The client end of the handshake may receive payload data after "server finished" message

[mscottmueller]

Because of receive segment coalescing and other TCP/IP behaviors (e.g. Nagle's algorithm), it is possible for some data payload to be present in the extra data buffers on the client end after the handshake is completed. We need a mechanism to detect and transfer this over to the input buffers for the subsequent receive cycle.

[laudrup]

This indeed sounds familiar and like something I've tried to handle but I must admit I can't remember the details.

Is this something you would be willing to look into or should I do that? In the latter case I would very much appreciate if you could explain this a bit more in details.

Thanks.