I don't understand why the follow code is written like this

[Huoke]

I don't understand why the follow code is written like this.
According to the MD5 algorithm, it is filled with first 1 and 0 at last. Why do we need this step?
` for (offset = initial_len + 1; offset < new_len; offset++)
msg[offset] = 0;

// append the len in bits at the end of the buffer.
to_bytes(initial_len*8, msg + new_len);

// initial_len>>29 == initial_len*8>>32, but avoids overflow.
to_bytes(initial_len>>29, msg + new_len + 4);`

[JialongWang1224]

in the above:
msg = (uint8_t*)malloc(new_len + 8); memcpy(msg, initial_msg, initial_len);``
so the 0 to initial of msg is the true msg, the initial to new_len should be the padding, so we need to msg[initial_len]=0x80 and the msg[offset]=0