In applications where we have virtual sequence to co-ordinate the stimulus generation, what is the recommended way to pass the response from the driver back to virtual sequence?
In reply to yourcheers:
Responses are always related to seq_items. Because a virtual sequence does not generate any seq_item it is useless to have this. But you can send back responses to the local sequences called in the virtual sequence.
In reply to chr_sue:
For example, in case of Read-modify-write operation to a hypothetical design, where Read interface is a separate agent & write interface is separate agent. Where the sequence coordination is happening in virtual sequence and the Read data should be seen by the Write seq, and we dont get the response back to virtual-sequence. In such cases, what is the recommended UVM env structure?
In reply to yourcheers:
Could you please show with a piece of (pseudo) code how you are controlling the read and write sequences?
BTW you have in your virtual sequence the handles of the read and write sequences. This allows you to see also the responses of these local sequences.