Hi Le mer. 10 oct. 2018 à 09:41, Felipe Balbi <felipe.balbi@xxxxxxxxxxxxxxx> a écrit : > > first of all, thanks for following the documentation when reporting a > bug, much appreciated. That's the least I can do. Thanks for take into account my issue. > Thomas JOURDAN <thomas.jourdan@xxxxxxxxx> writes: > > On the UP2 board, it is not possible in device mode to get the USB3.0 > > OTG port working at superspeed. It is unclear yet if it is a bug in > > the Linux DWC3 driver or the board itself. > > > > I have tried many cables, different host computers (Windows or Linux), > > and the connection always links at hi-speed. In host mode, it works > > fine at super speed. I tried many kernels, from upboard official one > > to latest RC version ans the problem persists. > > So, it doesn't work with UP's original kernel either? Then I'm pretty > sure either the board doesn't support super speed on the micro-b > connector, or the board is faulty. I have tested on 4 different boards and all exposes the same issue. The micro-b USB connector is the USB3.0 OTG port, and super speed works fine only in host mode. In parallel I will try to have a look at the hardware level to see if there is anything I can do to improve the signal quality. > > DCFG = 0x004e0814 > > maximum speed is set to super speed > > > [ 74.055170] g_ether gadget: high-speed config #2: RNDIS > > but it enumerates in high speed. Yes and that's also what reports the controller (DSTS = 0x00839fa8) > From your trace.txt, I don't see the early enumeration phase, so I can't > comment one what happened during that time. Care to rerun this so that > we have early enumeration? I need to see a Reset Event on the trace as > that's the first event when you first plug your usb cable. I don't know why it isn't there because I plugged the cable *after* I enabled the tracing. So I ran it a second time and despite the traces are smaller, there seem to be what you expect (messages for U0 state ?) Also I can enable the debug messages for the USB gadget layer if needed.
