We are looking (again) at developing a DisplayPort 1.2 Input + USB 3.0 Input to USB-C (DFP) adapter. The CCG3 Host reference design is more or less the design we are looking at, but we had one key question:
If we want to allow Pin Assignment D (2 lanes DP and 2 lanes USB 3), will that work when a upstream laptop is connected via legacy DisplayPort 1.2 cable? Will that laptop get information via AUX that it should provide only 2 lanes DP? Since the laptop will not be part of the PD signalling regarding C or D pin assignment.
What we "fear" is that the laptop will always default to 4 lanes DP on the DP output and not be informed about the need to only deliver 2 lane DP on lane 0 and 1.
Since Cypress host reference design has a DP 1.2 input and it apperently works when connecting a laptop via a DP cable to it, maybe it is not a problem?
If this product is going to integrate into Notebook/Laptop board directly, there is no problem. Since CCGx could have communication with EC to get more details information about the DP source selection.
Typically, The source side will offer more pin assignments options and select the best one as per attached device. If you are a dongle only, and you specified to 2 lane DP and USB3.0, the UFP_D may not support it. This will be a compliance issue.
Actually, CCG3 C to DP cable is bi-direction solution already, it is cover your case. Am I correct?
It is ok that the dongle support both C and D assignments so ensure compliance. Our worry is that the laptop with the DP source will not be informed correctly that is should serve 2 or 4 lanes DP since CCGx in this case will not be able to communicate with the laptops EC/GPU.
CCG3 C to DP doesn't support USB 3 right? And our adapter needs a Type-C receptacle, not a plug so we need to mux for that reason as well.
The question is if the laptop (DP Source) via the AUX lane via the DP cable will get the correct information to serve 2 or 4 lanes.