Import Arm Optimized Routines v21.02

This is the new replacement for the existing cortex-strings code which will
be replaced in a follow-up commit.
We should also be able to use some of the math functions to allow the
tests to pass on AArch64 (and other architectures) instead of just x86.
We might also be able to reuse some of the tests for the kyua testsuite.

Imported using

curl -L https://github.com/ARM-software/optimized-routines/tarball/e823e3abf5f89ecba58a10fc0fd82c13d9984b6b | tar --strip-components=1 -xvzf -
git add .

Differential Revision: https://reviews.freebsd.org/D29035
git-subtree-dir: contrib/arm-optimized-routines
git-subtree-mainline: e34c713b0e660a49b57fafc02c9eccb26b938220
git-subtree-split: f9f37c002ab5a580accfe26b731eef45e798b435