RFD Binding Table

Hi YiKai, thanks, please see the posts above, they are JPG files. The one shows the decoded values, and the other shows the raw values. I tried to upload the original workspace, but the browser would not let me because of the file extension. 

Hi YiKai, thanks for the help.

There are still 2 strange things happening with my code, if you could spare a few moments to read about them and comment, that would be greatly appreciated!:

1- For the ZDP_BindReq() function, it requires the zAddrType_t for the address parameters. When I declare & use this type and attempt to set the shortAddress parameter, the 2 LSB (Least Significant Bytes) of the IEEE address are replaced with the short address! They seem to have the same address. It does this even when I edit it in debug mode with the watch window, which shows that the addresses are the same.

The function still sends without errors, which is surprising because the address mode is set to afAddr64Bit, so I'd expected that the IEEE would have to be correct. Whats also strange is that it also sends a NWKAddrReq packet for the erroneous IEEE address.

Im not sure whats happening here, any comments / suggestions?

2- In the *_event_loop() function, I'm trying to read the signal strength of the incoming packet to determine the sending device's closeness to the receiving device, for ZDO commands specifically.  I attempt to read the incoming message of type afIncomingMSGPacket_t. I thought that LinkQuality / RSSI might help me.

The parameters are all misplaced, the data is a total mess. For example, I see the source-address in the place of the groupID parameter. And so all the parameters are garbage.

I tried to fix this by using "#pragma pack (push, 1)" and "#pragma pack (pop)" around the afIncomingMSGPacket_t definition, but nothing changed. Once again, I've no idea why it would do this. Any comments / suggestions?

Thanks again for taking time to read this long message! 

Sorry for the confusion: I will put two posts, one per query, to help with clarity: 

For query 1, I have shown the function that I use, and two JPG images showing the variables in real time, before and after I change the long address:

The function:

/*********************************************************************
 * @fn      sendBindReq
 *
 * @brief   Sends a bind request to a remote device and to this device
 *
 * @param   (uses global dstAddr)
 *
 * @return  none
 */
void sendBindReq( uint8 linkEP )
{
  if ( linkMode == true )   //Set elsewhere
  {
    uint8 outCome1, outCome2;
    zAddrType_t thisDeviceAddr;
    zAddrType_t otherDeviceAddr; 
    uint8 tempExtAddr[Z_EXTADDR_LEN];
    
    AddrMgrExtAddrLookup( remoteAddr.addr.shortAddr, tempExtAddr );
    
    osal_cpyExtAddr( (void *)&thisDeviceAddr.addr.extAddr, (void *)NLME_GetExtAddr() );
    thisDeviceAddr.addr.shortAddr = NLME_GetShortAddr();    //This step overwrites the last two bytes of IEEE address!!
    thisDeviceAddr.addrMode = afAddr64Bit;
    
    osal_cpyExtAddr( (void *)&otherDeviceAddr.addr.extAddr, (void *)tempExtAddr );
    otherDeviceAddr.addr.shortAddr = remoteAddr.addr.shortAddr;     //This step overwrites the last two bytes of the IEEE address!! 
    otherDeviceAddr.addrMode = afAddr64Bit;
    
    //Make entry into dstAddr's binding table ...
    outCome1 = ZDP_BindReq( &remoteAddr, tempExtAddr, linkEP,  
                            ZCL_CLUSTER_ID_GEN_ON_OFF, &thisDeviceAddr, LOADCONTROL_EP100_ENDPOINT, 0 );          
    
    //Make entry into local binding table ... 
    outCome2 = ZDP_BindReq( &otherDeviceAddr, tempExtAddr, linkEP, 
                            ZCL_CLUSTER_ID_GEN_ON_OFF, &thisDeviceAddr, LOADCONTROL_EP100_ENDPOINT, 0 );
    
    linkMode = false;
  }
}

The "Watch" window in IAR debugger BEFORE I manually edit the long address:

Notice for "thisdeviceAddr" and "otherDeviceAddr", that the last two bytes of the long address match the two bytes of the short address, making the IEEE address incorrect.

The "Watch" window in IAR debugger AFTER I manually edit the long address:

Notice that after I change the last two bytes of the long address for both "thisDeviceAddr" and "otherDeviceAddr" to what they should be, the short address also changes to these two bytes on its own!

I can see that the memory addresses of the short and long addresses are the same, surely this should not be the case?

For Query 2, I have shown the code of the  *_event_loop() function:

/*********************************************************************
 * @fn          loadControlEP100_event_loop
 *
 * @brief       Event Loop Processor for loadcontrol.
 *
 * @param       uint8 task id - load control task id
 * @param       uint16 events - event bitmask
 *
 * @return      none
 */
uint16 loadControlEP100_event_loop( uint8 task_id, uint16 events )
{
  afIncomingMSGPacket_t *MSGpkt;  
  //afDataConfirm_t *afDataConfirm;

  if ( events & SYS_EVENT_MSG )
  {
    while ( (MSGpkt = (afIncomingMSGPacket_t *)osal_msg_receive( loadControlTaskID )) )
    {
      switch ( MSGpkt->hdr.event )
      {
        #ifdef DANIEL_PCT_CI_SE_2013
         case MT_SYS_APP_MSG:
         // Message received from MT (serial port)
         loadcontrol_ProcessAppMsg( ( (mtSysAppMsg_t *)MSGpkt)->appData );
         break;
         
         case AF_INCOMING_MSG_CMD:
          customCommsCI( MSGpkt );
          break;
          
         case AF_DATA_CONFIRM_CMD:
          // This message is received as a confirmation of a data packet sent.
          // The status is of ZStatus_t type [defined in ZComDef.h]
          // The message fields are defined in AF.h
          //afDataConfirm = (afDataConfirm_t *)MSGpkt;
          
