On Fri, Jul 26, 2019 at 2:04 PM Hanno Zulla <abos@xxxxxxxx> wrote: > > Hi there, > > during a discussion of a Retropi issue [1] a question about the Linux > kernel gamepad spec compared to the actual gamepad drivers came up. > > The spec says in [2]: > > > "ABS trigger values start at 0, > > pressure is reported as positive values." > > The kernel drivers for the XBox 360 gamepad, for the Sony Sixaxis PS3 > gamepads and other drivers based on their example (like the BigBen > gamepad) however return trigger axis values with a value range that... > > - starts at -32767 (trigger resting position) and > - goes up to +32767 (full pressure on the trigger) > - with 0 being the middle position (half pressure) > > Please clarify, is the spec wrong or are the drivers in violation of it? > > If they are in violation, which should be changed, the spec or the > drivers' behavior? > Well some of the drivers are kind of in violation. The spec itself was added in 2013, while many of the devices or drivers are way older. The spec also states: "Legacy drivers often don’t comply to these rules. As we cannot change them for backwards-compatibility reasons, you need to provide fixup mappings in user-space yourself. " It is very, very hard to change the behaviour as it would break applications. (We actually went to this once for ds3 / ds4, but that's also because the driver was very incorrect and using illegal axes even. Was a huge pain..) We should pay attention to new drivers to support the spec as best as possible. For devices which just use the generic HID parser ranges can still be negative/postive for triggers, though mapping might also be off then. Thanks, Roderick
