amd64: bring back asm bcmp, shared with memcmp

Turns out clang converts "memcmp(foo, bar, len) == 0" and similar to
bcmp calls.

Reviewed by: emaste (previous version), jhb (previous version)
Differential Revision: https://reviews.freebsd.org/D34673

(cherry picked from commit fbc002cb72d2d9bb435cce99630d5d7da9f59390)