D-Input vs X-Input on 8Bitdo controllers

mitu

@bmn001 said in D-Input vs X-Input on 8Bitdo controllers:

Is there a way for me to confirm that xpadneo is actually being used by this controller?

Disconnect the controller, then run - from the terminal:

dmesg -w

Connect the controller again and see if the xpadneo driver takes over the connection. Press Ctrl+C to stop dmesg and go back to the terminal.

bmn001

@mitu Yeah that output is full of xpadneo messages referencing my gamepad. So I guess the driver is in use at least.

Jstest does show the two triggers as axis, not buttons, if that makes a difference.

WeirdH

@bmn001 said in D-Input vs X-Input on 8Bitdo controllers:

However...I'm helping out a not-so-computer-literate friend who I'm sure will manage to overwrite this setting at some point in the future by using Emulation Station's button configurator.

Just set Emulationstation to kiosk or kid mode after configuring. It will hide all the setup stuff.

mitu

@bmn001 said in D-Input vs X-Input on 8Bitdo controllers:

@mitu Yeah that output is full of xpadneo messages referencing my gamepad. So I guess the driver is in use at least.

Can you post the messages ? You can copy them by running the dmesg command over SSH.
The triggers being recognized as axis is ok and EmulationStation should be able to deal with them.

bmn001

@mitu Yep, here's a log. I start with the controller off, then power it on and this is the output I see as soon as it connects:

[34441.432303] xpadneo 0005:045E:02E0.000A: buggy firmware detected, please upgrade to the latest version
[34441.432314] xpadneo 0005:045E:02E0.000A: pretending XB1S Windows wireless mode (changed PID from 0x02E0 to 0x028E)
[34441.432322] xpadneo 0005:045E:02E0.000A: working around wrong SDL2 mappings (changed version from 0x00000903 to 0x00001130)
[34441.432330] xpadneo 0005:045E:02E0.000A: report descriptor size: 307 bytes
[34441.432338] xpadneo 0005:045E:02E0.000A: fixing up report descriptor size
[34441.433222] xpadneo 0005:045E:02E0.000A: battery detected
[34441.433230] xpadneo 0005:045E:02E0.000A: enabling compliance with Linux Gamepad Specification
[34441.433327] input: 8Bitdo SN30 Pro as /devices/platform/soc/fe201000.serial/tty/ttyAMA0/hci0/hci0:11/0005:045E:02E0.000A/input/input23
[34441.446011] xpadneo 0005:045E:02E0.000A: input,hidraw2: BLUETOOTH HID v11.30 Gamepad [8Bitdo SN30 Pro] on d8:3a:dd:1c:dd:84
[34441.446027] xpadneo 0005:045E:02E0.000A: controller quirks: 0x00000027
[34442.106805] xpadneo 0005:045E:02E0.000A: 8Bitdo SN30 Pro [e4:17:d8:7a:f4:73] connected

I've noted the first line about "buggy firmware," but I'm already running the latest 8Bitdo firmware available (2.02) so I'm not sure what that's looking at.

mitu

Yeah, it looks like xpadneo takes over and applies some quirks to the controller, so it's clearly managed by xpadne.

Not sure why the L2/R2 buttons don't get mapped in ES, but you can start EmulationStation using the debug option and it should print what input is activated during the input configuration dialog (log is printed on console and also written in $HOME/.emulationstation/es_log.txt).

emulationstation --debug

bmn001

@mitu The ES debug log seems to indicate that ES isn't seeing the triggers being pressed at all.

After "Mapping [Button 5] -> RightShoulder" when I was asked to press the left and right trigger buttons, I made sure to press L2 and R2 ten times each so they would be obvious in the log before I did a "hold button to skip" to move on.

Instead, there's nothing in the log about triggers at all, and indeed the UI gave no indication that any button or trigger was being pressed at that time.

