(Inconsistency) Violation of the requirement of Unit Identifier.

[ProtocolSpecCheck]

Hi,

I would like to bring your attention a violation between the protocol implementation and the specification regarding the MODBUS/TCP Unit Identifier requirement written in the Modbus Application Protocol Specification:

Requirements in the specification:

For Modbus TCP, the value of Unit Identifier must be defined as 0xFF or 0x00.

How does libmodbus violate this requirement?

Libmodbus lacks validation for the Unit identifier value, accepting request with invalid Unit Identifier value.

• Terminal-1(server): ./tests/bandwidth-server-many-up
• Terminal-2(client): send the request with invalid Unit Identifier value, send b'\x00\x01\x00\x00\x00\x06\x0a\x0f\x00\x00\x00\x01\x01\x01'.
• Server behaviour: Regardless of the value of the Unit Identifier, the server will ignore the check and treat it as valid, recv b'\x00\x01\x00\x00\x00\x06\x0a\x0f\x00\x00\x00\x01'.

[karlp]

Don't do this, it's unnecessary restriction on a field that is used realworld for relaying serial over tcp, the field is never used normally, so there's absolutely zero risk of any "illegal parsing"

Further, it's absolutely insane to add a printf for this into the library code. Please never merge this.

[ProtocolSpecCheck]

Hi @karlp,

Thank you for your feedback regarding the Unit Identifier value in the Modbus TCP implementation.
We will modify our PR and resubmit it.

While we understand that in some practical scenarios the Unit Identifier may be ignored, ensuring compliance with the specification contributes to consistency and clarity in the protocol.

Implementing validation for this field, can:

• Provide a safeguard against potential misuse.
• Ensure that the protocol implementation adheres to established standards.
• Prevent confusion for users who expect the implementation to reflect the official specifications, particularly in diverse application environments.

We believe this small adjustment can enhance the robustness of the protocol’s implementation without significantly impacting its functionality. We’d value your thoughts on this perspective and are open to further discussion.

[mhei]

I second @karlp 's assessment. As long as we don't have a usable mechanism to activate and/or disable a "strict standard compliant mode", this check should not be implemented.