Ferrari F1 Wheel support

[ReazerDev]

Hi, I've got a T300rs and the (old) Ferrari F1 Wheel attachment.
When using the normal PlayStation attachment, everything's working fine, but whenever I use my F1 Wheel, it's not getting picked up by the driver.
I'm willing to test/help with anything you need! :)

[Kimplul]

Hi, thanks for the report. First of all, could you clarify what you mean by

not getting picked up by the driver.

Do you mean that the buttons won't work, or that the device doesn't show up at all in lsusb? Or maybe something else?

[ReazerDev]

It shows up as "Thrustmaster FFB Wheel" in lsusb. But in Oversteer it doesn't show op at all

[Kimplul]

Thanks, that sounds like the attachment causes the wheel to report itself under a different ID than the PlayStation attachment. @scarburato has a table of known IDs over at https://github.com/scarburato/hid-tminit

Could you post dmesg output when you connect the wheel? There should be some output from hid-tminit in there with the wheel model ID.

[ReazerDev]

Unknown wheel's model id 0x203, unable to proceed further with wheel init

This one?

[Kimplul]

Yep, thanks. I'll try adding that ID to hid-tminit tomorrow, I'm not home at the moment. If you don't want to wait that long you can try adding {0x0203, 0x0005, "Thrustmaster T300RS (F1 attachment)"}, to hid-tminit.h, around line 48.

Although this does make me think that the lower byte represents possible wheel attachments and the higher byte is the base. So maybe the init model should be changed to only check the base? Not sure.

[ReazerDev]

And yeah, I guess the second byte is the base

[ReazerDev]

Even after adding the new line, I still get unknown model id in dmesg, but interestingly my mouse stops working

[Kimplul]

The mouse thing is weird, not sure what that's about. Did you run rmmod hid-tminit and modprobe hid-tminit to refresh the module loaded into the kernel?
Also, forgot to mention initially but th_wheels_infos_length should be changed to 6 after adding the line I mentioned earlier.

[ReazerDev]

I just rebooted my pc, as I didn't know how to reload the module. And yes, I forgot to change the length to 6🤦🏼‍♂️ I'll try tomorrow

[Kimplul]

Hi, I forked a version of hid-tminit that only looks at the wheel base ID instead of base + attachment. Could you check if it initializes your wheel?

https://github.com/Kimplul/hid-tminit

Should be just make && make install to build and install. If you've installed hid-tmff2 with DKMS, you'll need to temporarily remove the module with dkms remove hid-tmff2/0.8. Also if your kernel is newer than 5.14, you might need to run rmmod hid-thrustmaster before installing hid-tminit.

[ReazerDev]

Ah yes, the problem is that I'm on 5.13 and had to do rmmod hid-thrustmaster.
It's getting picked up as Thrustmaster Thrustmaster T300RS Racing wheel now.

I've got one more question, you can select an 'Advanced Mode' on Windows, which maps a few buttons differently. (There are two joysticks on the F1 Wheel and if in "normal" mode, it treats them as one, if you select advanced, it treats one as a joystick and the other as individual buttons).

I probably have to listen to the USB traffic using Wireshark, select advanced mode on windows and see what happens, right?
Since I've never done anything like that, could you send me a small guide on how I would do it?
(I can send you my Discord if that'd be easier)

[ReazerDev]

Oh btw. I just noticed that there is an error

[Kimplul]

Oh btw. I just noticed that there is an error

That's unfortunately expected. The wheel breaks the USB spec by completely killing off the connection as soon as it initializes, and that results in an error code. Me and @scarburato discussed whether we should suppress the error, but decided against it, as we couldn't figure out which error codes were 'good' and which were 'bad'.

I probably have to listen to the USB traffic using Wireshark, select advanced mode on windows and see what happens, right?

Yep, pretty much. I wrote up a quick guide on how I did my captures here:
#6 (comment)

If that's not detailed enough, I'd be glad to hop on to a discord call (or chat, whichever you prefer) and try to answer questions. There's a number of tidbits of information in past issues for this project, but (#25) is probably the most useful thread at the moment. I'm sure there are also a number of blog posts etc. about USB packet capture, but none that I can see that would directly touch on Thrustmaster devices.

[ReazerDev]

Nice, thank you! I'll try around a little and if I get stuck, I'll post a comment here and we can hop on a call👍

[ReazerDev]

Hi, I wanted to start trying stuff out tomorrow. Do I need a VM running Windows, or is it also possible to use a fresh windows install?
It'd be easier, since I already have a dual boot setup with Windows and Pop!OS running.

[Kimplul]

Should be doable from within native Windows with Wireshark, but the setup is a bit different from what I described earlier. The wireshark wiki has a brief section on it:
https://wiki.wireshark.org/CaptureSetup/USB#Windows

I'm not sure if the USB packet formatting is different depending on which platform is used, but the actual content of the packets should still be the same.

[ReazerDev]

t300_advanced_toggling.zip
Hi, I hope that this is correct^^
I've included two files, one when I switch from the normal mode to the advanced, and the other is from advanced to normal.

I don't know anything about USB, but I think the switching happens between or around the two URB_FUNCTION_ABORT_PIPE?

[Kimplul]

I don't know anything about USB, but I think the switching happens between or around the two URB_FUNCTION_ABORT_PIPE?

I think it might actually happen earlier. When switching from normal to advanced, it seems that the wheel's USB device ID changes from 0xb66e to 0xb66f. I suspect that packets 41 in normal_to_advanced.pcapng and 39 in advanced_to_normal.pcapng are what actually causes the switch to happen, sort of similar to how hid-tminit causes the switch from init to this driver.

Should be relatively simple to add in a flag of some kind that causes the driver to send out the mode switch command. I have exams coming up in a couple of days so I'll focus on them for the time being and come back to this issue next week. You're of course free to grok around the source code, just be sure to do it in a VM so you don't accidentally brick your OS.

[ReazerDev]

I'd try myself, but I don't know anything about drivers and I have no clue what happens in the code haha.

Good luck ony our exams!

[Kimplul]

Alright, hello again, I just pushed a commit into https://github.com/Kimplul/hid-tmff2/tree/adv that adds in an adv_mode attribute to the module. Essentially how it works is you write anything into /sys/bus/hid/devices/XXXX:044F:B66E.XXX/adv_mode and it should automatically switch modes. Try something like echo 1 > /sys/bus/hid/devices/XXXX:044F:B66E.XXX/adv_mode and report back what happens.

On my machine and my wheel the command above didn't do anything, but it could be because I don't have the required attachment, not sure. I quickly checked that the actual command that is sent out seems to be identical to the one that you captured, but I could very well be overlooking something.

[ReazerDev]

How do I figure out which id is correct? I have quite a lot of directories in the devices directory

[Kimplul]

The XXXX depend on which bus and device spot the wheel uses. If you run lsusb, the bus should be the first batch XXXX and device the second batch, so for example if my wheel is connected to bus 3 and device 2, the whole path would be /sys/bus/hid/devices/0003:044F:B66E.0002.

I usually just let autocomplete show me which devices are connected and look for 044F:B66E. Unless you have lots of wheels connected, there should only be one device with those numbers.

[ReazerDev]

Can you add me on discord Reazer#9411 ?

i don't have an adv_mode file in the /devices/XXX folder and when trying echo 1 > /sys/bus/hid/devices/XXXX:044F:B66E.XXX/adv_mode, I get ad permission denied, even when using sudo