Jul 12 08:30:51 lvl2:   Added known joystick '8Bitdo SN30 Pro' (instance ID: 0, device index: 0)
Jul 12 08:30:51 lvl3:   SystemView::getViewElements()
Jul 12 08:30:58 lvl2:   Configuring device 0 (8Bitdo SN30 Pro).
Jul 12 08:31:00 lvl2:     Mapping [Hat 0 up] -> Up
Jul 12 08:31:00 lvl2:     Mapping [Hat 0 down] -> Down
Jul 12 08:31:01 lvl2:     Mapping [Hat 0 left] -> Left
Jul 12 08:31:02 lvl2:     Mapping [Hat 0 right] -> Right
Jul 12 08:31:02 lvl2:     Mapping [Button 7] -> Start
Jul 12 08:31:03 lvl2:     Mapping [Button 6] -> Select
Jul 12 08:31:03 lvl2:     Mapping [Button 0] -> A
Jul 12 08:31:04 lvl2:     Mapping [Button 1] -> B
Jul 12 08:31:05 lvl2:     Mapping [Button 2] -> X
Jul 12 08:31:05 lvl2:     Mapping [Button 3] -> Y
Jul 12 08:31:06 lvl2:     Mapping [Button 4] -> LeftShoulder
Jul 12 08:31:06 lvl2:     Mapping [Button 5] -> RightShoulder
Jul 12 08:31:18 lvl2:     Mapping [Button 9] -> LeftThumb
Jul 12 08:31:19 lvl2:     Mapping [Button 10] -> RightThumb
Jul 12 08:31:19 lvl2:     Mapping [Axis 1-] -> LeftAnalogUp
Jul 12 08:31:20 lvl2:     Mapping [Axis 1+] -> LeftAnalogDown
Jul 12 08:31:20 lvl2:     Mapping [Axis 0-] -> LeftAnalogLeft
Jul 12 08:31:21 lvl2:     Mapping [Axis 0+] -> LeftAnalogRight
Jul 12 08:31:22 lvl2:     Mapping [Axis 4-] -> RightAnalogUp
Jul 12 08:31:22 lvl2:     Mapping [Axis 4+] -> RightAnalogDown
Jul 12 08:31:23 lvl2:     Mapping [Axis 3-] -> RightAnalogLeft
Jul 12 08:31:23 lvl2:     Mapping [Axis 3+] -> RightAnalogRight
Jul 12 08:31:24 lvl2:     Mapping [Button 6] -> HotKeyEnable
Jul 12 08:31:24 lvl3:   fireEvent: config-changed
Jul 12 08:31:24 lvl3:   fireEvent: controls-changed
Jul 12 08:31:24 lvl2:           /opt/retropie/supplementary/emulationstation/scripts/inputconfiguration.sh

Given that:

xpadneo identifies and drives my controller
jstest recognizes the triggers when pressed
The the triggers work within retroarch when I manually edit a config file
I've seen identical behavior using both xpad and xpadneo (so far)

...it's looking a lot like a possible bug in Emulation Station maybe?

I really do appreciate you taking the time to step through all this with me. Thank you.

mitu

@bmn001 said in D-Input vs X-Input on 8Bitdo controllers:

...it's looking a lot like a possible bug in Emulation Station maybe?

More likely with the SDL2 library used by EmulationStation underneath, which handles also the input. However, if things work with RetroArch, it's not a big deal, since L2/R2 are not really used by EmulationStation itself.

bmn001

@mitu I have only a vague grasp of this, so please bear with me.

I just installed some random SDL2 controller testing tool from github. It needed libsdl2-dev installed to build. The version I used was "2.0.10+5rpi". That testing tool does correctly see the triggers.

So I just pulled the latest Emulation Station from github and built it on the same machine (Ras Pi4). It still will not see the triggers.

mitu

@bmn001 said in D-Input vs X-Input on 8Bitdo controllers:

So I just pulled the latest Emulation Station from github and built it on the same machine (Ras Pi4). It still will not see the triggers.

What's the version of EmulationStation you have installed ?

bmn001

@mitu Whew. It's getting weird now.

I just built the latest version of SDL2 from github.

cmake sees SDL2 as 2.29.0 now instead of the previous 2.0.10.

So then I rebuilt Emulation Station.

The preinstalled Emulation Station was v2.11.2rp. The one I've built identifies as 2.12.0rp-dev.

The good news:

Emulation Station now recognizes my L2/R2 triggers!

The weird news:

When I launch retroarch emulators, they do not recognize my L2/R2 trigger presses.
While the bluetooth dialog still reports my controller name as "8Bitdo SN30 Pro" when I pair and connect it, this newly-built version of Emulation Station reports it as "Xbox 360 Controller".