power cycling Emerson EM229 headset caused phone to reboot

[Selene ToyKeeper]

While sanity testing krillin rtm image 139, I paired a bluetooth headset. I made a call with it, hung up, then turned the headset off. I noticed this caused the BT indicator icon to invert, which is a nice way to indicate whether something is connected. So, I turned the device back on, waited for the icon to invert, turned the device off, waited, and instead of inverting the icon... the phone rebooted.

Not sure what exactly failed here, like which component, but it's definitely not good when a paired headset can cause the phone to spontaneously reboot.

I don't see any relevant .crash files, and all I see in syslog from that time is this:

Oct 30 23:42:27 ubuntu-phablet NetworkManager[1429]: <warn> Activation (/ril_1) failed for connection '/310410695117999/context1'
Oct 30 23:42:27 ubuntu-phablet NetworkManager[1429]: <info> (/ril_1): device state change: failed -> disconnected (reason 'none') [120 30 0]
Oct 30 23:42:27 ubuntu-phablet NetworkManager[1429]: <info> (/ril_1): deactivating device (reason 'none') [0]
Oct 30 23:43:01 ubuntu-phablet kernel: [ 747.229681][Ker_PM][request_suspend_state]wakeup (3->0) at 747214006509 (2014-10-31 05:43:01.704759580 UTC)
Oct 30 23:43:02 ubuntu-phablet kernel: [ 747.779813]mtk-tpd: TPD wake up
Oct 30 23:43:02 ubuntu-phablet kernel: [ 747.816572]mtk-tpd: TPD wake up done
Oct 30 23:43:02 ubuntu-phablet kernel: [ 747.816593]<MAGNETIC> mag_context_obj ok------->hwm_obj->early_suspend=0
Oct 30 23:43:09 ubuntu-phablet dbus[769]: [system] Activating service name='org.freedesktop.systemd1' (using servicehelper)
Oct 30 23:43:09 ubuntu-phablet dbus[769]: [system] Successfully activated service 'org.freedesktop.systemd1'
Oct 30 23:43:09 ubuntu-phablet rsyslogd: [origin software="rsyslogd" swVersion="7.4.4" x-pid="771" x-info="http://www.rsyslog.com"] exiting on signal 15.
Oct 30 23:43:42 ubuntu-phablet rsyslogd: [origin software="rsyslogd" swVersion="7.4.4" x-pid="785" x-info="http://www.rsyslog.com"] start
Oct 30 23:43:42 ubuntu-phablet rsyslogd-2307: warning: ~ action is deprecated, consider using the 'stop' statement instead [try http://www.rsyslog.com/e/2307 ]
Oct 30 23:43:42 ubuntu-phablet rsyslogd-2307: warning: ~ action is deprecated, consider using the 'stop' statement instead [try http://www.rsyslog.com/e/2307 ]
Oct 30 23:43:42 ubuntu-phablet rsyslogd: rsyslogd's groupid changed to 103
Oct 30 23:43:42 ubuntu-phablet rsyslogd: rsyslogd's userid changed to 100
Oct 30 23:43:42 ubuntu-phablet kernel: [ 0.000000]Booting Linux on physical CPU 0
Oct 30 23:43:42 ubuntu-phablet kernel: [ 0.000000]Initializing cgroup subsys cpu
Oct 30 23:43:42 ubuntu-phablet kernel: [ 0.000000]Linux version 3.4.67 (root@android-barajas_1414177384) (gcc version 4.7 (GCC) ) #1 SMP PREEMPT Tue Oct 28 11:57:42 UTC 2014

[Brendan Donegan (brendan-donegan)]

Is this headset new? You might have hit an interop issue - if there was no crash it was probably a kernel panic or something like that. I've tried disconnecting and reconnecting a number of time with my headset and there are no issues.

[Selene ToyKeeper (toykeeper)]

This is a headset.

BTW, this issue is easily reproducible. All I have to do is turn the headset on, wait for the icon to go white, wait another couple seconds, then turn the headset off. Within a few seconds the phone will completely lock up, and a few more seconds later it reboots.

