Hard Fault Zephyr during scan with shell BLE command

RE: Hard Fault Zephyr during scan with shell BLE command

Hung Bui — Tue, 30 Aug 2022 10:31:38 GMT

Hi Pierre-Yves,
Yes they are basically the same.
However, when you do BLE you would need to request the tx_power/output_power change to the BLE controller, you can't access the radio directly.

Please follow the code on how to change tx_power using HCI command in the case I pointed to.

RE: Hard Fault Zephyr during scan with shell BLE command

Piwy — Tue, 30 Aug 2022 06:44:42 GMT

Thanks you for your reply.

But the radio_test sample can change output_power, is it the same as tx_power ?

regards,

Pierre-Yves

RE: Hard Fault Zephyr during scan with shell BLE command

Hung Bui — Mon, 29 Aug 2022 09:26:55 GMT

Hi Pierre,

For those command that's not available with bt shell, I would suggest to add your own shell commands which trigger your own function in the code.

The code to change TX Power, you can find here.

RE: Hard Fault Zephyr during scan with shell BLE command

Piwy — Mon, 29 Aug 2022 06:33:09 GMT

[quote userid="116326" url="~/f/nordic-q-a/91274/hard-fault-zephyr-during-scan-with-shell-ble-command"]another [/quote]

I want to do with NCS, a pc-ble-driver like software with shell command, that's why i want to be able to configure the radio directely, but the radio command like output power or data rate seems to works without the radio_cmd_init().

Maybe there is a conflict between the radio_cmd_init() and the configuration use for the scan.

Futhermore i want to have the possibility to add cmd() like update the peripheral software with DFU.

RE: Hard Fault Zephyr during scan with shell BLE command

Hung Bui — Fri, 26 Aug 2022 13:12:06 GMT

Hi Pierre,
Sorry that I didn't see your second reply.

The line
SYS_INIT(radio_test_init, APPLICATION, CONFIG_KERNEL_INIT_PRIORITY_DEVICE);

Will execute the radio_cmd_init() when the device booting up, right after the kernel is finished booting.

I don't think it's a good idea to combine the radio test with the bluetooth shell.

Could you let me know what exactly you want to achieve here ?

RE: Hard Fault Zephyr during scan with shell BLE command

Hung Bui — Fri, 26 Aug 2022 12:56:41 GMT

Hi Pierre-Yves,

I'm taking over the case from Jared as he is on vacation.

1. I'm not sure I'm entirely understand what you say. But could you confirm that after you disable MPSL you still see the same HARD FAULT error ?

3. 4. So my understanding is that only after you integrate the radio test example into the example you see the issue. We would need to look at what you do to integrate the 2 example together. I don't think they are built to be used together as both will try to access the radio.

5. Yes, by updating zephy using west update.

RE: Hard Fault Zephyr during scan with shell BLE command

Piwy — Fri, 26 Aug 2022 12:39:33 GMT

Hi,

I find the issue it was this :

static int radio_cmd_init(const struct device *dev)
{
	ARG_UNUSED(dev);

	return radio_test_init(&test_config);
}

SYS_INIT(radio_cmd_init, APPLICATION, CONFIG_KERNEL_INIT_PRIORITY_DEVICE);

But can you explain why tyhis create an assert ? the systeme work without it....

RE: Hard Fault Zephyr during scan with shell BLE command

Piwy — Fri, 26 Aug 2022 12:26:42 GMT

Hi,

1. Yes i try options, but disable it change anything to the issue, i disable it and will see in the furture

2. I use a custom board, but it's the same as nRF52840 DK, i use uart to communicate with serial port and not usb, in the futur i will need to use USB for the communication instead of UART. I saw that it's possible but it's not the issue here.

3. Yes i use the shell sample from ncs/zephyr/tests/bluetooth/shell and i mix it with radio_test sample from nordic.

4. I just use bt scan on, the same command as you..., but the issue occurs not directly, it occured 5-10 second after command and later if i do a coded scan ( there is less device in coded), I thought the fact that the assert came later in coded is related to the number of device detected in the scan.

But now ii tried just ncs/zephyr/tests/bluetooth/shell and it's working, i don't understand why the part of radio test creat this assert.... ?

I have exactly the same init as you (exept the identity but it's normal)

5. What do you call master branch, update of zephyr ?

regards,

RE: Hard Fault Zephyr during scan with shell BLE command

Jared — Fri, 26 Aug 2022 10:28:03 GMT

Hi,

I discussed this with our developers and here is the feedback:

Did you enable LLPM intentionally? Do you use it in your application? Not sure if this will work with shell.
Which board are you using? nRF52840 DK, Dongle or custom board?
Exactly which sample are you using? Are you using the shell sample from ncs/zephyr/tests/bluetooth/shell?

