bsd-user: Move per-cpu code into target_arch_cpu.h

Move cpu_loop() into target_cpu_loop(), and put that in
target_arch_cpu.h for each architecture.

Signed-off-by: Stacey Son <sson@FreeBSD.org>
Signed-off-by: Warner Losh <imp@bsdimp.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>