DECLSPEC void xts_mul2 (u32 *in, u32 *out)
{
  const u32 c = in[3] >> 31;

  out[3] = (in[3] << 1) | (in[2] >> 31);
  out[2] = (in[2] << 1) | (in[1] >> 31);
  out[1] = (in[1] << 1) | (in[0] >> 31);
  out[0] = (in[0] << 1);

  out[0] ^= c * 0x87;
}