https://gcc.gnu.org/bugzilla/show_bug.cgi?id=117800
--- Comment #6 from Andrew Pinski <pinskia at gcc dot gnu.org> --- (In reply to Andrew Pinski from comment #4) > Yes even fx comment > https://github.com/technion/libscrypt/issues/60#issuecomment-1895762926 > > Almost points out the aliasing issue. He said replace Blkcpy with memcpy the > issue goes away. Yes because that is where the alaising issue is located . He even mentioned the salsa20_8 function which is exactly where the aliasing issue shows up.
