System crash when both use softdevie and spi

RE: System crash when both use softdevie and spi

Kevin — Thu, 06 Apr 2017 12:54:13 GMT

Hi Kristin,

I debug using the method that you suggested. And the bug occurred when spi transfer too slow that appeared 0xFF in received data.

It will overflow and makes crash. After i give the client high priority and check data should not be 0xFF. It looks working normal until now.

I appreciate your kind assistance!!!

RE: System crash when both use softdevie and spi

FormerMember — Thu, 06 Apr 2017 08:44:22 GMT

Could you upload a screenshot of the Keil window when the hardfault occur?

RE: System crash when both use softdevie and spi

FormerMember — Wed, 05 Apr 2017 11:31:05 GMT

This app note from Keil, section "Debugging Faults with uVision" can be useful for determining where the hardfault occurs.

RE: System crash when both use softdevie and spi

Kevin — Wed, 05 Apr 2017 06:26:00 GMT

Hi Kristin,

I don't know how to see the hardfault originates from. I set break point at HardFault_Handler, i don't see any other message. Is there any example or method?

I have tried debug, and it occurred at a strange place.

I called a function in timer, i can see the debug print in timer. But i don't see the debug print in called function.

the code as follow:

In timer:

{

NRF_LOG_INFO("Send spi data\r\n");

send_spi_command(type, command, data, 128);

NRF_LOG_INFO("Send spi data\r\n");

}

static uint8_t send_spi_command(uint8_t type, uint8_t command, uint8_t* data, uint8_t length) { NRF_LOG_INFO("Command : %d(%d)\r\n", command, spi_send_buf);

memset(data_tmp[spi_send_buf - 1], 0xFF, 256);

data_tmp[spi_send_buf - 1][0] = spi_send_buf;

data_tmp[spi_send_buf - 1][1] = type;

data_tmp[spi_send_buf - 1][2] = command;

if (data != NULL) {

    memcpy(&data_tmp[spi_send_buf - 1][3], data, length);

}

spi_send_counter++;

return 0;

}

static void spi_schedule_handler(void * p_context)

{

if(spi_send_counter && spi_xfer_done) {

    memset(spi_rx_buf, 0, spi_length);

    nrf_drv_spi_transfer(&spi, data_tmp[spi_send_buf_count - 1], spi_length, spi_rx_buf, spi_length);

    spi_xfer_done = false;

    spi_send_counter--;

    goto end;
}

}

I call the function (send_spi_command) to send data, and i print message before and after the command. It only print the first message.

And it did't print NRF_LOG_INFO("Command : %d(%d)\r\n", command, spi_send_buf);

Thanks!!!

RE: System crash when both use softdevie and spi

FormerMember — Mon, 03 Apr 2017 08:56:46 GMT

If you run the chip in debug mode, and set a breakpoint in the hardfault handler which location does the hardfault originates from? (When doing this, do not single step, but make the application run until it hits the hardfault.)

RE: System crash when both use softdevie and spi

Kevin — Fri, 31 Mar 2017 13:31:57 GMT

Hi Kristin,

I print message and it always stoped after timer or spi handler. If I use debug mode step by step trace, it always happened fatal error.

I have searched, but I didn't see the same similar case. No one use timer to do long data transfer(200 bytes) and use both spi master and softdevice

Thanks!!

RE: System crash when both use softdevie and spi

FormerMember — Fri, 31 Mar 2017 12:46:30 GMT

At which location does your application hardfault?

Do you have tried to search the forum for a similar case?