https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67818
--- Comment #3 from Mikael Morin <mikael at gcc dot gnu.org> --- memcpy was fixed by Andre at r222477 on trunk. I think it fixes the regression, but I'm a little uncomfortable with backporting it, because of its medium size. So the question is whether r222477 should be backported, or r228361 reverted.
