Hi Sir,
We found connection issues while we connect 65988 in loopback mode(Port 1 connects to it Port2). We tested it in 3 scenarios:
(1) Port 1 – DRP, Port 2 – DRP: Fail with 2 symptoms
(2) Port 1 – DFP, Port 2 – UFP: Same as above
(3) Port 1 – UFP, Port 2 – DFP: Pass
In case (1) and case (2) above it failed in 2 symptoms, one is no connection always (always means that wait > 1.5mins) and one is connection built at a very late stage. Meaning that results are inconsistent.
Case 1 (Refer to DRP-DRP-FinalConnect.ccgx attached) starts at 401 Seconds. However, the connection keeps failing until 455 Seconds. It takes 55 seconds to build a valid connection.
While looks into the Type-C state we could see that the initial UFP (Port 2 in this case since it implemented Try.SNK and Port1 implement Try.SRC) side removes the Rd firstly (VBUS removal after). In this case, we could notice the UFP side is doing the removal rather than DFP.
(Zoom in look): I checked every disconnect is caused by Rd removal (No VBUS removal firstly in this case). And I see here’s a Try.SNK state at first Type-C connection. This shall be incorrect since we shall implement no Try on both sides and DPTX shall be DFP/Source preference.
Case2 below shows the same that Port 2 removes the Rd firstly:
We also tried to reset the PDC in the middle but it doesn't help the connection.
One more piece of information we could supplement is Port 2 0x30 register bytes 2-9 show all 0 even it indicates it receives two valid PDOs from Port1(Value 0x2 in the first byte).
Any suggestion of how to improve this will be welcome.
