RE: Flash data in Hardfault handler

Vidar Berg — Thu, 18 Jul 2019 13:20:39 GMT

I would just hard code a pointer to an unused part of RAM since this for debugging. You can check the *.map file in your build directory to find a free area between the globals and heap/stack

RE: Flash data in Hardfault handler

hongpal — Thu, 18 Jul 2019 12:56:19 GMT

Thank for your answer.

However, I am using eclipse with gcc.

Would you re-write the way ?

RE: Flash data in Hardfault handler

Vidar Berg — Thu, 18 Jul 2019 12:51:11 GMT

Hello,

Depending on the fault it may be difficult to reliably write the crash log to flash. I think it's easier to write the information to a non-initialized section of RAM, perform a reset, then on subsequent boot, check the RAM content and write it to flash.

If you use Segger embedded Studio you can use the ".non_init" section to place the data.

e.g.,

static char error_info[128] __attribute__((section(".non_init")));
static uint16_t crc __attribute__((section(".non_init")));

/* Call this function on startup to see if there is any new error 
 *  information available
 */
void check_error_info()
{
    uint32_t len;
    
    if ( crc == crc16_compute(error_info, sizeof(error_info), NULL))
    {
        crc = 0;
        len = strlen(error_info);
        if (len > sizeof(error_info)) len = sizeof(error_info);
        //store error information to flash
        //TODO: write error info to flash 
        NRF_LOG_INFO("%s", error_info);
    }
}

/*Read error information from flash */
void get_error_info(uint32_t * p_data)
{  
    *p_data = (uint32_t) error_info;
    memset(error_info, 0, sizeof(error_info));
    // flash_read(p_data, p_src)
}