[Bug tree-optimization/82847] [8 regression] gcc.dg/vect/slp-perm-9.c fail

[rguenther at suse dot de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82847

--- Comment #6 from rguenther at suse dot de <rguenther at suse dot de> ---
On Wed, 22 Nov 2017, rsandifo at gcc dot gnu.org wrote:

> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82847
> 
> --- Comment #5 from rsandifo at gcc dot gnu.org <rsandifo at gcc dot gnu.org> ---
> (In reply to rguenther@suse.de from comment #4)
> > On Tue, 21 Nov 2017, rsandifo at gcc dot gnu.org wrote:
> > 
> > > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82847
> > > 
> > > rsandifo at gcc dot gnu.org <rsandifo at gcc dot gnu.org> changed:
> > > 
> > >            What    |Removed                     |Added
> > > ----------------------------------------------------------------------------
> > >                  CC|                            |rsandifo at gcc dot gnu.org
> > > 
> > > --- Comment #3 from rsandifo at gcc dot gnu.org <rsandifo at gcc dot gnu.org> ---
> > > Isn't the problem here that vect_perm_byte and vect_perm_short don't return
> > > true for x86_64, even though the patterns seem to be there in sse.md?
> > 
> > vect_perm_byte is only available with SSEn+
> > 
> > The test only uses vect_perm_short though.
> 
> Sure, fixing only vect_perm_short would be enough for this PR.  I just
> meant that the same problem seems to apply to vect_perm_byte too.

We mostly use the "generic" vect_perm ... or maybe we used to do that.