Changing Eddystone beacon frame type and data

Question

Hi all, 
 Based on the EddystoneBeacon example (ble_app_eddystone), I try and change the advertising slot contents runtime from main.c on a button press. I read similar questions here in the Devzone, and as expected this pointed me to the usage of es_slot_on_write, but things don't work out as I want it to be. When I write the new data to slot 0, the application resets after reporting a Fatal error. When I write to slot 1, 2 or 3, nothing changes and the application keeps sending the earlier set Default data. Here is what I try to do in the Button event handle: 
 static uint8_t TestData[20] = {
APP_ES_UID_FRAME_TYPE, DEFAULT_FRAME_TX_POWER,
 0x33, 0x33, 0x44, 0x44, 0x55, 0x55, 0x44, 0x44, 0x33, 0x33, //10-byte UID namespace
 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, //6-byte UID ID
 0x00, 0x00
};

/**@brief Function for handling button events from app_button IRQ
 *
 * @param[in] pin_no Pin of the button for which an event has occured
 * @param[in] button_action Press or Release
 */
static void button_evt_handler(uint8_t pin_no, uint8_t button_action)
{
 if (button_action == APP_BUTTON_PUSH && pin_no == BUTTON_1) {
// nrf_ble_es_on_start_connectable_advertising();
 es_slot_on_write(0, 20, TestData);
 }
 if (button_action == APP_BUTTON_PUSH && pin_no == BUTTON_2) {
 es_slot_on_write(1, 20, TestData);
 }
} 
 
 With some debug prints I see it ending up in configure_slot as should be, but it does not work so I must do something wrong: 
 
 Do I need to call extra functions to get this working as should be? 
 Can it be the TestData frame contents, are the test-values for the UID-Frame for some reason invalid? 
 Some other cause? 
 
 I'm using nrf52840-DK and nRF5 SDK v.16.0 
 Thanks in advance for your reply

Slay · Accepted Answer

OK, I think I have figured it out myself: 
 Looking at the code for configure_slot in es_slot.c, it starts copying the data at 
 uint8_t * p_data_after_ranging_data = ((uint8_t *)(&m_reg.slots[slot_no].adv_frame.frame) +
 RANGING_DATA_INDEX + RANGING_DATA_LENGTH);
 
 This is 2 bytes from the frame-start. And the actual copying is done like: 
 case ES_FRAME_TYPE_URL:
 memcpy(p_data_after_ranging_data, &p_frame_data[1], length - 1);
 
 Which means that my TestData is one byte too long and should not include the Tx-Power. If I leave this byte out and pass a length of 19, no Fatal error occurs and the data is sent as should be. So apparently the es_slot_on-write does require a little different frame format as one (or at least I) would expect. For completeness the final (working) code: 
 static const uint8_t TestDataUid[19] = {
	APP_ES_UID_FRAME_TYPE,
	0x33, 0x33, 0x44, 0x44, 0x55, 0x55, 0x44, 0x44, 0x33, 0x33, //10-byte UID namespace
	0x07, 0x07, 0x07, 0x07, 0x07, 0x07, //6-byte UID ID
	0x00, 0x00
};

static void button_evt_handler(uint8_t pin_no, uint8_t button_action)
{
 if (button_action == APP_BUTTON_PUSH && pin_no == BUTTON_1) {
// nrf_ble_es_on_start_connectable_advertising();
 es_slot_on_write(0, 19, TestDataUid);
 }
} 
 Hope this could help someone else. Regards

Changing Eddystone beacon frame type and data

Top Replies