USB controllers aren’t mapped in order in SiDi

[kounch]

On a SiDi FPGA, at least with cores jtpcs1 and jtcps15 (Cadillacs and Dinosaurs, The King of Dragons, etc.), only the first two USB detected controllers work:

• Arcades are configured with tree players option
• Tested all controllers in pairs, with only two players, and they all work ok
• Also tested changing between all the USB ports, with two players, and they work aso ok
• Third player can be controlled without any problem with a keyboard

But, when having tree controllers plugged in, although the firmware sees them, the joystick test pass all ok, etc. the third controller (whichever is detected as third), does not work at all.

[kounch]

Tried also with several firmware versions between 200614 and 211121, and the issue persists

[jotego]

I have reviewed the connections and the third and fourth joysticks are connected all the way from the serial interface that links the FPGA and the MCU together to the core logic. I wonder if the firmware is communicating in the expected way. Do you know of a core in SiDi where three joysticks work?

[kounch]

You are right. After asking for some advice, i have tried with other cores (Dynablaster with minimig and Bomberman with tgfx16) some of them work ok with three controllers, and some do not.

[kounch]

After doing some more test with 5 controllers at the same time, i see that, what the jtcps cores are using, does not match what the firmware shows on the test screen.

USB controller 1 on test screen -> controller 1 in core USB controller 2 on test screen -> controller 2 in core USB controller 4 on test screen -> controller 3 in core USB controller 5 on test screen -> controller 4 in core

This way, having three controllers attached will never work with a three player game

[kounch]

When using the PC-Engine / SuperGrafx Core, the numbers match, when using 3, 4 or 5 players with Bomberman

USB controller 1 on firmware -> controller 1 in core USB controller 2 on firmware -> controller 2 in core USB controller 3 on firmware -> controller 3 in core USB controller 4 on firmware -> controller 4 in core USB controller 5 on firmware -> controller 5 in core

[kounch]

Changed issue text according to latest info

[jotego]

I have asked for help to the MiST firmware developers. That's the firmware used in SiDi.

[gyurco]

USB controller 1 on test screen -> controller 1 in core USB controller 2 on test screen -> controller 2 in core USB controller 4 on test screen -> controller 3 in core USB controller 5 on test screen -> controller 4 in core

Where is controller 3? Is it empty?

[gyurco]

It would be helpful to set up ARM debug logging: https://github.com/mist-devel/mist-board/wiki/Troubleshooting#usb Then it can be seen what happens when the controllers are enumerated.

[gyurco]

Actually I could reproduce it in Cadillacs & Dinosaurs input test screen: Controller 3 is ignored, and 3P is Controller 4. I check if it is a firmware or core error.

[gyurco]

The problem is in JTframe: https://github.com/jotego/jtframe/blob/master/target/mist/jtframe_mist_base.v#L212 You're using user_io signals joystick_0,_1,_3,_4, but they should be _0,_1,_2,_3

[jotego]

Blast! It was in front of me all the time. I've fixed it here: https://github.com/jotego/jtframe/commit/c059a6623a8f914eb8ec7d19b62a3d56a3e2fd45

Thank you, Gyurco!

[gyurco]

It happens to all of us :)