We tried to reproduce your issue by using shell sample and your kconfig and it didn't assert. Can you specify what shell commands that you have used, so to that we can use the same when we try to reproduce your issue? (see below for example)

ncs-int: nrf] west update
=== updating zephyr (zephyr):
M   tests/bluetooth/shell/prj.conf
HEAD is now at 0a904d0c17 [nrf fromlist] tfm: Add option for implementation custom reset handler with TF-M
=== updating mcuboot (bootloader/mcuboot):
HEAD is now at 129b6312 [nrf noup] bootutil: loader: Add check for netcore addr if NSIB enabled
=== updating mbedtls-nrf (mbedtls):
HEAD is now at b36261c55 [nrf noup]: Fix unset key bits for nrf_cc3xx v0.9.14
=== updating nrfxlib (nrfxlib):
HEAD is now at 1296d445 doc: version update for NCS 2.0.0

$ west build -b nrf52840dk_nrf52840
$ west flash
$ sudo minicom -D /dev/ttyACM9
uart:~$ bt init 
Bluetooth initialized
Settings Loaded
[00:00:11.907,928] <inf> fs_nvs: 8 Sectors of 4096 bytes
[00:00:11.907,958] <inf> fs_nvs: alloc wra: 0, fa0
[00:00:11.907,989] <inf> fs_nvs: data wra: 0, 64
[00:00:11.908,630] <inf> sdc_hci_driver: SoftDevice Controller build revision: 
                                         33 78 2a 18 20 f5 61 61  a6 8b 77 60 62 83 39 2a |3x*. .aa ..w`b.9*
                                         7c f1 14 e4                                      ||...             
[00:00:11.913,635] <inf> bt_hci_core: HW Platform: Nordic Semiconductor (0x0002)
[00:00:11.913,665] <inf> bt_hci_core: HW Variant: nRF52x (0x0002)
[00:00:11.913,665] <inf> bt_hci_core: Firmware: Standard Bluetooth controller (0x00) Version 51.10872 Build 1643454488
[00:00:11.914,367] <inf> bt_hci_core: No ID address. App must call settings_load()
[00:00:11.944,610] <inf> bt_hci_core: Identity: D3:E8:54:AA:01:B0 (random)
[00:00:11.944,641] <inf> bt_hci_core: HCI: version 5.3 (0x0c) revision 0x1136, manufacturer 0x0059
[00:00:11.944,641] <inf> bt_hci_core: LMP: version 5.3 (0x0c) subver 0x1136
uart:~$ bt scan on
Bluetooth active scan enabled
[DEVICE]: 5D:9F:44:D9:1B:B2 (random), AD evt type 0, RSSI -51  C:1 S:1 D:0 SR:0 E:0 Prim: LE 1M, Secn: No packets, Interval: 0x0000 (0 ms), SID: 0xff
[DEVICE]: 5D:9F:44:D9:1B:B2 (random), AD evt type 4, RSSI -49  C:1 S:1 D:0 SR:1 E:0 Prim: LE 1M, Secn: No packets, Interval: 0x0000 (0 ms), SID: 0xff
[DEVICE]: C1:1B:F3:1F:67:3D (random), AD evt type 0, RSSI -75  C:1 S:1 D:0 SR:0 E:0 Prim: LE 1M, Secn: No packets, Interval: 0x0000 (0 ms), SID: 0xff
[DEVICE]: C1:1B:F3:1F:67:3D (random), AD evt type 4, RSSI -74  C:1 S:1 D:0 SR:1 E:0 Prim: LE 1M, Secn: No packets, Interval: 0x0000 (0 ms), SID: 0xff
[DEVICE]: 38:E9:18:57:FA:9F (random), AD evt type 3, RSSI -57  C:0 S:0 D:0 SR:0 E:0 Prim: LE 1M, Secn: No packets, Interval: 0x0000 (0 ms), SID: 0xff
[DEVICE]: C1:1B:F3:1F:67:3D (random), AD evt type 0, RSSI -71  C:1 S:1 D:0 SR:0 E:0 Prim: LE 1M, Secn: No packets, Interval: 0x0000 (0 ms), SID: 0xff
...
...

There was a fix for a similar bug that has been added to the Master branch. Can you try to reproduce your issue using the Master branch?

regards

Jared

RE: Hard Fault Zephyr during scan with shell BLE command

Piwy — Thu, 25 Aug 2022 11:55:50 GMT

Hi,

I use NCS version 2.0.0.

