vrolq_n_u32
#define vrolq_n_u32(x, n) (vshlq_n_u32(x, n) | vshrq_n_u32(x, 32 - (n)))
#define vrolq_n_u32(x, n) vsriq_n_u32(vshlq_n_u32(x, n), x, 32 - (n))