PS5 doesn't like mouses

I tried to get used to the controller with its cute thumb stick thingies but I just couldn't. Sadly my logitech mouse and keyboards do not pair with the console for some unknown reason, and even if I paired them I was devasted to learn that most games ignore them anyways. Still, I don't imagine that as far as the software is concerned both input device supply the computer with pretty much the same data. Delta x/y, button press, whatever. So I decided to build a device that translates the mouse and keyboard inputs to gamepad inputs. 
               Mouse data         Gamepad data
┌──────────────────┐    ┌──────────────┐    ┌─────┐
│ Mouse & Keyboard ├───►│ Magic device ├───►│ PS5 │
└──────────────────┘    └──────────────┘    └─────┘

Basic plan

I had a pico laying around that sounds like a perfect match for this job! Sadly it has only one USB (type C in my case) port, so I had to solder a second one. Also, since the RP2040 has just one USB peripheral, I had to use this project that uses the mysterious PIO to emulate USB communication on GPIO pins. Then I bought a "connector" for $0.50 from the closest electronics and attached it to VBUS, GND, GPIO0 and GPIO1.

I'm not experienced with soldering by any stretch of the imagination, but I believe this is the first project where I'm not ashamed by the wiring. My handy $10 battery powered iron from china helped too.

Debugging

But wait! If I use the "internal" USB to emulate a gamepad and the secondary USB to emulate a host that reads a mouse, how would I do my printf debugging? How will I screen /dev/tty? I know there is something called SWD that allows you to attach a second pico with a proper debugger, but this felt too complicated.

I decided to buy one of those $2 OLED screen that you communicate over I²C with. The SSD1306 based one I got has a resolution of 128x64, which fits nicely around ~8 lines of text. I wrote a little function with the same signature as printf, but it internally uses something like a ring buffer. The first 8 logs go from top to bottom, and the ones after that "scroll" the text like in a normal terminal. 

ssd1306_t disp;
char logged[8][22];
uint32_t logcnt = 0;
void logstr(char *fmt, ...) {
    va_list argptr;
    va_start(argptr, fmt);
    vsnprintf(logged[logcnt % 8], 22, fmt, argptr);
    va_end(argptr);
    logcnt += 1;

    ssd1306_clear(&disp);
    int offset = logcnt - 8;
    if (offset < 0) offset = 0;
    for (int i = 0;i < 8;i ++) {
        int y = i * 8;
        ssd1306_draw_string_with_font(&disp, 0, y, 1, font_8x5, logged[(i+offset)%8]);
    }
    ssd1306_show(&disp);
}

As a bonus it also helped me test I was received all the HID mouse reports by simulating a little cursor that moves on the screen.

Testing

I plugged in my first version into my mac and testing it in a webbrowser. I was pleased to find out it was running very smoothly, although I had to stop drawing on the screen on every report by the mouse.

You'll notice that chrome recognizes it as a dualshock controller. The reason is that tinyUSB requires me choose a vendor and product ids for my product and I copied the ones from the real controller I had. If I made up my own ids, the test still worked, but the test app did not render the nice 2d image, just a bunch of raw numbers.

Authentication

Sadly I'm not done. When I plug it into my PS5 nothing happens. After some research it looks like the PS5 does not want to work with unnoficial / unlicensed controllers. There is some alternate handshake, probably happending with CDC instead of HID. I haven't got the time to figure it out, but I image I can turn my real dualshock controller into an authenticator and my mouse as the actual source of data. Stay tuned!