Backports to 6 (and 7, and 5)

[Jakub Jelinek]

On Tue, Jun 27, 2017 at 11:02:37AM -0500, Segher Boessenkool wrote:
> On Tue, Jun 27, 2017 at 09:18:07AM +0200, Richard Biener wrote:
> > On Mon, 26 Jun 2017, Segher Boessenkool wrote:
> > > https://gcc.gnu.org/ml/gcc-patches/2016-11/msg01853.html
> > > https://gcc.gnu.org/ml/gcc-patches/2016-11/msg01923.html
> > > https://gcc.gnu.org/ml/gcc-patches/2016-11/msg02048.html
> > > https://gcc.gnu.org/ml/gcc-patches/2016-11/msg02606.html
> > > bb-reorder: Improve compgotos pass (PR71785)
> > 
> > It's not clear this fixes a regression and as it is
> > a missed-optimization I'd not backport it at this point in time
> > (I understand it's in GCC 7 already).
> 
> It's a regression from 3.x, and a pretty severe one, but we can live
> with it for a bit longer, it's a bit invasive for a backport.

It is just a missed-optimizations, not wrong-code, and the regression is
pretty old, so I think it doesn't hurt it if it stays fixed only in 7+.

	Jakub