Show Posts

You can view here all posts made by this member. Note that you can only see posts made in areas to which you currently have access.

Topics - kaspro

Pages: [1]

Discussion - Software / FT90X Get an Interface's HID Report Descriptor

« on: January 06, 2021, 04:04:15 AM »

I'm trying to write some code on my FT90X to both get and parse multiple Interface HID Report Descriptors.

I'm doing this to make sure I'm only listening to interfaces that are Keyboard/Keypad and to parse out the data properly using the report descriptor.

Looking through the FT9xx API Programmers Manual and all the example code in the Bridgetek folder, I never see an example of getting the report descriptors. Only setting them (USB Device examples)

The end goal is that I listen to all keyboard-reporting interfaces, parse the data they send back so I can see what keys a user is pressing, and use that data for my main application.

How can I get this information? My USB Host code can currently listen for a HID device, initialize the interfaces, and listen to them. I just don't know where to get those Interface HID Report Descriptors.

Discussion - Software / FT90X USB Question: Handling Multiple Keyboard Endpoints (BOOT, NKRO, etc.)

« on: December 30, 2020, 04:07:08 AM »

I'm making a Keyboard-to-Controller adapter using the FT908. Pretty simple design: You plug a keyboard into one end, and the other end plugs into a Nintendo Switch. From there, you can use your keyboard as a controller.

Right now I'm trying to handle the USB keyboard inputs and this harder than I expected.

Right now I'm using a Corsair K65. It has 3 endpoints and I need to listen to them all simultaneously:

A "BOOT" endpoint (where it sends data matching the typical 6-KRO USB report for BIOS)
A "Media Controls" endpoint (things like volume up/down)
A "NKRO" endpoint (uses a bitmap so you can press ALL THE KEYS!)

I also made a video of me going through the keyboard and asking questions: https://photos.app.goo.gl/7j98istA1GyRfbjE6

The two questions I have:

I originally used the "USBH Example HID" code as a starting point, but realized that the "USBH_HID_get_report" is a blocking call to only 1 endpoint (which is a no-go - I need to listen to all 3 endpoints and receive from any as an interrupt.) I found out about "USBH_transfer_async" which seems like what I need. I combined it with looping over "USBH_Process" but it isn't yielding me any data in the buffer I'm sending in. I can't find any code examples using "USBH_transfer_async" - how should I be doing this? Figured this and it's working perfectly (although the documentation says NOT to re-call the "USBH_transfer_async" from within the callback, one of the examples does just that so so far it's been working pretty good)
When I do start receiving data, do I need to write unique code to decode all three of those different endpoints? The boot protocol, the volume All I want is the "keys" sent to me. If I do have to write unique code for each EP, this seems like an impossible task then right? There are hundreds of thousands of keyboards and I assume all of them send different kinds of reports. How could I possibly write code to interpret all of them?? Just looked at my Razer Tartarus and it also has an insanely unique way of reporting keys. HELP!!!

I've attached the USB HID data of my keyboard (interfaces, endpoints, etc.) Hope this helps someone answer my question 🙏

Discussion - Software / Setting bMaxPacketSize0 to anything other than "8" on a USB Slave doesn't work

« on: April 15, 2020, 05:48:40 AM »

Hey all, I'm currently working on hardware for the Nintendo Switch and one requirement is that the USB device needs to report back bMaxPacketSize0 of 64.

I'm attaching two files to this forum post: The descriptors of the controller I'm emulating, and my VNC2 solution. You'll notice they match 1:1 except for the bMaxPacketSize0.

Once I set bMaxPacketSize0, the hardware stops working. I've run the software with a debugger, and once I hit the last line below, it dies.

Code: [Select]

hidkbd_ioctl.ioctl_code = VOS_IOCTL_USBSLAVEHID_ATTACH;
hidkbd_ioctl.set.data = (void *)hUSBSLAVE_1;
vos_dev_ioctl(hUSBSLAVE_HID, &hidkbd_ioctl);

It works perfectly fine if I leave bMaxPacketSize0 at 8 (it even works in Windows!) but the Switch requires this to be 64.

Anyone know if this is a possibility with the VNC2? I've spent many hours on this one problem (close to 30) and already purchased a lot of hardware for it so I'm hoping this will work.

I've also looked through every sample and none of them ever modify bMaxPacketSize0 which is scary to me.