Bonding problem

RE: Bonding problem

Hung Bui — Wed, 28 Dec 2016 08:46:21 GMT

@Mohamed: It's not a black box, all the code are open, if it crash, it should be possible to figure out why it crashes and where.

I won't be able to help you if can't provide enough information on the issue or reproduce the issue on our example code.

RE: Bonding problem

Mohamed O.Abouzeid — Sun, 25 Dec 2016 00:05:49 GMT

Hi Hung bui, I tried proxmity application with nrf connect, it code is a little bit different because I depend on the bonding between the ble and phone, I mean you proxmity application didn't pop up window asking for pairing, I should go and press on "bond" to start bonding process, and also I've dfu in my application which is different than proxmity application,

anyways, I tried the code more and more & finished the debugging with that coclusion: if you've any problem with pairing, it crash, for example : you're connecting to the firmware, a pop up window asking you to pair or cancel, if you pressed on cancel, it will crash, I'm not doing any change in the device_manager_peripherial code & revised my code with the ble_app_hrs with dfu and found no difference.

the crash was at assembly command called DCW and behind it was written 2E2E the error code is very weird, it was 0x00000089

RE: Bonding problem

Mohamed O.Abouzeid — Thu, 22 Dec 2016 12:52:23 GMT

yeah yeah, you're right & I'm doing that already, I put break point inside the app_error_handler() and that's how I know that the crash in the file name and line num I had sent before, let me try it with ble app_proximity and come back to you.

RE: Bonding problem

Hung Bui — Thu, 22 Dec 2016 12:08:45 GMT

@mohamed: Please add a breakpoint inside app_error_handler() and find out which file name it is. 0x0001AB9C is an address in RAM, there should be some data there. You can take a screenshot showing how you get the file name, line, and err_code.

If you try to test with ble_app_proximity, and delete bond on the phone, do you have the same issue ?

RE: Bonding problem

Mohamed O.Abouzeid — Thu, 22 Dec 2016 10:07:57 GMT

but the problem is that there are two address one for the file name and one for the line which called the app_err_handler, which one to check? most problabably that there is a corruption while writing bonding information for a reason but I don't know what is it, what I'm doing is that: connect pair, when it bonded then I delete bond info from the mobile then try to connect again, it began to give hesitant behaviour, sometimes it connects and somtimes it refuses the connection and crash I guess it's pstorage file in SDK10 may cause that problem

RE: Bonding problem

Hung Bui — Wed, 21 Dec 2016 16:04:27 GMT

The assertion can't come from nothing. What is at address 0x0001AB9C in RAM ? You may want to track down what caused the assertion. Also test with the code that worked for you before.

You may want to describe how you get to the error, maybe a sniffer trace when you test with the DK can also help.

RE: Bonding problem

Mohamed O.Abouzeid — Wed, 21 Dec 2016 12:33:51 GMT

it's in the firmware I believe, I'm burning the firmware with FOTA to modules. I'm now debugging the DK and trying to replicate the case and I've found a bug randomly appears it showed me an error randomly appeared on file name (0x0001AB9C) with error in line 0x0000023D and err_code 0x00000088

RE: Bonding problem

Hung Bui — Wed, 21 Dec 2016 12:29:21 GMT

@mohamed: How often do you see the PTR51822A can bond and can not ?

It's not possible guess what could be wrong if you don't have access to the board and do some debugging, either using breakpoint or printing out trace on UART or RTT (SWD interface)

RE: Bonding problem

Mohamed O.Abouzeid — Wed, 21 Dec 2016 11:33:18 GMT

I'm using S110 anyways, but at the workingcase file, at packet 263, it's the same behaviour and the response of the master was different,

I don't understand what the revision will do with supporting s130 or something else.

another thing, I'm debugging the case again with nordic development kit, an arbitrarily crash pop up at file name in hexadecimal view(0x0001AB9C) which means that most probably the arbitrarily behaviour may be because of that crash.

RE: Bonding problem

Hung Bui — Wed, 21 Dec 2016 11:16:25 GMT

Could you explain this: "the slave is asking the master and then the master didn't respond for no reason" ? What I see is the master try to transmit but the the slave didn't response, so the master retransmit.

Another thing you have to check is which revision the chip is on the module. Could be that the chip was with revision 2 and won't support S130 v2.0 , please have a look here.

RE: Bonding problem

Mohamed O.Abouzeid — Wed, 21 Dec 2016 11:08:40 GMT

yeah, the sdk version I'm using is nRF51_SDK_10.0.0_dc26b5e, I'm using device manager to handle bonding and connection

you can look for example at packet 570 1st case, the slave is asking the master and then the master didn't respond for no reason, I guess something weird may be. may be an info is missed in the middle

here is the picture of module I'm using, it's name is PTR5518 eng V1.4 link text

RE: Bonding problem

Hung Bui — Wed, 21 Dec 2016 09:31:14 GMT

Please point me to the exact trace (file name) and if possible the packet number where I should check. The "workingnonworkingcase.pcapng" file showing your device crashed when it's trying to store encryption keys and then worked again after that meaning it's managed to store but for some reason it crashed after that.

Please state your SDK version, if you use pstorage or device manager.

Please point me to the schematic of the PTR51822a.

Depends on which SDK you use, re-bonding with a already bonded central could be rejected (for security reason) but it's not the case here, I don't see any rejecting. I suspect corrupted flash area, an erase all may help.

RE: Bonding problem

Mohamed O.Abouzeid — Tue, 20 Dec 2016 14:06:18 GMT

@hungbui, in the question, I'm attaching 6 different sniffer trace data, at different points, it didn't complete the connection, I don't know what is the reason for that. it has no big difference, only the PTR51822a module is used and in the DK, it's only the chip. you mean erase all bonds from the module, if it's already bonded to one of the phone, it should accept connection after that even if I erased the bond information from the phone.

RE: Bonding problem

Hung Bui — Tue, 20 Dec 2016 13:54:20 GMT

@mohamed: You mean you have no access to the programming interface ? Have you tested the device before you deploy it ?

How different the board is compare to the DK ?

Which exactly the trace you want me to have a loot at ?

Made sure you erase all bond before you test the device.

RE: Bonding problem

Mohamed O.Abouzeid — Tue, 20 Dec 2016 13:20:27 GMT

@HungBui, it's already inside the package, I don't know how to debug it while it's in the package, but it didn't happen with the development kit, if you've time and looked at the six files I'm attaching, you will find that it sometimes do it independtly of pairing or anything, it's just try to connect, then terminate it.

I don't know how to replicate this case in the DK because it's already the same code burnt to DK and the modules.

RE: Bonding problem

Hung Bui — Tue, 20 Dec 2016 11:45:43 GMT

What you observed usually happens when the firmware on the chip crashes and got restarted. You see the master keep transmitting but it's not "keeps sending two much packets with no waiting for the module to respond ". It's simply resend the packet be cause the slave didn't response to it.

What you need to do is to figure out why the system crash, you can follow the debug guide here.

I suspect it has something to do with how you handle bonding, the last command from the central to the peripheral is a pairing request.

RE: Bonding problem

Mohamed O.Abouzeid — Tue, 20 Dec 2016 10:23:57 GMT

any help please !!!