Asetek LaPrima Wheelbase support

uweeby commented 1 month ago

Rotation is detected on a gamepad tester and seems to show 99% range as intended from -1 to +1 (I have a bump stop enabled in the firmware).

iRacing calibration tool detected it as going from 50%-100% rotation which is odd. Gives error that the wheel does not support force feedback.

Windows raw values go from 0 to 65535

JacKeTUs commented 1 month ago

Can you please give kernel logs when you connect the wheelbase to the PC? Also, please, install Wireshark + USBMon and create .pcap file from Linux, when you connect the wheelbase to the PC. That's how we can look into the descriptor

uweeby commented 1 month ago

Having a hard time figuring out how to do the pcap but here are the kernel logs.

[ 133.129908] usb 1-6.2: new high-speed USB device number 9 using xhci_hcd [ 133.264040] usb 1-6.2: New USB device found, idVendor=2433, idProduct=fa00, bcdDevice= 1.38 [ 133.264047] usb 1-6.2: New USB device strings: Mfr=1, Product=2, SerialNumber=3 [ 133.264050] usb 1-6.2: Product: Gaming HUB [ 133.264052] usb 1-6.2: Manufacturer: Asetek SimSports [ 133.274088] hub 1-6.2:1.0: USB hub found [ 133.276933] hub 1-6.2:1.0: 7 ports detected [ 134.183260] usb 1-6.2.1: new full-speed USB device number 10 using xhci_hcd [ 134.524436] usb 1-6.2.1: New USB device found, idVendor=2433, idProduct=f303, bcdDevice= 2.00 [ 134.524443] usb 1-6.2.1: New USB device strings: Mfr=1, Product=2, SerialNumber=3 [ 134.524446] usb 1-6.2.1: Product: LaPrima Wheelbase [ 134.524448] usb 1-6.2.1: Manufacturer: Asetek SimSports [ 134.524450] usb 1-6.2.1: SerialNumber: 365134633131 [ 134.614975] Asetek SimSports LaPrima Wheelbase: Invalid code 768 type 1 [ 134.614980] Asetek SimSports LaPrima Wheelbase: Invalid code 769 type 1 [ 134.614982] Asetek SimSports LaPrima Wheelbase: Invalid code 770 type 1 [ 134.614983] Asetek SimSports LaPrima Wheelbase: Invalid code 771 type 1 [ 134.614985] Asetek SimSports LaPrima Wheelbase: Invalid code 772 type 1 [ 134.614986] Asetek SimSports LaPrima Wheelbase: Invalid code 773 type 1 [ 134.614988] Asetek SimSports LaPrima Wheelbase: Invalid code 774 type 1 [ 134.614989] Asetek SimSports LaPrima Wheelbase: Invalid code 775 type 1 [ 134.614991] Asetek SimSports LaPrima Wheelbase: Invalid code 776 type 1 [ 134.614992] Asetek SimSports LaPrima Wheelbase: Invalid code 777 type 1 [ 134.615063] input: Asetek SimSports LaPrima Wheelbase as /devices/pci0000:00/0000:00:01.3/0000:02:00.0/usb1/1-6/1-6.2/1-6.2.1/1-6.2.1:1.0/0003:2433:F303.0007/input/input25 [ 134.615259] hid-generic 0003:2433:F303.0007: unknown set_effect report layout [ 134.615264] hid-generic 0003:2433:F303.0007: input,hiddev99,hidraw6: USB HID v1.11 Joystick [Asetek SimSports LaPrima Wheelbase] on usb-0000:02:00.0-6.2.1/input0

JacKeTUs commented 1 month ago

Usbmon is what you need

Looks like PIDFF again, i think it should work just fine with our patches to pidff, but first we need to see the descriptor :D

moonrail commented 4 days ago

Hey, from my experience, this Wheelbase "just works" under Proton without any additional driver (so hid-pidff?). Seems to be pretty similiar to Simucube: https://granitedevices.com/wiki/Using_Simucube_wheel_base_in_Linux This should be the same for all other Asetek Wheelbases (Forte, Invicta) as well.

What does not work (or at least I cannot get to work) is RaceHub - so no configuration from Linux possible other than mimicking USB commands sniffed while using RaceHub in Windows. I've "reverse engineered" some commands such as unlocking High Torque and am able to send them via pyusb Python Scripts, but this approach is highly fragile and may break with any firmware update.

