crc16_ccitt optimization in XMODEM protocol

[Cuda-Chen]

Hi community,

For checking the crc16_ccitt(), I think it uses a half-byte technique to achieve the balance between performance and space.
I have done optimization of CRC32 in sse2neon which uses a half-byte LUT that only uses 64 bytes.
For crc16_ccitt(), the LUT just needs 32 bytes, and I think it is feasible on many embedded systems, and I think I can create a PR to optimize this.

Before I begin, I would like to ask whether there are any concerns crc16_ccitt() uses only bit operations.

[fifieldt]

@jp-bennett - thoughts?

[jp-bennett]

I'll admit, I'm slightly out of my depth here. Generally speaking, we won't turn down a PR over bitwise operations. Make sure to add comments so someone else can try to understand what's going on.