word_out
# define word_out(x,c,v) { ((uint8_t*)(x)+4*c)[0] = bval(v,0); ((uint8_t*)(x)+4*c)[1] = bval(v,1); \
# define word_out(x,c,v) (*((uint32_t*)(x)+(c)) = (v))
# define word_out(x,c,v) (*((uint32_t*)(x)+(c)) = aes_sw32(v))