Pi Zero 2W Freezes after about 30 seconds #134

Activity

1. chazzawalla

   added

   bugSomething isn't workingSomething isn't working

   on Feb 8, 2024
2. danielskowronski commented on Feb 7, 2025

   

   danielskowronski

   on Feb 7, 2025

   More actions

   I had the same problem with Raspberry Pi 3B+ running Ubuntu 24.04 64-bit when interfacing Stream Deck Neo. It was running for several seconds, then all subsequent runs were failing with StreamDeck.Transport.Transport.TransportError and only RPi reboot could help.

   It turns out that Ubuntu images ship with dwc2,dr_mode=otg as default setting instead of dwc2,dr_mode=host (default on Raspbian), which impacts the onboard USB controller (DWC2) ability to handle large HID Feature reports.

   To solve the issue, you need to append dtoverlay=dwc2,dr_mode=host to /boot/firmware/config.txt and reboot.

   ---

   A few words on my debugging, maybe something will help maintainers.

   Output from failing example_neo.py usually looks like that (but sometimes it can pass reset and fail on first write command):

   Found 1 Stream Deck(s).
   
   Traceback (most recent call last):
     File "/home/daniel/streamdeck/example_neo.py", line 145, in <module>
       deck.reset()
     File "/usr/local/lib/python3.12/dist-packages/StreamDeck/Devices/StreamDeckNeo.py", line 120, in reset
       self.device.write_feature(payload)
     File "/usr/local/lib/python3.12/dist-packages/StreamDeck/Transport/LibUSBHIDAPI.py", line 400, in write_feature
       return self.hidapi.send_feature_report(self.device_handle, payload)
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
     File "/usr/local/lib/python3.12/dist-packages/StreamDeck/Transport/LibUSBHIDAPI.py", line 260, in send_feature_report
       raise TransportError("Failed to write feature report (%d)" % result)
   StreamDeck.Transport.Transport.TransportError: Failed to write feature report (-1)
   python3: ../../libusb/os/threads_posix.h:46: usbi_mutex_lock: Assertion `pthread_mutex_lock(mutex) == 0' failed.
   python3: ../../libusb/os/threads_posix.h:58: usbi_mutex_destroy: Assertion `pthread_mutex_destroy(mutex) == 0' failed.
   Aborted (core dumped)
   

   Those assertions failures (last 3 lines) are not the root cause, rather symptom of improper exit. I suspect there's something out of order in handling HID device closure, but I know just a few things about USB. Anyway, removing self.HIDAPI_INSTANCE.hid_exit() didn't impact Failed to write out report (-1), but assertions were gone. This could be fixed, as it's easy to trigger an ugly error with Ctrl+C even if things still work.

   I also tried making payloads of hidapi.send_feature_report and hidapi.write truncated, but writes were always failing.

   The problem is not with the device, I could make it fail on RPi, disconnect it, connect to some other computer with Elgato software, run it there, reconnect back to RPi, and it'd behave like before.

   OP has "Stream Deck original V2" which seems to be "Stream Deck Mk.2" and I guess it may be similar to my Neo (both use USB-C). Thus, both devices may be similarly too new or too heavy for OTG mode.

   If needed, I could do some USB traffic sniffing to help further, but for my scenario it's enough to keep dwc2,dr_mode=host (dwc_otg does not work on my board).

   Some more context on https://raspberrypi.stackexchange.com/a/77061