MPU FAULT

[bhargav_gondaliya]

I'm working with nRF Connect SDK (Zephyr). In that I am using 255 byte array to get response from nrf52832 through BLE. my code is getting crash in response back. when I decrease size of array to 50 byte it's working.

but when I compile same in nrf5 sdk code it's working properly with 255 bytes array. Can you provide any solution or workaround to get it work?

Error Log:

*** Booting nRF Connect SDK d96769faceca ***
GPIO pins initialized correctly
[00:00:00.002,777] <inf> adc_module: gpio_init: pin reset low: 0

I2C device ready
ADC device ready
Device MCP7940N is ready.
Device 1 is ready.
[00:00:06.940,979] <inf> bt_sdc_hci_driver: hci_driver_open: SoftDevice Controller build revision:
                                            36 f0 e5 0e 87 68 48 fb  02 fd 9f 82 cc 32 e5 7b |6....hH. .....2.{
                                            91 b1 5c ed                                      |..\.             
[00:00:06.943,237] <inf> bt_hci_core: hci_vs_init: HW Platform: Nordic Semiconductor (0x0002)
[00:00:06.943,267] <inf> bt_hci_core: hci_vs_init: HW Variant: nRF52x (0x0002)
[00:00:06.943,298] <inf> bt_hci_core: hci_vs_init: Firmware: Standard Bluetooth controller (0x00) Version 54.58864 Build 1214809870
[00:00:06.944,122] <inf> bt_hci_core: bt_dev_show_info: Identity: E4:FA:82:90:32:AC (random)
[00:00:06.944,152] <inf> bt_hci_core: bt_dev_show_info: HCI: version 5.4 (0x0d) revision 0x118f, manufacturer 0x0059
[00:00:06.944,183] <inf> bt_hci_core: bt_dev_show_info: LMP: version 5.4 (0x0d) subver 0x118f
Bluetooth initialized
Advertising started
[00:00:06.946,380] <err> MCP7940N: read_time: RC = 0

localtime_buff = Thu 2024-08-29 19:27:49
[00:00:07.946,685] <err> MCP7940N: read_time: RC = -5

[00:00:07.946,685] <err> adc_module: sec_timer_handler: Failed to read time, error: 251

[00:00:08.000,549] <err> os: mem_manage_fault: ***** MPU FAULT *****
[00:00:08.000,549] <err> os: mem_manage_fault:   Data Access Violation
[00:00:08.000,579] <err> os: mem_manage_fault:   MMFAR Address: 0x0
[00:00:08.000,579] <err> os: esf_dump: r0/a1:  0x20002bb0  r1/a2:  0x00000000  r2/a3:  0x00000000
[00:00:08.000,610] <err> os: esf_dump: r3/a4:  0x00000000 r12/ip:  0x00000000 r14/lr:  0x0002130b
[00:00:08.000,610] <err> os: esf_dump:  xpsr:  0x21000221
[00:00:08.000,640] <err> os: esf_dump: Faulting instruction address (r15/pc): 0x000211a2
[00:00:08.000,671] <err> os: z_fatal_error: >>> ZEPHYR FATAL ERROR 19: Unknown error on CPU 0
[00:00:08.000,671] <err> os: z_fatal_error: Fault during interrupt handling

[00:00:08.000,732] <err> os: z_fatal_error: Current thread: 0x20001a78 (unknown)
[00:00:08.439,880] <err> fatal_error: k_sys_fatal_error_handler: Resetting system

 

[Amanda Hsieh]

Hi, 

For the first error log in the figure, the fault message indicates that your device is running out of memory, which is caught by the MPU. Try adjusting CONFIG_MAIN_STACK_SiZE and see if this still occurs.

For the second error log, you can use Addr2line to get the line. See the Fatal errors part in this course and Debugging with addr2line in this course. 

Regards,
Amanda H.

