp@bitrange.com> writes:
>
> > I can't think of a reason where you'd have different
> > ELF_DYNAMIC_INTERPRETER values for one and the same bfd vector.
>
> There are many systems which can use the elf32-i386 BFD vector, but
> which use different dynamic interpreters.
Let me rephrase: I can't think of a reason to define one and the
same named bfd vector differently, by defining items such as
ELF_DYNAMIC_INTERPRETER differently (when the effect can be
reached by defining different bfd vectors). I see I didn't
write it that way, though. :-]
> I think that the difference between two linkers which both use
> elf32-i386 but which should use different dynamic interpreters is
> exactly the sort of thing which linker emulations were designed to
> handle. Therefore, I think that any non-default dynamic interpreter
> should be specified in the linker emulation.
>
> In fact, for standard ELF targets, I see that this is already possible
> using the ELF_INTERPRETER_NAME variable in the ld/emulparams file.
Great, I totally agree. Is that ok for everyone; no bfd changes
but instead setting ELF_INTERPRETER_NAME in ld/emulparams/* when
it needs to be different from the one in the ELF ABI supplement?
brgds, H-P