Choose which USB and USB HID devices to include at compile-time

[dhalbert]

Make it easy to enable and disable which USB and USB HID devices to include at build time, using two makefile variables. Defaults are:

USB_DEVICES = "CDC,MSC,AUDIO,HID"
USB_HID_DEVICES = "KEYBOARD,MOUSE,CONSUMER,GAMEPAD"

For instance, in mpconfigboard.mk, you can define those variables:

USB_DEVICES = "CDC,MSC"    # Omit MIDI and HID
USB_HID_DEVICES="KEYBOARD"    # Use only keyboard

A couple of new HID devices have been added, but are not in the default set. The entire set is KEYBOARD,MOUSE,CONSUMER,SYS_CONTROL,GAMEPAD,DIGITIZER,XAC_COMPATIBLE_GAMEPAD,RAW

If only one HID device is chosen, then it will not have a report ID in the generated HID descriptor.

XAC_COMPATIBLE_GAMEPAD is an alternate gamepad that is compatible with the Microsoft XAC controller. RAW is a 64-bit raw HID device, usage page 0xFFAF, usage 0xAF (AF is "AdaFruit"). RAW cannot be used with other HID devices, because it cannot have a report id.

The mechanism for generating the table of usb_hid.Device entries has also been simplified. More code is generated and less is done by tricky Makefile variables.

This partially addresses some needs stated in #1015 but does not allow generation of USB or USB HID descriptors at run-time, in boot.py.

Tagging @ATMakersBill and @hierophect for interest.

[hierophect]

Low endpoint devices such as the F411 will still have issues with the descriptor file, as they need the MSC endpoints reduced from 2 to 1.

tandard.EndpointDescriptor(
                description="MSC in",
                bEndpointAddress=0x0 | standard.EndpointDescriptor.DIRECTION_IN,
                bmAttributes=standard.EndpointDescriptor.TYPE_BULK,
                bInterval=0),
            standard.EndpointDescriptor(
                description="MSC out",
                bEndpointAddress=0x1 | standard.EndpointDescriptor.DIRECTION_OUT,
                bmAttributes=standard.EndpointDescriptor.TYPE_BULK,
                bInterval=0)

Could we introduce a flag or other condition for this case?

[hierophect]

Looks good! Thanks for including this.

[ATMakersBill]

Excited to see this! Will test tomorrow (hopefully) and then share w/friends in the XAC community!

[ATMakersBill]

@dhalbert is there a version of Gamepad that matches the XAC gamepad?

[dhalbert]

@ATMakersBill Yes, see the top post. You can do (in, say, mpconfigboard.mk

USB_HID_DEVICES = XAC_COMPATIBLE_GAMEPAD

and get a single XAC gamepad without a report ID. Use your XAC-compatible Gamepad.py. I plan to test this soon.

[ATMakersBill]

Sorry, I meant was there an updated HID library with an XAC gamepad. I can test with mine, but was wondering how that would be added to the libs

[dhalbert]

Nope, I didn't add anything yet - just use your library for now. It will probably be a separate library, because adding it to frozen HID will make HID even bigger and take more flash, which is a problem on CPX.

[ThomasAtBBTF]

Yes, a way to control of which devices the USB composite device consists of is very important!
Also, a second UART CDC would be extremely helpful. I hate to use the primary CDC for communication and give the customer access to the REPL!