[bhargav_gondaliya]

Hi,Amanda Hsieh 

Thanks for the reply,

I adjust the CONFIG_MAIN_STACK_SIZE but still facing same error. Here you can see the peripheral I initialize in the main for nrf5 SDK and nrf connect sdk both are same configuration. I created a my custom ble service and added tx and rx characteristic. also I checked memory report, RAM 58% used.

size of tx and rx buffers are 255

Initialized in nrf connect sdk

Initialized in nrf5 sdk ( working without any error )

int main(void)
{
    bool erase_bonds;

    // Initialize.
    log_init();
    timers_init();
    pins_init(&erase_bonds);
    // EPD Setup start
    epd_init();
    //displayGuide();
    power_management_init();
    ble_stack_init();
    adc_configure();
    twi_master_init();
    read_eeprom();
   // display_guide();
    gap_params_init();
    gatt_init();
    advertising_init();
    services_init();
    conn_params_init();
    MCP7940_begin();
    MCP7940_now(&current_epoch_time);
    weekly_adherence_calculate();
    nrf_delay_ms(3000);
    MCP7940_now(&current_epoch_time);
    nrf_drv_saadc_sample();

    // Format time, "ddd yyyy-mm-dd hh:mm:ss zzz"
    latest_minute = *localtime(&current_epoch_time);
    latest_minute.tm_year = latest_minute.tm_year - 100;

    // Start execution.
    application_timers_start();
    advertising_start();

  for(;;){

  }

}

[Amanda Hsieh]

What value of CONFIG_MAIN_STACK_SIZE did you try?

Please enable the following configs and provide the log again.

• CONFIG_DEBUG_OPTIMIZATIONS – Limits the optimizations done by the compiler to only those that do not impact debugging.
• CONFIG_DEBUG_THREAD_INFO – Adds additional information to the thread object so the debugger can discover the threads.

[bhargav_gondaliya]

Hi Amanda,

sorry for the late reply.
I checked CONFIG_MAIN_STACK_SIZE with different values. max valued was 8192. but still facing same issue. I have created custom BLE service and added RX and TX characteristic for sending data from nrf connect app to nrf52832 device.

Below is my custom service, that I define

BT_GATT_SERVICE_DEFINE(my_service,

BT_GATT_PRIMARY_SERVICE(BT_UUID_MY_SERVICE),

BT_GATT_CHARACTERISTIC(BT_UUID_MY_SERVICE_RX,BT_GATT_CHRC_WRITE | BT_GATT_CHRC_WRITE_WITHOUT_RESP,

BT_GATT_PERM_READ | BT_GATT_PERM_WRITE,

NULL, on_receive, NULL),

BT_GATT_CHARACTERISTIC(BT_UUID_MY_SERVICE_TX,

BT_GATT_CHRC_NOTIFY,

BT_GATT_PERM_READ,

NULL, NULL, NULL),

BT_GATT_CCC(on_cccd_changed,

BT_GATT_PERM_READ | BT_GATT_PERM_WRITE),

);

for your understanding.

I send a command 0201 and 4bytes user id from nrf connect app to a nrf52832 device and I got successfully response from device.

but got an error of buffer overflow detected.

* buffer overflow detected *

[00:00:45.221,405] <err> os: r0/a1:  0x00000002  r1/a2:  0x0000000b  r2/a3:  0x0000000a

[00:00:45.221,435] <err> os: r3/a4:  0x00000004 r12/ip:  0x00000004 r14/lr:  0x000178b5

[00:00:45.221,435] <err> os:  xpsr:  0x61000000

[00:00:45.221,435] <err> os: Faulting instruction address (r15/pc): 0x000178c0

[00:00:45.221,496] <err> os: >>> ZEPHYR FATAL ERROR 2: Stack overflow on CPU 0

[00:00:45.221,527] <err> os: Current thread: 0x20002368 (BT RX)

[00:00:45.698,181] <err> os: Halting system

Due to above error my device is getting disconnect from application.

also facing GATT errors. ERROR 133(0x85): GATT ERROR and ERROR 22(0x16): GATT CONN TERMINATE LOCAL HOST

below is my switch case that will be execute when I send command.

for response to client(application)

also Increased RX & TX stack size in prj.conf file.

provide a solution if any.

[Amanda Hsieh]

bhargav_gondaliya said:

[00:00:45.221,496] <err> os: >>> ZEPHYR FATAL ERROR 2: Stack overflow on CPU 0

[00:00:45.221,527] <err> os: Current thread: 0x20002368 (BT RX)

The log indicated that the BT RX thread caused the fault.

Try increasing the stack size for that thread, using for instance CONFIG_BT_RX_STACK_SIZE=4096 in your prj.conf. 

[bhargav_gondaliya]

thread analyze logs.

main stack usage is 536 out of 4096 cpu : 99% why?

I increased CONFIG_BT_RX_STACK_SIZE to 4096 still facing same error.