Fix Xbox 360 controller support

Doesn't work out of the box in linux using xboxdrv
<altname>Xbox Gamepad (userspace driver)</altname>
has to be added

Before I add that altname, are you sure that name is exclusively for the 360 controller? According to this, http://pingus.seul.org/~grumbel/xboxdrv/ it sounds like it uses the same device name for both the 360 controller and the xbox 1 controller. I don't want to create a conflict. If that device name is used for either xbox 1 or 360 controllers, what happens when someone uses an xbox 1 controller with xboxdrv?

Okay, thats a point.
As I only have 360 Gamepads I cant check that.

Anyone with a linux install, xboxdrv, and an xbox 1 controller, please let us know what device name shows up and how it behaves with the keymap, if it shares the same device name as the 360 controller. Also, MastaofDesasta, there is something you can check. Does the 360 controller using that altname with xboxdrv behave properly with the keymap in this thread? Use these images as a comparison: http://i.imgur.com/d5kZa.jpg http://i.imgur.com/lDqMz.jpg

Hey Guys. Just recently dove into XBMC with a new PC after having last used it on XBOX some time ago. Just set up and noodling now and trying to decide on a controller/keyboard input solution and saw the recent posting on the progress on the Xbox 360 controller. I have alot of those!!! I also came across some old posts regarding the MS Xbox 360 Controller ChatPad. Was wondering if this looks like a doable/viable solution to the ultimate controller for XBMC?

Here is a link to that initial thread:
http://forum.xbmc.org/showthread.php?tid=71179

There was an Xbox Chatpad Driver challenge contest originally started and someone posted on the XBMC forums in the link above. As far as I can tell the contest is still going
http://www.mp3car.com/hardware-developme...ge-70.html

There is also a FaceBook page for this contest apparently.

Also, there are some links in the thread pointing to work on ChatPad drivers for Windows:
http://www.neogaf.com/forum/showthread.php?t=392607
http://code.google.com/p/chatpad-super-d...ingStarted

Not sure of it's viability yet with Windows/XBMC, but for me, I would get a backlit ChatPad connected to a 360 controller for sure.

p.s. - sorry for the lengthy post

Dondi: Chatpad support would be excellent. I think it would act like an actual keyboard though (if you have proper working modded drivers), so neither a new device name nor keymap adjustment would be necessary to support it. Thanks for sharing though. That being said, you can use the onscreen keyboard via a regular xbox 360 controller with relative ease. Don't forget to try all those xbox 360 controllers with the keymap in this thread to see if you can unearth some new device names for us.

I will be noodling soon.... setting up my RAID and NAS stuff this week (need a refresher on some XBMC stuff ATM).

I only brought up the ChatPad thing because it looks like the thread about it on here died... I think it would be perfect!!

Not sure that it works as a keyboard as far as Win OS et al are concerned, which is probably why there is a contest to get this thing to work with any HTPC (thats the gist of what I got from the contest thread and the drivers for ChatPad threads)

Thanks!!

(2012-12-04 01:14)Dondi Wrote:  Not sure that it works as a keyboard as far as Win OS et al are concerned, which is probably why there is a contest to get this thing to work with any HTPC (thats the gist of what I got from the contest thread and the drivers for ChatPad threads)

I know it isn't supported natively. Whichever driver they create to support the chatpad will have it behave like a keyboard to the system, rather than as a secondary joystick device that needs a keymap written for xbmc. That's what I meant.

Following MastaofDesasta's post about linux and xboxdrv, if I start xboxdrv without any config, xbmc lists my xbox 360 wireless controller as
04:46:06 T:3035742944 INFO: opened device 'Xbox Gamepad (userspace driver)' (file name /dev/input/event0), m_bSkipNonKeyEvents 0

By the way, my controller is listed as Microsoft Xbox 360 Wireless Controller (PC) if I don't change the name and tried to add it as an altname without any effect.

But if I specify a config like the following:
[xboxdrv]
ui-clear=true
extra-devices=false
extra-events=false
deadzone=4000
device-name="Xbox 360 Wireless Receiver"
(...)

I can change the name to match the ones in the /xbmc/system/keymaps/joystick.Microsoft.Xbox.360.Controller.xml and xbmc seems to recognize it correctly in the log:
04:53:28 T:3036242656 INFO: Found input device /dev/input/event0
04:53:28 T:3036242656 INFO: opened device 'Xbox 360 Wireless Receiver' (file name /dev/input/event0), m_bSkipNonKeyEvents 0

The issue is that it doesn't work at all with xbmc, but I receive all the events in the ssh sessions from the controller... Is it supposed to work out of the box? Do I need to povide such xboxdrv config?

Thanks.

maboivin: xboxdrv does send input to xbmc in my testing. Did you restart xbmc after starting xboxdrv?

I installed Frodo beta 1 from the unstable ppa in a new install of Xubuntu 12.10 x64. The default device name of the built in 360 driver on Linux is called "Microsoft X-Box 360 pad", and the xboxdrv userspace driver is called "Xbox Gamepad (userspace driver)".

XBMC needs a second redesigned keymap file for these two device names (and others as they are discovered), representing fundamentally different numbering and axes. There's no other way to support these drivers on *nix. And yes, in Linux, the Guide button is mapped to a button number, and using the unmodified keymap in this thread, any use of the triggers destroys the whole session, causing future inputs to fail or behave poorly. To have the buttons behave the same way crossplatform, *nix device names need a separate keymap (possible a 3rd for OSX, but I don't have that system to test; so chime in OSX users). If the triggers' buggy behavior can't be fixed by a keymap, it will require either code change or being excluded from the *nix keymap entirely.

If we discover that some device names appear on both Windows and *nix, OS detection will be required to determine which keymap to use. Or a different keymap bundled with each OS version of XBMC would solve that problem.

I'm reverting the parent post back to v1.95, as my keymap is now known to only be capable of supporting Windows accurately. A separate *nix keymap is required.