This is the mail archive of the
mailing list for the binutils project.
Re: PING Re: [RFA] Linker script extension SECTION_FLAGS
- From: Nick Clifton <nickc at redhat dot com>
- To: Catherine Moore <clm at codesourcery dot com>
- Cc: binutils at sourceware dot org, Tristan Gingold <gingold at adacore dot com>
- Date: Tue, 28 Jun 2011 12:37:35 +0100
- Subject: Re: PING Re: [RFA] Linker script extension SECTION_FLAGS
- References: <4DD41EB0.email@example.com> <firstname.lastname@example.org> <4DDC37D2.email@example.com> <4DED4E64.firstname.lastname@example.org> <4DEE2390.email@example.com> <4E025F3A.firstname.lastname@example.org>
Thanks for the patch review. I've developed a new patch that addresses
the comments that both you and Tristan made regarding the original
patch. I've now associated the INPUT_SECTION_FLAGS with the input
section specifications instead of the output sectionas you and others
suggested. I've tested arm-coff, mips-elf and ppc-elf. What do you
think? Is this okay to commit?
I like this patch, but I am afraid that I cannot approve it yet. There
are still several problems that need to be addressed:
* You did not test building a toolchain configured with
"--enable-targets=all --enable-64-bit-bfd". There were several targets
that did not build because they did not define an appropriate
* The new linker tests fail for targets that do not support the
INPUT_SECTION_FLAGS feature. (Ie the ones that use
* The ld.patch file contained the new linker tests, so the
ld-test.patch file could not be applied afterwards.
* The code did not check for contradictory input section flags, eg:
INPUT_SECTION_FLAGS (SHF_TLS & !SHF_TLS)
* There was no mention of this new feature in the ld/NEWS file.
* You have not provided a patch for GOLD. I would like to keep the
two linkers in sync, at least as far as features go.
I hope that you will be willing to resolve these issues and submit a