Hi, Thanks for your reply.
> Currently there is not so much to try, but you can start poking in > sys/dev/usb/controller/xhci*.[ch] . I have not made any USB-C specific > changes to the XHCI driver in FreeBSD, so yes, it might not work. > > There are also some debug knobs: sysctl -a | grep xhci USB-C here is a red herring, there is a i2c controlled mux involved before it will work for USB data. I think we are missing a second usb instance (instances? I am not sure of the terminology) on the xhci controller. Looking at Linux and NetBSD, they both launch two USB from their xhci_pci drivers. Linux creates the first here: https://github.com/torvalds/linux/blob/master/drivers/usb/host/xhci-pci.c#L326 the second here https://github.com/torvalds/linux/blob/master/drivers/usb/host/xhci-pci.c#L334 NetBSD manages two instances of bus in the softc, sc_bus and sc_bus2 sc_bus2 is configured here, the same as the first bus but with an explicit usb revision set to USBREV_2_0. They share the same dmatag. https://github.com/NetBSD/src/blob/trunk/sys/dev/usb/xhci.c#L929 Both instances are then registered with the usb subsystem here: https://github.com/NetBSD/src/blob/trunk/sys/dev/pci/xhci_pci.c#L281 I seem a similar pattern in the tegra xhci drivers for NetBSD and Linux. I am going to have a dive into the intel xHCI Interface docs to get more background. https://www.intel.com/content/dam/www/public/us/en/documents/technical-specifications/extensible-host-controler-interface-usb-xhci.pdf Could you tell me if this multiple usb instance interface is something we are missing or something we implement in a different manner? - Tom _______________________________________________ freebsd-usb@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-usb To unsubscribe, send any mail to "freebsd-usb-unsubscr...@freebsd.org"
