Some controllers that mimic the Xbox 360 Controller are not initialized correctly

[dnmodder]

I have a fake Xbox 360 controller, specifically a Fantech GP11 Shooter, it happens that this type of controllers have problems to work on Linux because they are first recognized by usbhid and then disconnected and reconnected as controllers of xbox 360, to then not generate any key event.

In this case my problem is identical to this: Linux gamepad - no input events

And this is very similar: #118

Using VirtualBox, Wireshark and usbmon I analyzed the traffic between the device and the linux driver looking for the initialization codes, after finding them I wrote the following script: fixcontroller.py

I do not provide more data, like the output of dmesg, because they are the same as the first link I left.

But if necessary I can provide more information or try any solution they propose, in advance I apologize if something is not well understood, since this was translated from Spanish.

[zeunus]

Hablas español?

[MirceaKitsune]

I just got a new gamepad: Trust GXT 540 Yula. It has a button to switch between D-input and X-input modes: dinput works OOTB with no issues, albeit it has support for less features and some games don't accept it. xinput refuses to work, the device is recognized by the OS but no buttons or events are sensed. I asked why that is and someone pointed me to this issue.

I can confirm the problem reported here is the cause: The python script provided as a workaround fixes the problem and gets the device fully working in this input mode. Of course it must be ran each time after the controller is plugged in or the machine is restarted, which is very annoying and problematic to do (especially since it wants root access).

This is a concerning problem: Any information on when it might be resolved? Feel free to ask me for more data if it helps fix this... my OS is Manjaro KDE for the record. In the meantime is there an easier workaround to that python script, such as a bash script that does the same thing which would be easier to automate, or some OS setting we can change to fix the bad detection?

[MirceaKitsune]

Some basic info. First the device is listed as Bus 005 Device 024: ID 045e:028e Microsoft Corp. Xbox360 Controller by the lsusb command; It's not a real Xbox controller and IMO it was a bad design choice to label it as such.

The usb-devices command lists the controller as follows:

T:  Bus=05 Lev=02 Prnt=03 Port=00 Cnt=01 Dev#= 26 Spd=12  MxCh= 0
D:  Ver= 2.00 Cls=ff(vend.) Sub=ff Prot=ff MxPS=64 #Cfgs=  1
P:  Vendor=045e ProdID=028e Rev=01.10
S:  Manufacturer=SHANWAN
S:  Product=Controller
C:  #Ifs= 1 Cfg#= 1 Atr=a0 MxPwr=500mA
I:  If#=0x0 Alt= 0 #EPs= 2 Cls=ff(vend.) Sub=5d Prot=01 Driver=xpad

Lastly for now here's the dmesg log piece after plugging in the device:

1336a1337,1351
> [ 3317.060917] usb 5-2.1: new full-speed USB device number 6 using xhci_hcd
> [ 3317.184666] usb 5-2.1: New USB device found, idVendor=2563, idProduct=0575, bcdDevice= 2.00
> [ 3317.184669] usb 5-2.1: New USB device strings: Mfr=1, Product=2, SerialNumber=0
> [ 3317.184671] usb 5-2.1: Product: PS3/PC Gamepad
> [ 3317.184672] usb 5-2.1: Manufacturer: SHANWAN
> [ 3317.286799] input: SHANWAN PS3/PC Gamepad as /devices/pci0000:00/0000:00:08.1/0000:0c:00.3/usb5/5-2/5-2.1/5-2.1:1.0/0003:2563:0575.0004/input/input22
> [ 3317.286852] hid-generic 0003:2563:0575.0004: input,hidraw3: USB HID v1.10 Gamepad [SHANWAN PS3/PC Gamepad] on usb-0000:0c:00.3-2.1/input0
> [ 3317.861781] usb 5-2.1: USB disconnect, device number 6
> [ 3318.177585] usb 5-2.1: new full-speed USB device number 7 using xhci_hcd
> [ 3318.303668] usb 5-2.1: New USB device found, idVendor=045e, idProduct=028e, bcdDevice= 1.10
> [ 3318.303672] usb 5-2.1: New USB device strings: Mfr=1, Product=2, SerialNumber=0
> [ 3318.303673] usb 5-2.1: Product: Controller
> [ 3318.303674] usb 5-2.1: Manufacturer: SHANWAN
> [ 3318.922012] input: Microsoft X-Box 360 pad as /devices/pci0000:00/0000:00:08.1/0000:0c:00.3/usb5/5-2/5-2.1/5-2.1:1.0/input/input23
> [ 3318.922074] usbcore: registered new interface driver xpad

[dnmodder]

@MirceaKitsune I think it can be automated with a udev rule.

[MirceaKitsune]

@dnmodder Someone on Reddit suggested this. Unfortunately I never had to manually mess with udev before. Could anyone explain what needs to be edited and where and how? I can tell this obviously involves messing with root / system stuff and I'm always weary not to risk breaking something.

If that's easier, a script or command written in bash (no python) would also be helpful until this is resolved. I tried stuff like echo 1 > /dev/bus/usb/005/011 as a test to see what that does, but it just throws an invalid argument error and that's it.

[dnmodder]

@MirceaKitsune Write the following to a text file:
ACTION=="add", ATTRS{idProduct}=="028e", ATTRS{idVendor}=="045e", DRIVERS=="usb", RUN+="/path/to/script.py"
changing the script path, then save this file as 99-fixcontroller.rules (I chose that name at random) in /etc/udev/rules.d/99-fixcontroller.rules, then run the following command in a terminal:
sudo udevadm control --reload
That should be enough.

[MirceaKitsune]

Guess that script is the only way then. Mainstreamed it to a more permanent configuration. I don't mind executing it manually when plugging in the controller for now, but will keep that solution in mind too for ease. Thanks for all the help!

Also what happened to PR #120 if I may ask? I took a look and everyone says it's supposed to fix this issue, however it just dropped dead two years ago with no updates since. Was it just forgotten or is there a reason why it's waiting for so long?

[dnmodder]

@MirceaKitsune It is functional, since it does basically the same as the script but directly from the driver, but it is not written in the most adequate way, but it was the best the author of the PR could do, and I do not know how to program in C ++.

[MirceaKitsune]

Hmmm. Does it break anything in its current form? I support clean code, but over letting such an issue persist I'd take even an official workaround that's less ideal. Hope a solution in some form is found and accepted soon.