Unknown function at 0xA60

RE: Unknown function at 0xA60

whoosh — Sat, 26 Jun 2021 16:14:23 GMT

Just a heads up, I found someone having the exact same problem as me. A more detailed problem description is that when I attempted to connect using NRF Connect Mobile app, I was unable to connect due to Error 133 (0x85): Gatt error.

I realized by tracking local variables that the evt_handler variable was 0x00 upon entry into this code. I resolved the issue as described in this previous post: devzone.nordicsemi.com/.../error-133-0x85-gatt-error-with-custom-service-tutorial

RE: Unknown function at 0xA60

whoosh — Fri, 25 Jun 2021 14:30:59 GMT

I am expecting that when I connect to the SoftDevice using the mobile NRF Connect application, it will actually connect and display some of the client data such as Services, Characteristics, etc.. In the current state, the connection does not get established.

I'll continue to add more logging in my code to better root cause the issue per your suggestion.

Thanks!

RE: Unknown function at 0xA60

Einar Thorsrud — Fri, 25 Jun 2021 13:59:49 GMT

What do you see that points to an issue? Don't get me wrong, I am not saying there is no issue in your code (I have no knowledge about it). But the fact that you see the debugger complain about "Unknown function at 0xA60" is not an indication of an issue. And as stated, you cannot step in the code when using a SoftDevice.

To debug if you see something unexpected I suggest you enable debug logging. If you are using SES, then select the Debug build target and test. In most cases the debug log will show you if an error has been detected. You can also add more logs in your code to see what happens.

RE: Unknown function at 0xA60

whoosh — Fri, 25 Jun 2021 13:54:27 GMT

Hello, thank you for your reply.

I've read this suggestion from similar posts. The issue still persists even when running the NRF52 DK without Segger IDE. In other words, even in the circumstances where I am not stepping through the SoftDevice, I am getting the same problem (on connect, the mobile application just shows a blank client tab). That leads me to believe that this might be an actual issue with the code, but I could be mistaken.

RE: Unknown function at 0xA60

Einar Thorsrud — Fri, 25 Jun 2021 09:43:53 GMT

Hi,

The reason the function is unknown is that this is within the MBR, and the MBR and SoftDevice is only provided as binary without debug symbols, so your debugger cannot make sense of it. That is not an error, it is just how it is.

Stepping while the SoftDevice is enabled is not supported and will cause problems. If you do this while advertising or in a BLE connection the pause will make the SoftDevice assert when it continues. So to do something similar with a SoftDevice you need to use a breakpoint and run to the breakpoint. I you want to continue a bit / step, then you instead need to move the breakpoint and reset.

RE: Unknown function at 0xA60

whoosh — Fri, 25 Jun 2021 03:40:24 GMT

/**@brief Function for handling the Connect event.
 *
 * @param[in]   p_cus       Custom Service structure.
 * @param[in]   p_ble_evt   Event received from the BLE stack.
 */
static void on_connect(ble_cus_t * p_cus, ble_evt_t const * p_ble_evt)
{
    p_cus->conn_handle = p_ble_evt->evt.gap_evt.conn_handle;

    ble_cus_evt_t evt;

    printf("ble_cus.c on_connect entry\r\n");   
    evt.evt_type = BLE_CUS_EVT_CONNECTED;
    printf("ble_cus.c: Set event type to BLE_CUS_EVT_CONNECTED\r\n");
    p_cus->evt_handler(p_cus, &evt); //PROBLEM HERE
    printf("ble_cus.c on_connect exit\r\n");
}

I was able to narrow down the problem to the following line of code, (line 15) but still need help in understanding what 0xA60 means.