Merge pull request #5703 from wentaoshang/fix-sha256 hashes: support unaligned memory access in be32enc_vect
