Hi,
We are experiencing an issue with detecting the USB line state through the USB3320 PHY.
It appears that the AM37xx is NOT receiving the 0004 PCD interrupt to detect the line state, but the PHY does send a RXCMD with 01b as the line state.
This is same processor and reference design used from the BeagleBoard xM Rev B. The only board difference is that the transceiver is attached to a downstream hub instead of the Ethernet combo chip.
The same code works properly on the Beagle xM, but not in our custom design.
The only dmesg log difference is shown below where the Beagle xM registers the interrupt as "irq status 0004 PCD" then detects the device, but our board does not register the interrupt or detect the downstream hub.
Here is a low-res plot showing the power-on sequence. Triggered at RBIAS going high after power-on. The plots shows DIR is properly asserted during the reset sequence until CLKOUT starts.
After DIR falls, we zoom in to grab a high-res plot. This is triggered on the falling edge of DIR. That plot is shown below.
As expected, we see DIR first fall, then rise again to begin transmitting the first RXCMD with line state information. At the rising edge of the REFCLK, the line state on BUS1 is shown as 01, which is what I would expect for hi-speed hub attached.
Note: STP was never asserted again after the REFCLK was started.
What could be wrong with the AM37xx initialization sequence preventing the interrupt from happening on the AM37xx side?
Thanks in advance,