[Selene ToyKeeper (toykeeper)]

Sometimes I have to turn the headset off and on twice, but I've never seen it take more than two tries.

[Selene ToyKeeper (toykeeper)]

The headset I'm using is what they directed me to at Walmart when I asked for an inexpensive bluetooth headset, the Emerson EM229. It has to be a pretty popular device in the states in order to be carried at Walmart, especially at under $20.

[Alex Kaluzhny (alex-kaluzhny)]

Krilin, r140, BT headset: Rocketfish RF-MAB2
On the first attempt, I was not able to connect to BT headset, which always connected and worked reliably with previous releases. I was able to connect after I rebooted the phone and power cycled BT headset. I tried about 5 more times power-cycling BT headset after that and it did connect reliably every time. It did not cause phone to reboot while BT headset was power cycled.

[Ricardo Salveti (rsalveti)]

Please get the content available at /proc/last_kmsg after a device reboots, that's the only way to get the kernel crash trace.

[Selene ToyKeeper (toykeeper)]

If I can manage to reproduce this again, I will. It seems quite a bit more stable in the more recent images though.

[Selene ToyKeeper (toykeeper)]

I finally managed to trigger this bug again. The last_kmsg file is attached, and confirms it is indeed a kernel panic. The traceback looks promising too.

[Selene ToyKeeper (toykeeper)]

Oh, um, this was on krillin rtm 176.
Linux ubuntu-phablet 3.4.67 #1 SMP PREEMPT Tue Dec 9 16:37:17 UTC 2014 armv7l armv7l armv7l GNU/Linux

[Ricardo Salveti (rsalveti)]

Nice, thanks for the trace, it should hopefully be enough to indicate the problem (which is clearly in the krillin bluetooth driver).

[John McAleely (john.mcaleely)]

I've raised #1403477 so that the driver team can look at this.

[Jean-Baptiste Lallement (jibel)]

Thank you for taking the time to report this bug and helping to make Ubuntu better. You reported this bug a while ago and there hasn't been any activity in it recently. We were wondering if this is still an issue for you?
Thanks in advance.

[Simon Fels (morphis)]

From the last kernel log this looks like a bug which was recently fixed upstream in the linux kernel.

See

- https://git.kernel.org/cgit/linux/kernel/git/bluetooth/bluetooth-next.git/commit/net/bluetooth?id=75e34f5cf69bd731d3b7375a786d4a15494fb8c6
- https://git.kernel.org/cgit/linux/kernel/git/bluetooth/bluetooth-next.git/commit/net/bluetooth?id=435c513369768f5840cd57101e398bc450fd26de
- https://git.kernel.org/cgit/linux/kernel/git/bluetooth/bluetooth-next.git/commit/net/bluetooth?id=1da5537eccd865b83fedbbb7ea704669f6d255fd
- https://git.kernel.org/cgit/linux/kernel/git/bluetooth/bluetooth-next.git/commit/net/bluetooth?id=2c501cdd6823a644cf35ac594c635310bf077125
- https://git.kernel.org/cgit/linux/kernel/git/bluetooth/bluetooth-next.git/commit/net/bluetooth?id=75e34f5cf69bd731d3b7375a786d4a15494fb8c6

Will prepare them to get imported into our trees at some point.

[Selene ToyKeeper (toykeeper)]

I'm not sure why this is marked as incomplete now. Has the new upstream kernel made it into our devices? Is a fix included and ready to test?

[Tony Espy (awe)]

@Selene

Sorry for updating the status without a comment. I think my intention was to get this re-tested on a current image and see if it was still reproducible.

Looks like Simon has commented about kernel changes that may address the original issue, so re-assigned to him.

[Selene ToyKeeper (toykeeper)]

Yes, this still happens with today's phone image. It took about a dozen power cycles to trigger.

[Daniel van Vugt (vanvugt)]

Reduced priority for Ubuntu Touch bugs.

[Daniel van Vugt (vanvugt)]

Ubuntu Touch is no longer supported.