Asetek has a lot of vendor specific USB HID reports in its descriptor (both La Prima wheelbase and wheel) and my dmesg also shows USB HID errors about unsupported effects. I've attached the descriptor: asetek_la_prima_wheelbase_hid_report.txt

I have have used this Wheelbase successfully with:

Assetto Corsa Competizione
AMS2
Beamng.drive
Dirt Rally 2.0 (requires adding the wheelbase device configuration file, just as with EA WRC: https://www.overtake.gg/news/ea-sports-wrc-ffb-guide-moza-asetek-fix.1456/) - this is required for Windows as well, nothing Linux specific, its just how the developer decided to implement device support

Steps are simple:

One time setup

Create an udev rule allowing unprivileged users access to wheelbase and wheel via hidraw e.g. a file /etc/udev/rules.d/60-steam.rules with content:

# Asetek La Prima Wheelbase
KERNEL=="hidraw*", ATTRS{idVendor}=="2433", ATTRS{idProduct}=="f303", MODE="0660", TAG+="uaccess"
# Asetek La Prima Steering Wheel
KERNEL=="hidraw*", ATTRS{idVendor}=="2433", ATTRS{idProduct}=="f203", MODE="0660", TAG+="uaccess"

Reload udev rules: sudo udevadm control --reload-rules && sudo udevadm trigger
Per game setup

Add Proton setting PROTON_ENABLE_HIDRAW=0x2433/0xf303,0x2433/0xf203 to your environment. This can be done via Steam via Game Properties > General > Launch Options, so e.g.: PROTON_ENABLE_HIDRAW=0x2433/0xf303,0x2433/0xf203 %command%

Forte and Invicta have different PID values, so adjust accordingly.

Alternative: Per Proton environment setup

Alternatively to per-game one can set this setting once for each Proton installation via user_settings.py: https://github.com/ValveSoftware/Proton/wiki/Proton-FAQ

JacKeTUs commented 3 days ago

@moonrail, thank you! Did you encounter original error from the first message, with odd rotation range detection? Did you try to launch RaceHub with similar fixes as in GraniteDevices guide? (DisableInput, Enable SDL, etc...) If it won't work, then we maybe could bring these reverse engineered commands to the Oversteer or Boxflat. And no custom driver needed at all, that's good :D

moonrail commented 3 days ago

@JacKeTUs

OPs issue

I am not sure I completely understand OPs issue. Even without udev rules (hence without hidraw) I can use the full range in jstest: asetek_la_prima_wo_hidraw

However I could manufacture a seemingly similiar behaviour (range only from 50 to 100%) if I unset all calibration (Button "Raw Events" in jstest). Once I've done that, Wheel-center was reported in jstest as 100% (as if fully turned to the right) and Wheel-left until bumpstop was reported as 50% (as if wheel was centered). Maybe OP is struggling with uncalibrated events?

I would assume OPs issue could be solved by configuring permissive hidraw udev rules and (if using Proton) defining hidraw usage for the Wheelbase as described in my initial comment.

RaceHub

As for RaceHub: Yes, I've tried all documented Simucube methods and tried all seemingly related Wine settings - no luck yet. I've used Wine 8 & 9, setup manually as in Simucube docs as well as with bottles. Of course I also tested with udev rules, even some that would allow any and all USB access, but it made no difference. RaceHub logs the following: 240819_213952 - RaceHub Log.log I mean - "Skipped USB check" sounds very much like the issue, but it may be a result of the previous Exception where a value is null but is not expected to be while discovering installed games. Therefore I've asked Asetek via Email for any clue to go further or maybe even support, but the polite answer was simply, that they do not support Linux and they have to resources to look into it.

Oversteer & Boxflat

Regarding Oversteer & Boxflat - I am not content with my current lack of understanding - as I said, its only mimicking seen USB traffic, not really understanding its contents. Setting High Torque for example is possible this way exactly once. If its disabled again (by own commands) the same payload does not enable it again.

The following enables it once: Report endpoint to send to: 0x6b Report endpoint to wait for answers on: 0x6c

Sending HID data: 97006bb71700000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
Wheelbase does not answer
Sending HID data: 97006bb71700000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
Wheelbase answers, but I do not know what - maybe a simple confirmation?
Sending HID data: 96006bb71700000000000000000000000000000000e28a63090000000000000000000000000000000000000000000000000000000000000000000000
Wheelbase does not answer, but plays its "High Torque enabled" sound and is now in High Torque mode

Disabling High Torque is an unanswered send of HID data: 96006bb71700000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000

Profiles can be switched with HID data: 26{profile_id:02x}00000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000

Current configuration can be requested via HID data: 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 And wheelbases answer has selected profile at byte 19 and high torque state at byte 55, where 00 means disabled and 04 enabled.

And so on. I do not think that this is nearly enough for Oversteer or Boxflat.

JacKeTUs / linux-steering-wheels