Wrong button mappings reported by evdev with PowerA Nintendo Switch gamepad

Wrong button mappings reported by evdev with PowerA Nintendo Switch gamepad

[Cameron]

Hello all,

This is my first time reporting a kernel issue, so please let me know if this is not the right place to report it. I am using a third party PowerA USB wired Switch controller on Void Linux. Games and applications are able to successfully detect and use the controller, however the button mapping is very wrong in every program I have tested. I tested this with evtest-qt to see what button values evdev is reporting. Since there appears to be no way to remap these from userspace, I suspect the problem lies in the kernel driver.

Pressing the two analog sticks (which should be BTN_THUMBL and BTN_THUMBR) incorrectly generate events for BTN_SELECT and BTN_START, and pressing the + and - buttons (which should be mapped to BTN_START and BTN_SELECT) instead result in BTN_TL2 and BTN_TR2 events.

The device in question (as reported by lsusb) is:
Bus 001 Device 003: ID 20d6:a711 Core (Plus) Wired Controller
and I am using kernel version 5.15.26_1 on Void Linux.

I understand that this gamepad is intended to be used with the Nintendo Switch console, so it may not be as HID compliant as, say, a Logitech. However the kernel has a standard for how these buttons are to be mapped, which the driver used for this gamepad appears to violate. https://www.kernel.org/doc/html/v4.13/input/gamepad.html
Section 4.3 states: "All new gamepads are supposed to comply with this mapping. Please report any bugs, if they don’t."

I can provide more info if needed.

Thanks,
Cameron

Update: Wrong button mappings reported by evdev with PowerA Nintendo Switch gamepad

[Cameron]

On Wed, 02 Mar 2022 12:45:17 -0600
Cameron <cameronghall@cox.net> wrote:

> Hello all,
> 
> This is my first time reporting a kernel issue, so please let me know
> if this is not the right place to report it. I am using a third party
> PowerA USB wired Switch controller on Void Linux. Games and
> applications are able to successfully detect and use the controller,
> however the button mapping is very wrong in every program I have
> tested. I tested this with evtest-qt to see what button values evdev
> is reporting. Since there appears to be no way to remap these from
> userspace, I suspect the problem lies in the kernel driver.
> 
> Pressing the two analog sticks (which should be BTN_THUMBL and
> BTN_THUMBR) incorrectly generate events for BTN_SELECT and BTN_START,
> and pressing the + and - buttons (which should be mapped to BTN_START
> and BTN_SELECT) instead result in BTN_TL2 and BTN_TR2 events.
> 
> The device in question (as reported by lsusb) is:
> Bus 001 Device 003: ID 20d6:a711 Core (Plus) Wired Controller
> and I am using kernel version 5.15.26_1 on Void Linux.
> 
> I understand that this gamepad is intended to be used with the
> Nintendo Switch console, so it may not be as HID compliant as, say, a
> Logitech. However the kernel has a standard for how these buttons are
> to be mapped, which the driver used for this gamepad appears to
> violate. https://www.kernel.org/doc/html/v4.13/input/gamepad.html
> Section 4.3 states: "All new gamepads are supposed to comply with
> this mapping. Please report any bugs, if they don’t."
> 
> I can provide more info if needed.
> 
> Thanks,
> Cameron

So, here's an update on this situation. I managed to hack around
this situation by writing a simple kernel module to override the button
mapping for this device. The code for the module can be found on my
GitHub here:
https://gist.github.com/camthesaxman/af7099505103a555518741b4083eaea8
I have almost zero experience with kernel development, so I don't know
how this kind of thing would be properly fixed in-tree (perhaps
something with hid-quirks.c?) and make a proper patch for the kernel.

Thanks,
Cameron