On Tue, Feb 13, 2018 at 05:07:26PM -0600, Peter Bergner wrote: > On 2/13/18 4:34 PM, Segher Boessenkool wrote: > >> This patch passed bootstrap and retesting on powerpc64le-linux with > >> no regressions. Ok for mainline? > > > > Okay, thanks! Does this need backports? > > Committed with your suggested change below. Thanks! > > It'd be easy to backport and should be fairly harmless. That said, I was > never able to create a simpler test case, using __builtin_altivec_lvx() > that would end up being re-recog'd as vsx_movv4si_64, so the only way I > know we can hit this is with Kelvin's optimization that replaces aligned > vsx loads/stores with altivec loads/stores and that optimization is only > on trunk. > > It's up to you whether you want the backport because you don't trust > me being able to create a failing test case. :-)
We can backport without having a failing testcase. Let's do that for 7 at least? Segher