[PATCH v2] x86_64/gcc: Add gcc patch for LIB_SPEC (rtems.h in gcc)

Mon May 7 17:58:39 UTC 2018

On Mon, May 7, 2018 at 12:16 PM, Amaan Cheval <amaan.cheval at gmail.com>
wrote:

> I'm not quite sure how GCC's releases work; it seems like GCC 8.1 has
> already been released:
> https://gcc.gnu.org/releases.html
> https://ftp.gnu.org/gnu/gcc/gcc-8.1.0/
>
> However, the "gcc-7-branch" is still being pushed to:
> https://gcc.gnu.org/git/?p=gcc.git;a=shortlog;h=refs/heads/gcc-7-branch
>
> What would pushing to that branch entail? Is it simply to make
> applying the patch to a 7.3 release easier?
>
> I only meant to make the patch to the RSB because:
> - I'm not sure when the RSB will move to using GCC 8.1+ (we still use 7.3)
> - It's better to have the RSB build the tooling the same for my dev
> environment as the community will have, without needing manual patches
> for building the tools
>
> I'm okay to just make a note about the patch if we'd rather not have
> it in the RSB in case by the end of the port, GCC has released a
> version including the patches which we can simply switch to using.
>
> Let me know if you prefer the latter!
>

Unless gcc makes a release from the branch, we will have to continue
to have a patch.

Chris is on holiday. Let's see what he wants done when he gets back.

>
> On Mon, May 7, 2018 at 10:28 PM, Joel Sherrill <joel at rtems.org> wrote:
> >
> >
> > On Mon, May 7, 2018 at 11:56 AM, Amaan Cheval <amaan.cheval at gmail.com>
> > wrote:
> >>
> >> Thanks a ton! I'll likely make the corresponding RSB patch to backport
> >> this to GCC 7.3 tomorrow for Chris to review :)
> >
> >
> > If need be, we can address pushing it to open gcc branches. I didn't see
> it
> > as critical as this was a new port but am happy to do it.
> >
> > If it goes on the 7 branch, it needs to go on the 8 branch though.
> >
> >>
> >>
> >> On Mon, May 7, 2018 at 10:03 PM, Joel Sherrill <joel at rtems.org> wrote:
> >> >
> >> >
> >> > On Fri, May 4, 2018 at 4:25 PM, Joel Sherrill <joel at rtems.org> wrote:
> >> >>
> >> >>
> >> >>
> >> >> On Fri, May 4, 2018 at 4:58 AM, Amaan Cheval <amaan.cheval at gmail.com
> >
> >> >> wrote:
> >> >>>
> >> >>> On Fri, May 4, 2018 at 7:54 AM Joel Sherrill <joel at rtems.org>
> wrote:
> >> >>>
> >> >>> > I am now confused. Do I need to get a patch into GCC?
> >> >>>
> >> >>> Hey!
> >> >>>
> >> >>> Sorry about the confusion. Technically the answer is yes, but it's
> >> >>> unrelated to _this_ patch.
> >> >>>
> >> >>> To clear things up:
> >> >>>
> >> >>> - [DONE] You already committed a previous patch for GCC for -qrtems,
> >> >>> -qnolinkcmds, etc.
> >> >>>
> >> >>>
> >> >>>
> >> >>> https://gcc.gnu.org/git/?p=gcc.git;a=commitdiff;h=
> 602fa1e9d3ea5e87d4d6e17e3e91fc2647e42da3
> >> >>>
> >> >>> - [DONE] This email thread backports the above patch to
> >> >>> rtems-source-builder so that we can use it in GCC 7.3. Chris
> accepted
> >> >>> this
> >> >>> patch into rtems-source-builder here
> >> >>>
> >> >>>
> >> >>>
> >> >>> https://git.rtems.org/rtems-source-builder/commit/?id=
> defa958301215995b0fa41d8e65cb23c9a28a847
> >> >>>
> >> >>> - [PENDING - Joel] I recently made a GCC patch to have GCC build
> >> >>> crti.o
> >> >>> and
> >> >>> crtn.o for the x86-64 RTEMS tools, so we can have symbols like
> "_fini"
> >> >>> added automatically through our bsp_specs file, the way we do for
> >> >>> other
> >> >>> architectures. As Sebastian mentioned in an email to you and me, I
> >> >>> believe
> >> >>> he'd prefer if you reviewed the patch here:
> >> >>>    https://gcc.gnu.org/ml/gcc-patches/2018-05/msg00126.html
> >> >>>
> >> >>> - [PENDING - Amaan, and later Chris] Once the crti.o patch above is
> >> >>> merged
> >> >>> into GCC, I'll make a corresponding backport patch to
> >> >>> rtems-source-builder
> >> >>> to make that available to GCC 7.3 as well, and then Chris can
> review +
> >> >>> merge that in after.
> >> >>>
> >> >> I meant to deal with this today.
> >> >>
> >> >> I am leaving myself a note to get this merged next week. Ping me in
> the
> >> >> middle
> >> >> of next week if you haven't heard action.
> >> >
> >> >
> >> > This is now committed on the gcc master.
> >> >
> >> > https://gcc.gnu.org/ml/gcc-cvs/2018-05/msg00226.html
> >> >
> >> > Hopefully that helps clean up the RSB.
> >> >
> >> >>
> >> >>
> >> >>>
> >> >>> Hope that clears things up!
> >> >>
> >> >>
> >> >> It does.
> >> >>>
> >> >>>
> >> >>> > On Thu, May 3, 2018, 8:38 PM Chris Johns <chrisj at rtems.org>
> wrote:
> >> >>>
> >> >>> >> On 07/04/2018 08:35, Amaan Cheval wrote:
> >> >>> >> > Adds -qnolinkcmds, -qrtems, and defaults to pass to linker, for
> >> >>> >> > gcc targeting x86_64.
> >> >>>
> >> >>> >> Pushed, thank you.
> >> >>>
> >> >>> >> Chris
> >> >>> >> _______________________________________________
> >> >>> >> devel mailing list
> >> >>> >> devel at rtems.org
> >> >>> >> http://lists.rtems.org/mailman/listinfo/devel
> >> >>>
> >> >>> > _______________________________________________
> >> >>> > devel mailing list
> >> >>> > devel at rtems.org
> >> >>> > http://lists.rtems.org/mailman/listinfo/devel
> >> >>
> >> >>
> >> >
> >
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/devel/attachments/20180507/408f0d1d/attachment-0002.html>