          /*(void)sentEP;
          (void)sentTransID;

          // Action taken when confirmation is received.
          if ( sentStatus != ZSuccess )
          {
            // The data wasn't delivered -- Do something
          }*/
          break;
         
        #endif  //DANIEL_PCT_CI_SE_2013

        case ZDO_CB_MSG:
          
          if ( (MSGpkt->rssi < 90) && (MSGpkt->wasBroadcast == true) ) //To stop every single LCN responding to a IPD/Remote broadcast, 
          {   }                                                        //  but still allowing unicast ZDO messages through.                                                                
          else  {                                                     // Therefore broadcasting device must be close to LCN!                
            loadcontrol_ProcessZDOMsgs( (zdoIncomingMsg_t *)MSGpkt );  }            
          break; 
...

Have a look at case "ZDO_CB_MSG": I have attempted to filter out calls to the function "loadcontrol_ProcessZDOMsgs(...)" based on the two variables "MSGpkt->wasBroadcast" and "MSGpkt->rssi". The idea here is to only allow broadcasting messages that are very close to this device, to call the function.

This is what the "Watch" window shows when execution enters the case:

As you can see the short address (0xE0E8) is at the groupId parameter, and the short address, endPoint (0xBD) and clusterId (0x9E21) parameters are all some unknown value, along with all the others. This tells me that all the data in the struct "afIncomingMSGPacket_t" is completely mis-aligned.

So, when I try to test the two parameters in question, the values that they refer to are completely inconsistent: moving the calling device further or nearer to this device make absolutely no difference to the rssi value, which is simply some random number every time this code is called.

Thanks for taking time to understand these problems!

I see that the reply to Query 1 came out in a very confusing way! Here it is as it should have been:

>>>

Sorry for the confusion: I will put two posts, one per query, to help with clarity:

For query 1, I have shown the function that I use, and two JPG images showing the variables in real time, before and after I change the long address:

The function:

/*********************************************************************
 * @fn      sendBindReq
 *
 * @brief   Sends a bind request to a remote device and to this device
 *
 * @param   (uses global dstAddr)
 *
 * @return  none
 */
void sendBindReq( uint8 linkEP )
{
  if ( linkMode == true ) 
  {
    uint8 outCome1, outCome2;
    zAddrType_t thisDeviceAddr;
    zAddrType_t otherDeviceAddr; 
    uint8 tempExtAddr[Z_EXTADDR_LEN];
    
    AddrMgrExtAddrLookup( remoteAddr.addr.shortAddr, tempExtAddr );
    
    osal_cpyExtAddr( (void *)&thisDeviceAddr.addr.extAddr, (void *)NLME_GetExtAddr() );
    thisDeviceAddr.addr.shortAddr = NLME_GetShortAddr();  //This line changes the last two bytes of the IEEE address to its own values!!
    thisDeviceAddr.addrMode = afAddr64Bit;
    
    osal_cpyExtAddr( (void *)&otherDeviceAddr.addr.extAddr, (void *)tempExtAddr );
    otherDeviceAddr.addr.shortAddr = remoteAddr.addr.shortAddr;   //This line changes the last two bytes of the IEEE address to its own values!!
    otherDeviceAddr.addrMode = afAddr64Bit;
    
    //Make entry into dstAddr's binding table ...
    outCome1 = ZDP_BindReq( &remoteAddr, tempExtAddr, linkEP,  
                            ZCL_CLUSTER_ID_GEN_ON_OFF, &thisDeviceAddr, LOADCONTROL_EP100_ENDPOINT, 0 );          
    
    //Make entry into local binding table ... 
    outCome2 = ZDP_BindReq( &otherDeviceAddr, tempExtAddr, linkEP, 
                            ZCL_CLUSTER_ID_GEN_ON_OFF, &thisDeviceAddr, LOADCONTROL_EP100_ENDPOINT, 0 );
    
    linkMode = false;
  }
}

The "Watch" window in IAR debugger BEFORE I manually edit the long address:

Notice for "thisDeviceAddr" and "otherDeviceAddr", that the last two bytes of the long address match the two bytes of the short address, making the IEEE address incorrect.

The "Watch" window in IAR debugger AFTER I manually edit the long address:

Notice that after I change the last two bytes of the long address for both "thisDeviceAddr" and "otherDeviceAddr" to what they should be, the short address also changes to these two bytes on its own!
I can see that the memory addresses of the short and long addresses are the same, surely this should not be the case?

Thanks YiKai, 

Yes I know the memory is shared, but surely it should not be sharing the memory? If we look at the definition in ZComDef.h they are separate variables, so they should have separate addresses in memory. Or did I miss something?

typedef struct
{
  union
  {
    uint16      shortAddr;
    ZLongAddr_t extAddr;
  } addr;
  byte addrMode;
} zAddrType_t;