If we could use zvbc with SEW32, we could get better performance for CRC32 algorithm.
We could still use the current SEW32 instructions, but we need the additional widening from SEW32 to 64 and masking the unused MSB parts.