Hardware/Software Environment:
- Module: Verdin iMX8M Mini DualLite 1GB WB IT
- OS: TorizonCore 6.8.3+build.37 (kirkstone)
- Kernel: Linux 5.15.183-6.8.3+git.defb81960f4f #1-TorizonCore SMP PREEMPT Tue Jun 3 13:26:20 UTC 2025 aarch64
- Bluetooth/WiFi Module: AzureWave AW-CM276NF (SDIO, Marvell/NXP chipset)
- Application: Python container using BLE via BlueZ over D-Bus (e.g. python-dbus, pydbus)
- Docker Variant: Yes
Problem Description:
After several hours of BLE operation in Central mode (LE scans), the AzureWave/Marvell-NXP SDIO controller enters a frozen state:
- Symptoms:
-
dmesglog shows:Bluetooth: hci0: unknown advertising packet type: 0xff Bluetooth: hci0: command 0x200c tx timeout Bluetooth: hci0: failed to disable LE scan: status 0x1f -
The Python application (using BlueZ via D-Bus) reports:
Can't init device hci0: Connection timed out (110) -
Sometimes
hciconfig hci0still shows “UP RUNNING”, other times it disappears (“No such device”). -
No software reset (
hciconfig hci0 reset,down,up) restores BLE functionality. -
Only a system reboot or power cycle recovers the BLE interface.
-
Questions for Toradex/NXP Support:
- Is there any known software workaround to recover BLE in this situation, without rebooting or power cycling the SOM?
- Is there a firmware fix or update for AzureWave/Marvell-NXP BLE SDIO to prevent this freeze (especially after “unknown advertising packet type” events)?
- Are there recommended monitoring or automated recovery strategies for this specific hardware, OS, and Python/BlueZ D-Bus setup?
Attachments:
-
Log excerpts:
Sep 08 09:19:48 kernel: Bluetooth: hci0: unknown advertising packet type: 0xff Sep 08 09:27:19 kernel: Bluetooth: hci0: command 0x200c tx timeout Sep 08 10:44:36 [Python]: Can't init device hci0: Connection timed out (110) -
Output from
hciconfig hci0,hcitool dev.
Thank you for your support!