XHC WHB04B development?

I don't think I would trust a USB device under these circumstances. I would recommend you go back to XHC's website and look for a 433 MHz pendant that hard wires in to send/receive unit. After years  of use, these have never let me down 

Do you have a link to the one you use or a model number?

I had a feeling that it could be possible to read out the factory firmware from the dongle due to some bugs in the firmware. And that is indeed the case, at least for a partial readout. It is possible to read the firmware from where the USB device descriptor is stored in program memory until and including the bootloader. I added this tool to the github repository:
github.com/finngineering/whb04b/blob/mas...s/firmware_readout.c
While I think this tool should be safe to use, I give no guarantees. With a partial readout, I could compare that to my firmware, and if it matches, the likelihood that the whole firmware is the same is very high.

Actually, now that I've done this, I remember that there is a much easier way to read out the firmware from CH55x with bootloader version lower than 2.50 (which is the version that my dongle has). The above tool should print the bootloader version, and if it's less than 2.50, I could maybe make a tool to read out the firmware through the bootloader and over USB.

Here is what happens when I run the tool. My whb04b is from the era when they just stopped selling the predecessor. The output from the command is attached,

Thank you for testing. The output indicate that the program was working as expected. The firmware is not exactly the same as mine, but it looks similar. How big differences/similarities there are would require disassembly comparing those. I can see that at least memory addresses are a bit different. Here is the output from mine (not the real dongle, I put the factory firmware on another device which has bootloader version 2.40): The addresses and sizes are close to each other, but not exactly the same.

Because you have bootloader version 2.40, I believe we would be able to read out the firmware through the bootloader. But it requires creating a similar program for that as the one we just used. And to get to bootloader mode, a 10k resistor needs to be added to the pendant. So maybe better to wait until I have tested the new firmware a bit more. And then you could read out the firmware and flash the new one at the same time, if you are still interested at that point.

I used the mill for a couple of hours today with my firmware in the pendant. I tried my best to cause some issues, but no disconnects from the dongle nor any other issues as far as I could tell. To be honest, I'm surprised that it works this well already.

I found a sticker with version information on the back on my pendant (underneath the rubber):
 
The top row is the firmware version and the bottom row the hardware version, I believe. And the RX should be the dongle, because there is a V09 marking on the main board (underneath the RF module).

And I attach a couple of photos from the dongle as well, in case somebody wants to compare. Not the best photos. I took them through the microscope some months ago. And right now I have a thousand wires connected for debugging, so not possible to take good photos at the moment.
 
 

The sticker on my pendant 
 

Had two crashes yesterday. Mode switching. libusb_fill_interrupt_transfer() isn't the solution.
 
I had commented out the assert() in setupAsyncTransfer and let it continue
Failed here  github.com/LinuxCNC/linuxcnc/blob/0673e3...whb04b-6/usb.cc#L541