CASE 1: Two separate TPS65988DHRSHR chips with port 2 of first chip and port 1 of second chip setup as DisplayPort Sink (GPIO4 HPD input) and Source (GPIO3 HPD output), respectively.
CASE 2: Single TPS65988DHRSHR with port 2 and port 1 setup as DisplayPort Sink (GPIO4 HPD input) and Source (GPIO3 HPD output), respectively.
CASE 3: Single TPS65988DHRSHR with port 2 setup as DisplayPort Sink (GPIO4 HPD input) and single TPS65982ABZQZR setup as DisplayPort Source (GPIO3 HPD output).
CASE 4: Single TPS65982ABZQZR setup as DisplayPort Sink (GPIO4 HPD input) and Single TPS65988DHRSHR with port 1 setup as DisplayPort Source (GPIO3 HPD output).
In each case, the CC bus of the DisplayPort Sink is connected to the DisplayPort Source via USB Type-C in DisplayPort Alternate Mode, while we inject a 4 to 4.8 millisecond HPD pulse into GPIO4 HPD input and look for the same pulse to be echoed by the GPIO3 HPD output. We have found that, in cases 2, 3, & 4, everything works as expected and we see the 4 to 4.8 millisecond HPD pulse echoed at the GPIO3 HPD output. However, in case 1, the TPS65988DHRSHR chip setup as DisplayPort Sink (GPIO4 HPD input) is pulse width sensitive. As long as one does NOT send a HPD with pulse width of 4 to 4.8 millisecond, everything will continue to work as expected, but upon sending a HPD with pulse width 4 to 4.8 milliseconds, the DisplayPort Sink (GPIO4 HPD input) will "get stuck" and will not issue HPDs anymore. The only way to clear this condition is to either do a cycle power or issue a 'GAID' or 'Gaid' 4CC.
