nrf52840dk halting when using hci_usb sample to connect to multiple devices "ASSERTION FAIL lll_preempt_calc"

[aferm]

Hello,

I'm trying to evaluate using the zephyr hci_usb sample as a temporary solution to connecting to several ble sensors. I have modified the prj.conf to allow multiple connections, and everything appears to work fine with a single device, but a soon as I try to connect a second sensor the hci device locks up.

I am using zephyr-sdk-0.15.1 and the nrf52840dk.

Here are my changes to the prj.conf for the hci_usb sample:

CONFIG_STDOUT_CONSOLE=y
CONFIG_GPIO=y
CONFIG_SERIAL=y
CONFIG_UART_INTERRUPT_DRIVEN=y

CONFIG_BT=y
CONFIG_BT_HCI_RAW=y
CONFIG_BT_MAX_CONN=4

CONFIG_LOG=y
CONFIG_RTT_CONSOLE=y
CONFIG_USE_SEGGER_RTT=y
CONFIG_LOG_BACKEND_RTT=y
CONFIG_LOG_BACKEND_UART=n
CONFIG_UART_CONSOLE=n

CONFIG_USB_DEVICE_STACK=y
CONFIG_USB_DEVICE_BLUETOOTH=y
CONFIG_USB_DEVICE_BLUETOOTH_VS_H4=n

# Workaround: Unable to allocate command buffer when using K_NO_WAIT since
# Host number of completed commands does not follow normal flow control.
CONFIG_BT_BUF_CMD_TX_COUNT=10

And here are the resulting RTT logs:

*** Booting Zephyr OS build zephyr-v3.2.0-882-g1fb23b6509d2  ***
[00:00:00.371,978] <inf> bt_hci_raw: Bluetooth enabled in RAW mode
Bluetooth over USB sample
[00:00:53.757,354] <inf> bt_ctlr_ull_conn: incoming Tx: count= 0, len= 0, rate= 0 bps.
[00:00:54.077,819] <inf> bt_ctlr_ull_conn: force_md_cnt: old= 0, new= 212.
[00:00:54.258,361] <inf> bt_ctlr_ull_conn: force_md_cnt: old= 0, new= 229.
[00:00:54.770,050] <inf> bt_ctlr_ull_conn: force_md_cnt: old= 0, new= 15.
[00:00:54.770,080] <inf> bt_ctlr_ull_conn: incoming Tx: count= 44, len= 422, rate= 3454 bps.
ASSERTION FAIL [0] @ WEST_TOPDIR/zephyr/subsys/bluetooth/controller/ll_sw/nordic/lll/lll.c:476
  lll_preempt_calc: Actual EVENT_OVERHEAD_START_US = 335
[00:00:54.866,638] <err> os: r0/a1:  0x00000003  r1/a2:  0x00000000  r2/a3:  0x00000004
[00:00:54.866,638] <err> os: r3/a4:  0x200008f0 r12/ip:  0x00000010 r14/lr:  0x00001b3b
[00:00:54.866,668] <err> os:  xpsr:  0x41000011
[00:00:54.866,668] <err> os: Faulting instruction address (r15/pc): 0x00001b46
[00:00:54.866,699] <err> os: >>> ZEPHYR FATAL ERROR 3: Kernel oops on CPU 0
[00:00:54.866,729] <err> os: Fault during interrupt handling

[00:00:54.866,760] <err> os: Current thread: 0x20001738 (unknown)
[00:00:55.184,906] <err> os: Halting system

[Øivind]

Hi,

Are you using the nRF Connect SDK? If so which version are you using?

[aferm]

I was using the zephyr SDK version 0.15.0, which I set up by following these instructions. When I ran into the error above.

 https://developer.nordicsemi.com/nRF_Connect_SDK/doc/latest/zephyr/develop/getting_started/index.html

I just tried to build the same sample from the Connect SDK v2.1.1, and am unable to even scan for devices.

Here is a quick grab from btmon. I see this repeatedly when trying to enable scanning.

< HCI Command: LE Set Scan Enable (0x08|0x000c) plen 2                                                                                                                                         #4426 [hci1] 6.736679
        Scanning: Enabled (0x01)
        Filter duplicates: Enabled (0x01)
> HCI Event: Command Complete (0x0e) plen 4                                                                                                                                                    #4427 [hci1] 6.738113
      LE Set Scan Enable (0x08|0x000c) ncmd 1
        Status: Invalid HCI Command Parameters (0x12)