Hi,
I'm using SDK v3.2 and I'm extending the lighting example. I'm developing an extension to the provisioner, to inform a back-end about the configuration of the subnet it is part of. For that, I use the existing serial interface handler. However, I also want to know the Network ID derived from the NetKey. For that, I use the Application serial command. Within that command, I define a more specific opcode, which leads to the underlying code:
static void serial_interface_cb(const uint8_t* p_data, uint32_t length)
{
NRF_MESH_ASSERT(length > 0);
switch(p_data[0])
{
case BLUESS_OPCODE_NETWORK_ID_DISCOVERY_SERIAL_COMMAND:
; // Empty statement
mesh_key_index_t p_key_list[DSM_SUBNET_MAX_LIMIT];
uint32_t p_count = DSM_SUBNET_MAX_LIMIT;
dsm_subnet_get_all(p_key_list, &p_count);
uint32_t length = sizeof(bluess_serial_opcode_t) + p_count * NRF_MESH_NETID_SIZE;
uint32_t index = 0;
uint8_t p_serial[length];
p_serial[index] = BLUESS_OPCODE_NETWORK_ID_DISCOVERY_SERIAL_EVENT;
index += sizeof(bluess_serial_opcode_t);
for(uint32_t i = 0 ; i < p_count ; i++){
nrf_mesh_beacon_secmat_t * p_secmat;
uint8_t key = 0;
uint32_t status = dsm_subnet_key_get(dsm_net_key_index_to_subnet_handle(p_key_list[i]), &key);
if(status == NRF_SUCCESS){
status = nrf_mesh_keygen_beacon_secmat(&key, p_secmat);
}
memcpy(&p_serial[index], &p_secmat->net_id[0], NRF_MESH_NETID_SIZE);
index += NRF_MESH_NETID_SIZE;
}
nrf_mesh_serial_tx(p_serial, length);
break;
}
}
As the provisioner is only part of one subnet, p_count will become equal to 1. This is the case from the start of the for-loop. However, I notice that during the for-loop, suddenly p_count becomes much larger, causing the loop to run endlessly. When commenting out the whole code within the for-loop, the callback is completed as it should, without p_count changing in between. At first glance, it seems that to many method calls within a callback, causes issues... Is there any explanation for this?
Kind regards,
Mathias
