Hi all,
I have a spurious crash in the softdevice I guess. It occurs sometimes when the master (an HTC-M8 android phone) rapidly submits characteristic changes (acknowledgement turned off) - and cancel the android app asynchronously.
Here is a dump of a backtrace:
Program received signal SIGTRAP, Trace/breakpoint trap.
0x00013b0a in ?? ()
(gdb) bt
0 0x00013b0a in ?? ()
1 0x00013b42 in ?? ()
2 0x00013b42 in ?? ()
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
The addresses are quite different. Another dump was:
0 0x00002ade in ?? ()
1 0x000034bc in ?? ()
2 0x000034bc in ?? ()
It's reproducible when the handler "ble_evt_dispatch" takes about 20ms in BLE_GATTS_EVT_WRITE.
I'm using softdevice version 7.0.0.3-alpha. Is this a known issue? Any ideas? Side note: I know that this handler should finish as fast as possible but I'm unable to guarantee hard real time behavior here because of external hardware I'm actual using.
Thanks, Peer