Issue on the server with rejected requests and timed out requests

[Eastwo00d]

Hi,

I think there is a problem concerning the management of refused tcp requests. When I start a modbus server on a specific IP and a specific Unit Identifier, if I use several MODBUS clients to request the same IP but on other unit IDs (that are not handled by the server), the CPU usage skyrockets till the server is not able to respond to requests he is supposed to...

Also, when I am debugging my program, if I stay long enough in a breakpoint so the tcp timeout is reached, the server refuses all the requests after resuming the program. Stopping and restarting the server solves this problem.

Regards

[Eastwo00d]

I spotted the problem : in both ModbusTcpRequestHandler and ModbusRtuRequestHandler, the TryReceiveRequestAsync method returns 'false' if the request is not addressed to the server, but in the case we are in asynchronous mode, no response will be sent and till the request timout is reached the thread is stuck !

I propose to trigger a cancel token in addition to return false in that case so the connection will be closed as it is not handled by the server. It will solve the problem