Hi, I was trying out newer kernels to see if updates to the logitech hidpp drivers would fix a problem I was seeing with the force feedback command queue filling up and causing delays. I found that my G920 was no longer recognized as an HID++ device at all with kernel 5.2 or 5.2.1, resulting in a complete lack of force feedback functionality. I've run across two other reports of a similar issue on the steam community forums. I'm running the gentoo patched kernels nominally, but see the same behavior between gentoo's 5.2 and vanilla 5.2. after the usb_modeswitch command which results in a flip of the product id from c261 to c262, the 5.1.17 kernel (most recent working I've tested) shows me this in my kernel log: Jul 20 10:23:02 frisco kernel: usb 1-12: USB disconnect, device number 6 Jul 20 10:23:03 frisco kernel: usb 1-12: new full-speed USB device number 7 using xhci_hcd Jul 20 10:23:03 frisco kernel: usb 1-12: New USB device found, idVendor=046d, idProduct=c262, bcdDevice=96.01 Jul 20 10:23:03 frisco kernel: usb 1-12: New USB device strings: Mfr=1, Product=2, SerialNumber=3 Jul 20 10:23:03 frisco kernel: usb 1-12: Product: G920 Driving Force Racing Wheel for Xbox One Jul 20 10:23:03 frisco kernel: usb 1-12: Manufacturer: Logitech Jul 20 10:23:03 frisco kernel: usb 1-12: SerialNumber: 000010a4ade826d3 Jul 20 10:23:03 frisco kernel: input: Logitech G920 Driving Force Racing Wheel for Xbox One as /devices/pci0000:00/0000:00:14.0/usb1/1-12/1-12:1.0/0003:046D:C262.0004/input/input16 Jul 20 10:23:03 frisco kernel: logitech-hidpp-device 0003:046D:C262.0004: input,hiddev97,hidraw3: USB HID v1.11 Joystick [Logitech G920 Driving Force Racing Wheel for Xbox One] on usb-0000:00:14.0-12/input0 Jul 20 10:23:03 frisco kernel: logitech-hidpp-device 0003:046D:C262.0004: HID++ 4.2 device connected. Jul 20 10:23:03 frisco kernel: logitech-hidpp-device 0003:046D:C262.0004: Force feedback support loaded (firmware release 1). with 5.2: I get this (I believe this was after a few replugs, which is why the device numbers are so much higher, same end result every time though): Jul 20 09:45:34 frisco kernel: usb 1-12: USB disconnect, device number 12 Jul 20 09:45:34 frisco kernel: usb 1-12: new full-speed USB device number 13 using xhci_hcd Jul 20 09:45:35 frisco kernel: usb 1-12: New USB device found, idVendor=046d, idProduct=c262, bcdDevice=96.01 Jul 20 09:45:35 frisco kernel: usb 1-12: New USB device strings: Mfr=1, Product=2, SerialNumber=3 Jul 20 09:45:35 frisco kernel: usb 1-12: Product: G920 Driving Force Racing Wheel for Xbox One Jul 20 09:45:35 frisco kernel: usb 1-12: Manufacturer: Logitech Jul 20 09:45:35 frisco kernel: usb 1-12: SerialNumber: 000010a4ade826d3 Jul 20 09:45:35 frisco kernel: input: Logitech G920 Driving Force Racing Wheel for Xbox One as /devices/pci0000:00/0000:00:14.0/usb1/1-12/1-12:1.0/0003:046D:C262.0007/input/input19 Jul 20 09:45:35 frisco kernel: logitech-hidpp-device 0003:046D:C262.0007: input,hiddev97,hidraw3: USB HID v1.11 Joystick [Logitech G920 Driving Force Racing Wheel for Xbox One] on usb-0000:00:14.0-12/input0 with 5.2 I don't get those last two lines. and evtest doesn't show the "Event type 21 (EV_FF)" which 5.1.17 and earlier do. I configured each kernel I tested using olddefconfig on the .config from my original 4.19.57 kernel, with nothing else changed. I added some additional hid_warn's to try to track down where it was failing, and it looks like the hidpp_validate_report function is getting a 0 when testing REPORT_ID_HIDPP_SHORT. REPORT_ID_HIDPP_LONG does get the expected length though. Not really sure what I'm looking at here. Happy to provide any further details/testing. Regards, Peter
