Re: Wrong inputs on DualSense Edge Wireless Controller

[Rahul Rameshbabu <sergeantsagara@protonmail.com>]

On Tue, 28 Nov, 2023 01:50:37 +0000 "Alexander Koskovich" <AKoskovich@protonmail.com> wrote:
> I tested in evtest, and it's reporting sane results (cross south, square west, etc). So this seems to be an issue outside of the playstation driver itself?
>

Right, I believe the problem stems from Proton -> Windows userspace
Xinput drivers or game/Steam controller test specific code though I am
not 100% sure.

  https://github.com/ValveSoftware/Proton/blob/proton_8.0/docs/CONTROLLERS.md

A easy way to confirm would be to run the Steam controller test from
Windows (if you have access) and see if it is weird. You will see from
the diagram that the driver is bypassed by hidraw on the linux side. If
hidraw is not usable on the system, a fallback occurs SDL2's js backend
is used to map the device into a "generic" Xbox controller on the
wine/Windows land.

>
> On Monday, November 27th, 2023 at 2:40 PM, Roderick Colenbrander <thunderbird2k@gmail.com> wrote:
>
>
>> This seems a little weird to me. Would you be able to test in using evtest? It would rule out some of the other middle layers which can do their own remapping.
>> Thanks,
>> Roderick
>>
>> On Mon, Nov 27, 2023, 6:42 AM Alexander Koskovich <AKoskovich@protonmail.com> wrote:
>>
>> > I was testing this in the Steam controller test, Cyberpunk 2077 (through GOG, not through Steam), and SuperTuxKart. All of these have wrong mappings for the Edge controller exclusively.
>> > For additional context I have a PS5 controller (non Edge) and it works out of the box. This just seems to be an issue with the Edge variant.
>> >
>> >
>> > On Monday, November 27th, 2023 at 9:37 AM, Roderick Colenbrander <thunderbird2k@gmail.com> wrote:
>> >
>> >
>> > >
>> > >
>> > > Hi Alexander,
>> > >
>> > > Sorry for the late reply, but I have been out for a few days. I'm not
>> > > aware of any button/axis mapping change between Edge and regular
>> > > DualSense. The HID reports stayed the same.
>> > >
>> > > I just did a quick check on my laptop also on Fedora 38 / kernel 6.5
>> > > and the mapping seems to be fine. In evtest, triangle reports
>> > > BTN_NORTH, square BTN_WEST, etcetera.
>> > >
>> > > The sticks, triggers and buttons seem to be fine as well. How are you
>> > > testing this?
>> > >
>> > > Thanks,
>> > > Roderick
>> > >
>> > > On Wed, Nov 22, 2023 at 4:18 PM Alexander Koskovich
>> > > AKoskovich@protonmail.com wrote:
>> > >
>> > > > [Resending email due to lack HTML message rejected]
>> > > >
>> > > > Hello,
>> > > >
>> > > > I am currently on Fedora 39 (6.5.12-300.fc39.x86_64) and I am noticing that the inputs for this controller are wrong primarily on the right side of the controller.
>> > > >
>> > > > playstation 0005:054C:0DF2.000C: hidraw11: BLUETOOTH HID v1.00 Gamepad [DualSense Edge Wireless Controller] on 10:3d:1c:fd:30:bc
>> > > > playstation 0005:054C:0DF2.000C: Registered DualSense controller hw_version=0x01000208 fw_version=0x01000036
>> > > >
>> > > > This is the current mapping that I'm seeing with the hid_playstation module loaded:
>> > > > "X" is actually "Square"
>> > > > "Square" is "Triangle"
>> > > > "Triangle" is "Circle"
>> > > > "Circle" is "X"
>> > > >
>> > > > Also the right joystick seems to be controlling "R2" instead of the right joystick. "L2" and "R2" triggers control the joystick instead. It's overall very weird. Has there been any similar reports to this?