P2 Eval --> Microsoft Serial BallPoint?

Comments

• cgracey Posts: 14,133

  2019-01-30 22:50

  To solve this, you need to get a chainsaw and cut your motherboard in half, preserving the part with the CPU. The remaining logic will figure out what is left and adapt, accordingly. This also gives malware less places to hide.
• Rayman Posts: 13,893

  2019-01-30 23:25

  Seems this used to happen to GPS devices with USB interfaces too:
  https://answers.microsoft.com/en-us/windows/forum/windows_7-hardware/windows-detects-usb-gps-as-serial-ballpoint-please/e0e03b9b-e9ae-4645-8b3c-5754f06ec3b5
• Rayman Posts: 13,893

  2019-01-30 23:31

  Here's a more recent incident:
  https://www.cnet.com/forums/discussions/jumping-cursor-in-windows-10/
• localroger Posts: 3,451

  2019-01-30 23:55

  This can happen if the computer is booted while any serial device that transmits continuously is connected. The Ballpoint driver init misinterprets the data stream as the output of a mouse and installs the driver, both blocking that port for other uses and screwing up the real mouse movement. The solution is to right-click the ballpoint mouse device in Device Manager and disable it.
• Cluso99 Posts: 18,069

  2019-01-31 00:09

  Are you using a wireless mouse?
  When my mouse batteries need replacing the mouse starts to jump occasionally.
• Rayman Posts: 13,893

  2019-01-31 00:10

  I think you're right, that should do it...
  Having Cordic Crickets running when I rebooted today was maybe what started this...
  
  It actually took me several tries to reboot before it would start. It would start and then immediately shut down. Hopefully, this is a related event...
• localroger Posts: 3,451

  2019-01-31 03:10

  Cordic Crickets blows debug data out the programming port. Not a problem for programming because the downloader force-resets the P2. But if it's plugged in when you boot your PC, the FTDI driver loads first, then the Ballpoint driver sees what it thinks is a serial port with incoming "mouse" data and installs its driver on that serial port, munging your mouse and blocking the USB-serial port from further use.
  
  I've seen this hundreds of times. Industrial measuring devices which transmit continuous serial data have had this problem since the early 1990's. What I can't figure out is why the driver for serial mice is still part of Windows at all, but it is. Even though your device (the P2 Eval) is "USB" it's really a "USB to serial" converter, and once ITS driver is installed as far as your computer is concerned it's a serial port, and THEN the ballpoint driver comes along and can say HEY A MOUSE!
  
  Fortunately disabling the Ballpoint driver in Device Manager solves it. Unfortunately there's no way to automate the fix, and it always grabs my customers who swap out a computer without knowing about the trap.
• Rayman Posts: 13,893

  2019-01-31 11:16

  I think I've never noticed this before because I'm usually powering the Prop boards by the computer's USB port. This time, it's powered by a monitor with USB ports, so it's on when the computer is off...
  
  I think it's tied to FTDI chips and maybe to that 115200 baud rate.
  
  If we knew the protocol, I think we could use this to communicate without the FTDI drivers...
  
  
• localroger Posts: 3,451

  2019-01-31 14:38

  It's not tied to either the FTDI drivers or the baud rate. I've seen the ballpoint driver attach itself to real serial ports receiving data at 9600,n,8,1. Any installation which is blowing continuous serial data at the PC as it boots is vulnerable. Of course, one solution is to make sure the device isn't plugged in while the computer boots, but in a lot of cases this is impractical.
• Rayman Posts: 13,893

  2019-01-31 15:02

  It's amazing I haven't run into this before...