Here there is my Config.

CONFIG_CONSOLE_HANDLER=y
CONFIG_SHELL=y
CONFIG_FLASH_SHELL=y

CONFIG_DYNAMIC_INTERRUPTS=y

CONFIG_NRFX_TIMER0=y

# Enable the BT NUS shell
CONFIG_SHELL_BT_NUS=y

CONFIG_SHELL_BACKEND_SERIAL=y
#CONFIG_HAS_HW_NRF_RADIO_IEEE802154=y

CONFIG_BT_MAX_CONN=5
CONFIG_MPSL=y
#CONFIG_TEST=y
#CONFIG_TEST_LOGGING_DEFAULTS=n
CONFIG_BT=y
CONFIG_BT_DEBUG_LOG=y
CONFIG_BT_CENTRAL=y
#CONFIG_BT_PERIPHERAL=y
CONFIG_BT_PRIVACY=y 
CONFIG_BT_SMP=y
CONFIG_BT_SIGNING=y
#CONFIG_BT_FIXED_PASSKEY=y
#CONFIG_BT_ATT_PREPARE_COUNT=32

CONFIG_BT_GATT_DM=y
CONFIG_BT_GATT_CLIENT=y

CONFIG_BT_L2CAP_DYNAMIC_CHANNEL=y

#CONFIG_BT_TINYCRYPT_ECC=y

CONFIG_BT_SHELL=y
CONFIG_BT_DEVICE_NAME="test shell"
CONFIG_BT_DEVICE_NAME_DYNAMIC=y
CONFIG_BT_L2CAP_TX_BUF_COUNT=10
CONFIG_BT_CTLR_EXT_SCAN_FP=y

CONFIG_BT_EATT=y
CONFIG_BT_L2CAP_ECRED=y

CONFIG_BT_GATT_DYNAMIC_DB=y
#CONFIG_BT_HRS=y
CONFIG_BT_FILTER_ACCEPT_LIST=y
CONFIG_BT_REMOTE_INFO=y
CONFIG_BT_REMOTE_VERSION=y
CONFIG_BT_GATT_AUTO_DISCOVER_CCC=y

CONFIG_BT_SETTINGS=y
CONFIG_FLASH=y
CONFIG_FLASH_PAGE_LAYOUT=y
CONFIG_FLASH_MAP=y
CONFIG_NVS=y
CONFIG_SETTINGS=y
CONFIG_BT_CTLR_PHY_CODED=y
CONFIG_BT_CTLR_SYNC_PERIODIC=y

CONFIG_BT_SCAN=y
CONFIG_BT_SCAN_FILTER_ENABLE=y
CONFIG_BT_SCAN_UUID_CNT=1



CONFIG_BT_EXT_ADV=y
CONFIG_BT_EXT_ADV_LEGACY_SUPPORT=y
CONFIG_BT_PER_ADV=y
CONFIG_BT_PER_ADV_SYNC=y

CONFIG_BT_USER_DATA_LEN_UPDATE=y
CONFIG_BT_AUTO_DATA_LEN_UPDATE=y

CONFIG_BT_USER_PHY_UPDATE=y
#CONFIG_BT_AUTO_PHY_UPDATE=y

#CONFIG_BT_ISO_BROADCASTER=y
#CONFIG_BT_ISO_SYNC_RECEIVER=y
#CONFIG_BT_ISO_UNICAST=y


#CONFIG_MAIN_STACK_SIZE=4096


CONFIG_BT_CTLR_ADV_EXT=y
CONFIG_BT_LL_SW_SPLIT=y
CONFIG_BT_BROADCASTER=y
CONFIG_BT_OBSERVER=y
#CONFIG_DESKTOP_BLE_QOS_ENABLE=y
CONFIG_BT_LL_SOFTDEVICE=y
#CONFIG_QOS=y
#CONFIG_ZERO_LATENCY_IRQS=y
CONFIG_BT_CTLR_SDC_LLPM=y
CONFIG_BT_CTLR_SDC_SCAN_BUFFER_COUNT= 20

The only things that don't work is the scan.

I tried to increase the stacks but it don't change anything.

Reagrds

RE: Hard Fault Zephyr during scan with shell BLE command

Jared — Thu, 25 Aug 2022 11:45:25 GMT

Hi there,

Which NCS version are you using, master branch?

The assert indicates that something is not correct with your configs, can you share your project config file?

[quote user=""]And i have another question on another topic, it's about the name of a peripheral device, the name appear when the peripheral is in 1M PHY but he don't apprear when the device is in Coded PHY, do you know why ?[/quote]

Can you make a new ticket for this problem?

regards

Jared