Hi Alan, On Thu, Oct 8, 2020 at 4:42 PM Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> wrote: > > Jiri and Benjamin: > > Please look at Bugzilla # 209137: > > https://bugzilla.kernel.org/show_bug.cgi?id=209137 > > Somewhere between 5.0 and 5.4, the logitech-djreceiver driver stopped > binding to the user's Logitech wireless device. Here are relevant parts > of the kernel logs. From 5.0 (working): > > [ 2.857770] logitech-djreceiver 0003:046D:C52B.0003: hiddev0,hidraw0: USB HID v1.11 Device [Logitech USB Receiver] on usb-0000:00:14.0-4/input2 > [ 2.985339] input: Logitech Unifying Device. Wireless PID:1024 Mouse as /devices/pci0000:00/0000:00:14.0/usb3/3-4/3-4:1.2/0003:046D:C52B.0003/0003:046D:1024.0004/input/input13 > [ 2.985410] hid-generic 0003:046D:1024.0004: input,hidraw1: USB HID v1.11 Mouse [Logitech Unifying Device. Wireless PID:1024] on usb-0000:00:14.0-4:1 > [ 2.995266] input: Logitech K520 as /devices/pci0000:00/0000:00:14.0/usb3/3-4/3-4:1.2/0003:046D:C52B.0003/0003:046D:2011.0005/input/input17 > [ 3.026985] logitech-hidpp-device 0003:046D:2011.0005: input,hidraw1: USB HID v1.11 Keyboard [Logitech K520] on usb-0000:00:14.0-4:2 > [ 3.033173] input: Logitech M310 as /devices/pci0000:00/0000:00:14.0/usb3/3-4/3-4:1.2/0003:046D:C52B.0003/0003:046D:1024.0004/input/input18 > [ 3.033358] logitech-hidpp-device 0003:046D:1024.0004: input,hidraw2: USB HID v1.11 Mouse [Logitech M310] on usb-0000:00:14.0-4:1 > > From 5.4 (non-working): > > [ 2.908901] hid-generic 0003:046D:C52B.0001: input,hidraw0: USB HID v1.11 Keyboard [Logitech USB Receiver] on usb-0000:00:14.0-4/input0 > [ 2.909003] hid-generic 0003:046D:C52B.0002: unknown main item tag 0x0 > [ 2.909016] hid-generic 0003:046D:C52B.0002: unknown main item tag 0x0 > [ 2.909019] hid-generic 0003:046D:C52B.0002: unknown main item tag 0x0 > [ 2.909023] hid-generic 0003:046D:C52B.0002: unknown main item tag 0x0 > [ 2.909030] hid-generic 0003:046D:C52B.0002: unexpected long global item > [ 2.909038] hid-generic: probe of 0003:046D:C52B.0002 failed with error -22 > [ 2.909196] hid-generic 0003:046D:C52B.0003: hiddev0,hidraw1: USB HID v1.11 Device [Logitech USB Receiver] on usb-0000:00:14.0-4/input2 > [ 2.910586] logitech-djreceiver 0003:046D:C52B.0002: unknown main item tag 0x0 > [ 2.910601] logitech-djreceiver 0003:046D:C52B.0002: unknown main item tag 0x0 > [ 2.910605] logitech-djreceiver 0003:046D:C52B.0002: unknown main item tag 0x0 > [ 2.910609] logitech-djreceiver 0003:046D:C52B.0002: unknown main item tag 0x0 > [ 2.910616] logitech-djreceiver 0003:046D:C52B.0002: unexpected long global item > [ 2.910620] logitech-djreceiver 0003:046D:C52B.0002: logi_dj_probe: parse failed > [ 2.910623] logitech-djreceiver: probe of 0003:046D:C52B.0002 failed with error -22 > > FYI, intf 0 is a boot-interface keyboard, intf 1 is a boot-interface > mouse, and intf 2 is the wireless HID interface. That is weird. I also have (several) C52B devices here, and I don't see the problem (though arguably there are several firmware versions out in the open). The first part of failures show that even hid-generic can not parse the report descriptor. This means that the provided report descriptor is incorrect, which is weird because the one I just tested works fine. The second part (with logitech-dj) is showing the same errors, so that's expected given the previous problem. To actually get the report descriptors, `lsub -v` works fine if the HID subsystem is not bounded. So there we need to ask the reporter to run the command with the HID subsystem not handling the device, or simply in the working case use hid-recorder from https://gitlab.freedesktop.org/libevdev/hid-tools/. Another solution if the report descriptor is identical to what I have here (in the working case), would be to write a BPF kprobe to dump what hid_parse sees. It should be straightforward to write one, I got everything in place for i2c-hid at https://gitlab.freedesktop.org/bentiss/hid-bpf (though for using the CO-RE BPF program, I am not sure 5.4 is recent enough). > > It's not immediately obvious to me where the problem lies. Not obvious to me either... Cheers, Benjamin > > Thanks, > > Alan Stern >
