Hi Dialog,
We have found a weird problem when connecting Sony Xperia XZ with Bluetooth 5.0 as master
connecting to Murata ZY (DA14580) with SDK 5.0.3 with ble_app_peripheral example template (unchanged).
After several repeated connection and disconnection, the module would either :
1. Loop inside the stack (between address 0x25XXX, 0x26XXX and 0x33XXX), never went to main loop (breakpoint never activated),
no advertisement.
2. Hard Fault-ed.
As for no.1, we checked with bluetooth sniffer, and just before it occurs, it would have a CONNECT_IND packet with invalid WindowOffset and WindowSize parameters.
We can send the sniffer log if you're interested.
We don't have any other Bluetooth 5.0 enabled device so we can't confirm that this happen in other BLE 5.0 device beside Sony Xperia XZ.
Has this ever occurred before? (Is it known bug?).
Thanks
Note that we already deployed more than 100000 devices with 5.0.3, and we cannot afford those to just fail with the incoming BLE 5.0 smartphones..
嗨meriororen,
I am not aware of any issues between the Bluetooth 5 spec and the 580 devices although i am waiting for the team in order to verify this. It would be a good idea to test this with another BLE 5 enabled device. Also a sniffer log would help as well. But as far as i can understand since there is a CONNECT_IND packet send from the central device with an invalid parameter that means that the mobile device doesn't comply with the spec.
Thanks MT_dialog
Thanks for the confirmation.
Unfortunately, currently we only have one BLE 5 device in our hand.
具体而言,电话我s Xperia XZ Premium
http://www.sonymobile.co.jp/xperia/docomo/so-04j/
https://www.sonymobile.com/global-en/products/phones/xperia-xz-premium/
Please found the attached logs. It's a zip file with log from teledyne lecroy's BPA low energy sniffer.
You can view the log with the software downloaded from their website.
http://www.fte.com/support/CPAS-download.aspx?demo=ComProbe%20BPA%20low%...
Thanks
嗨meriororen,
I can see the errors in the connection indications on the sniffer logs that you have attached, i dont think that this has anything to do with the BLE 5 and i dont think that you will face any problems with new BLE 5 enabled devices as long as they stick with the protocol (i can see that the device also generate an access address invalid to the BLE specification 0x00000000 and also the CRC value is incorrect). If you would like to do negative testing and handle cases where devices send out of spec data then you can always recover via unexprected cases with a reset from the hardfault or the NMI. I ve also informed the QA on this regarding the phone and the abnormalities during repeated connections.
Thanks MT_dialog