Why the value of afAddrMode_t change from Addr16Bit to AddrBroadcast with shortAddr equal to 0xfffd?

The address 0xFFFD is used to send to all devices on which macRxOnWhenIdle is set to TRUE. The address 0xFFFC is used to send to all routers and the coordinator. When you say "coordinator send message to enddevice the value of afAddrMode_t change from Addr16Bit to AddrBroadcast with shortAddr equal to 0xfffd", does this message send by your application?

When the communication have no problem,the broadcast with address 0xFFFD do not appear.Once the communication have problem,the odd
broadcast appear

Which p.nbr do you think there is a problem?

The p.nbr is 125,131,137,142 and so on.Originally they are unicast message,but after APSDE_DataReq() they change to broadcast message,and I can not find the reason because some code is not open.

this is normal communication:

this is abnormal communication:

By the way, I cannot know what this packet is because the message is encrypted and I don't know the key. Do you use a specific network key?

thank you for helping to solve the problem,
this is the whole situation here:
I have two devices,one coordinator and one enddevice.I want to test the point-to-point communication with them.
I set a timer in enddevice,every the arrival time the enddevice will be woken and send a message to coordinator.Once the coordinator receive the message the led in the board change it's state and send back a message to enddevice.Once the enddevice receive the message the led in the board also change it's state.
At the beginning,the communication is well.I move the coordinaor to a place beyond power can reach.After a while,I move the coordinator to the original place.Then the prlblem arise.

Z-Stack version is z-stack ha 1.2.2,I don't change the key,it is the original key in z-stack.
the packet from enddevice to coordinator is character string "device",from coordinator to enddevice is hexadecimal{0x01,0x02,0x03}.

Do you enable NV_RESTORE in your test?

Yes,I enabled NV_RESTORE in my test,what's more,the enddevice setting is POLL_RATE=0,after enddevice send message,it wait 500ms to use NLME_SetPollRate(1) to request data from father node(coordinator).

You say "NLME_SetPollRate(1) means to do polling every 1 milli-second" is misapprehensive.

this is the annotation in z-stack:

Yes, I forget this and you are correct. Can you do the test again and use sniffer to log all the process including ZED joins ZC?

Hi,Yk Chen,

this is the new capture:

capture_package-point_to_point.psd

I still cannot decrypt your sniffer file. It is wired. What example in Z-Stack Home 1.2.2 do you use?

I forget that I have changed something,I must use the original z-stack to test again,please wait for a moment.

Hi,this is the new capture:

capture.psd

Hi,YK Chen,
Can you decrypt my new sniffer file this time?

Yes, I can read it now. But, what is your problem?

After coordinator move back to the original place,enddevice can send to coordinator,but coordinator can not send back to enddevice.
I check the packet,it has change unicast to broadcast.

Actually, it is a route request which means ZC is look for routing path to send message to ZED. It is very strange that actually the ZED does polling to ZC but ZC cannot recognize ZED. Do you change any thing in Z-Stack?

I don't change anything important in the z-stack.
Do you know what can make father node cannot recognize it's child node?
They had disconnected in a period of time,but when they connected again,problem arised.

Can you reproduce this problem every time? By the way, what Z-Stack Home example do you test on?

Yes,I can reproduce this problem easily.
I do not use the example in z-stack ha,but using the SampleApp in older z-stack version.
I am sure that the problem has nothing to do with the upper application,it is a problem related with z-stack itself.

I will try to invite some TI guys to have a look at your problem.

By the way, can you share your test projects?

Thanks a lot for helping me to solve the problem.I am Chinese and my English is bad.Please be considerate!

OK,just wait for a moment

Don't worry about the English. It is OK.

Hi,YK Chen,

this is the whole project:

Z-Stack Home 1.2.2.test.rar

I only use the coordinator and enddevice,please check!

Hi,
when the NWK layer receive the packet from its aps layer(APSDE_DataReq),the NWK layer will search the destination address whether it is in neighbor list.If it is not found in the neighbor list,then find it in the router list.If not found in both the two table list,a route discovery is initiated .I think the broadcast message is a route discovery.
Now I suppose that the enddevice address which is in the neighbor list of the coordinator originally is deleted or covered in somewhere.
What's going on after you invited some TIs to have a look at this issue?

I still don't get their response about this issue.

Hi,many days have gone,what about my issue?

I think it is almost a month passed. TI had replied that they will response to this but I don't know when. Or, you can try to contact TI local representative for this issue.

Ok,thanks,I will try to contact with TI local representative.

You are welcome.

Do you get any feedback to contact TI local representative? I also tried to contact TI for this and they told me they were still working on it. I think this might be really a difficult issue for them. I would keep asking them and hope them would reply soon.

I have contact Ti local representative,and they have confirmed the issue is a bug in zstack.They told me they will give me a solution around June 5th.
Thank you very much.

OK, it is a good news.

Do you have any feedback from TI?

no

This is not good.

For me and for my team it is not good,do you have feedback from Ti?

No, I ask TI several times but no reply.

It is not a good new

Yes, it is no good. I think this might be a big issue in Z-Stack and not easy to fix.

Hi shuming,
Do you have any feedback from TI?