usbfs: add support for the USBFS peripheral

[rmsyn]

Adds support for the USB full speed device peripheral.

Implements the Device function using the usb-device crate.

Creates a new usbfs module to allow for adding the OTG and Host functions in later work.

Depends on #56

[rmsyn]

Moving this work to a draft PR.

Currently, having issues with getting the device to detect VBUS line. The device will power on when the VBUS wire is plugged directly into a 3.3V/5V power line, but not the VBUS line.

There is also a read error when the device gets detected by the host, and I still need to make the TX FIFO sizes configurable based on the type of device.

I'll continue debugging, and update the PR with fixes. Any help/advice is greatly appreciated.

[Disasm]

Hi! There is a PR in a different repo which adds support for this chip: stm32-rs/synopsys-usb-otg#36 It would be nice to re-use it, but testing is needed for different chips to avoid regressions. WBR, Vadim Kaushan

…

On 18 June 2023 08:12:53 EEST, rmsyn ***@***.***> wrote: Moving this work to a draft PR. Currently, having issues with getting the device to detect `VBUS` line. The device will power on when the `VBUS` wire is plugged directly into a 3.3V power line, but not the `VBUS` line. There is also a read error when the device gets detected by the host, and I still need to make the TX FIFO sizes configurable based on the type of device. I'll continue debugging, and update the PR with fixes. Any help/advice is greatly appreciated. -- Reply to this email directly or view it on GitHub: #57 (comment) You are receiving this because you are subscribed to this thread. Message ID: ***@***.***>

[rmsyn]

It would be nice to re-use it, but testing is needed for different chips to avoid regressions.

So, I rebased the work in #50 on the latest gd32vf103-pac, and did some minor updates to get it to work with the longan-nano USBFS example. It still doesn't work, it actually is less functional (it doesn't respond to Setup Address commands). edit: The issues I was noticing seem to be from something else in the stack.

edit: So, I think something else is the issue with my example in longan-nano. I tested against your example in https://github.com/Disasm/usb-otg-workspace, and it seems to work fine. Going to try using a full HIDClass example based on yours, see if I can actually get it to send keyboard reports to the host.

I still can't get the VBUS line to power the device. Working on getting the hardware setup to proxy the device, because the device doesn't register the vid/pid (via lsusb), and I can only get read errors/setup address errors by plugging the D+/D- lines into the wrong pins (errors via dmesg).

Added an HID example to usb-otg-workspace, but it also doesn't register as a USB device via lsusb. Tried the test-class example, but no luck there either.

It feels like I'm missing something. I am using a USB-C breakout board to connect VBUS, GND, D-, and D+ to the longan-nano's 5v, GND, PA11, and PA12 pins (alternatively trying PA9 for VBUS). Nothing produces a listing on the host...