CFG_steps for inputs

[clemensschiffer]

Some implementation will want to send a CFG_steps PDU not just for outputs but also for inputs. If this is desired the standard should explicitly allow for it. Currently all mentions of steps refer to outputs, such as on p40:

In PDU CFG_steps it defines the communication step size of an output for SRT and HRT operating modes

I think the idea is: In SRT a master configures an output of slave A to send data to slave B with steps=2. Slave A will send every second time step. Slave B has no way of knowing if this is intentional or if there is some slowdown. (It does however know that no PDUs are missed due to seq_id.) Sending CFG_steps to slave B during configuration could clarify this and help in error handling.

Change proposal: Either on page 40 or on page 47 under CFG_steps add something like:

Optionally a CFG_steps PDU may be send for inputs to a DCP slave to inform it of the configuration of the corresponding output.

Possibly also: This my also be done for the default value of steps=1.

[martinkrammer]

Options for not breaking the protocol (towards 1.0.1) are:

• Add error code for variable id not an output OR
• Allow it, make it optional OR
• do nothing

Options that are breaking the protocol:

• Require sending CFG_steps to input OR
• do nothing