Re: [PATCH, rs6000] Fix ELFv2 homogeneous float aggregate ABI bug

["Ulrich Weigand" <uweigand at de dot ibm dot com>]

Jakub Jelinek wrote:
> On Wed, Jul 09, 2014 at 02:19:36PM -0400, David Edelsohn wrote:
> > > This is an ABI change for the affected corner cases of the ELFv2
> > > ABI.  However, those cases should be extremely rare; the full
> > > compat.exe and struct-layout-1.exp ABI compatibility test suite
> > > passed, with the exception of two tests specifically intended
> > > to test multiple homogeneous float aggregates.
> > >
> > > Tested on powerpc64le-linux.
> > >
> > > OK for mainline?
> > > [ The patch should then also go into the 4.8 and 4.9 branches for
> > > consistency. ]
> > 
> > Okay everywhere.
> 
> IMHO it should only go to the trunk, same as
> https://gcc.gnu.org/ml/gcc-patches/2014-07/msg00540.html (see also related
> PRs, has been discussed on IRC too).
> 
> If it is a corner case, the better, fewer people will be affected when
> transitioning from 4.9.x to 5.0 (or 4.10?).  But 4.9.0 is for a few months
> in the wild already, I don't think ABI changes are something people would
> like to see between 4.9.0 and 4.9.1 or any other patchlevel revision,
> most people aren't prepared to rebuild the world just because they have
> upgraded to a newer compiler snapshot or new patchlevel version.

Richard Biener wrote:
> Can you add -Wpsabi warnings for all the issues you found?  That way
> a world re-build will at least show if anything important is affected.

OK, to summarize the off-line discussions on these concerns, what I'm
now planning to do is the following:

- Add -Wpsabi warnings for all cases changed to the mainline patch
- Backport *only* the warnings, but not the actual change in behavior
  to the 4.8 and 4.9 branches.

This should match existing precent of how to handle ABI changes.

Doing a world-rebuild with the warnings enabled would certainly be
extremely useful feedback ...

Thanks,
Ulrich

-- 
  Dr. Ulrich Weigand
  GNU/Linux compilers and toolchain
  Ulrich.Weigand@de.ibm.com