Re: [PATCH 3/5] gas: Default to ELF for RTEMS targets

[Sebastian Huber <sebastian dot huber at embedded-brains dot de>]

On 19/01/17 23:59, Joel Sherrill wrote:
> On 1/19/2017 4:55 PM, Alan Modra wrote:
> > On Thu, Jan 19, 2017 at 09:39:39AM -0600, Joel Sherrill wrote:
> > > One of the reasons this was duplicated was to ensure
> > > that when someone tinkered with a target they knew
> > > XXX-rtems was impacted. With this patch, there is
> > > no record of each cpu-rtems being a used target.
> > >
> > > I know binutils doesn't remove targets often but this
> > > makes the individual RTEMS targets invisible.
> > >
> > > It definitely means we likely don't have to touch
> > > these files for new targets but it also means
> > > we aren't obvious when deprecation discussions
> > > occur.
> > >
> > > It would be nice to have a binutils maintainer wade
> > > in. I can see how this is a nice clean up patch but
> > > it loses information.  What are the thoughts on this?
> >
> > I would prefer to lose that information.  We're talking about bfd/
> > here, and given that there is no difference between bfd support for
> > <target>-elf and <target>-rtems, it doesn't make much sense to
> > single out rtems targets in config.bfd.  The only concern would be if
> > a future change removed support for <target>-elf but you wanted to
> > keep <target>-rtems.  However, I can't see us deprecating an ELF
> > target unless all support for <target> is no longer wanted.
>
> That's the second opinion I wanted. Just wanted to
> confirm that someone with a broader binutils thought
> the information loss would be OK.
>
> I am OK with it. 

Thanks for your review. One side-effect of this patch set is support for 
RISC-V and 64-bit PowerPC for RTEMS.

Would someone mind committing this and

https://sourceware.org/ml/binutils/2017-01/msg00318.html

for me?

--
Sebastian Huber, embedded brains GmbH

Address : Dornierstr. 4, D-82178 Puchheim, Germany
Phone   : +49 89 189 47 41-16
Fax     : +49 89 189 47 41-09
E-Mail  : sebastian.huber@embedded-brains.de
PGP     : Public key available on request.

Diese Nachricht ist keine geschäftliche Mitteilung im Sinne des EHUG.