Bug 154258

Summary:	kkpilotdaemon crashes always when polling usb: device for sync
Product:	[Unmaintained] kpilot	Reporter:	dukat <dukat>
Component:	kpilotDaemon	Assignee:	groot
Status:	RESOLVED UNMAINTAINED
Severity:	crash	CC:	broeksema, philsf79, robert, vR
Priority:	NOR
Version:	unspecified
Target Milestone:	---
Platform:	Ubuntu
OS:	Linux
Latest Commit:		Version Fixed In:
Sentry Crash Report:
Attachments:	Log file Backtrace

Description dukat 2007-12-18 00:09:51 UTC

Version: 4.9.2-358 (moribund alleycat) (using KDE KDE 3.5.8)
Installed from: Ubuntu Packages
OS: Linux

kpilot / Daemon seems to have a problem with the new usb: device sync (not using the visor module). I'm trying this with my Treo 680 on Kubuntu 7.10. I had these problems already on Kubuntu 7.04.

These errors occur on a fresh install - I just installed kpilot via apt-get - no previous config lying around.

1. kpilot crashes at the auto detection wizard.

Treo 680 in my cradle (the offcial Palm cradle), launched kpilot after fresh install. I chose wizard and then let Kpilot detect my setup. I press the Hotsync button, the counter in kpilot increases fast until 30%, then stops, ca. 10 seconds nothing, then crashes with signal 11. Always reproducible.
Interestingly, I get the same crash even when my Treo is not in the cradle. Not good.

2. kpilot-daemon crashes

Not I setup my device with the dialog (While the Treo is not connected to the cradle!). I change the default device from /dev/pilot to usb:, enter my Palm Sync name and change default sync to "Copy Handheld to PC". All other options stay the same. Kpilot starts up normally, the daemon is running in the tray. Kpilot ejects sporadically "Unable to read System information from Pilot", the progress is stays at 10%, and jumps to 30% whenever a new messages is printed, and immediately returns to 10%. This goes on and on.
Now I place the Treo into the cradle, NOT pressing the hotsync button - no sync is initiated, or has been performed beforehand.. Exactly at the next jump to 30% (but before printing its "Unable to read System information from Pilot"), kpilotdaemon crashes with signal 11.
Restarting kpilot without Treo in cradle. When I insert the Treo, I get again the crash at 30%. Always reproducible.

3. Crash during sync

Next attempt, now I press the hotsync button before the first jump to 30%, kpilot starts its sync just normal. Everything seems to work, as this is the initial sync, it does a full backup (ca. 20MB) to the PC. Somewhere during the road kpilot-daemon crashes suddenly (this time with signal 6, after ca. 30s synctime). After a restart I get a new crash during sync with signal 6. However, this time it was much earlier than before (after ca. 10s sync time). The third time the sync lasted roughly as long as the first one, but still not until the end. I give up.

I've also tested pilot-link with device usb: and J-Pilot with device usb: Neither had any problems.

Comment 1 jpo 2008-01-06 02:08:39 UTC

Same as problem 1 with Tungsten T, Palm Cradle, Gutsy

Comment 2 Mike Jolley 2008-02-23 20:52:35 UTC

I also have a Treo 680.  When I use usb: with Kpilot open, the daemon seg faults randomly.  It doesn't matter what screen I'm in.

[KCrash handler]
#5  0xb7b83aeb in __dynamic_cast () from /usr/lib/libstdc++.so.6
#6  0xb7bd4c38 in KPilotDeviceLink::customEvent ()
   from /usr/lib/libkpilot.so.0
#7  0xb6f1eeff in QObject::event () from /usr/lib/libqt-mt.so.3
#8  0xb6eb59b6 in QApplication::internalNotify () from /usr/lib/libqt-mt.so.3
#9  0xb6eb77df in QApplication::notify () from /usr/lib/libqt-mt.so.3
#10 0xb75fab52 in KApplication::notify () from /usr/lib/libkdecore.so.4
#11 0xb6e4616f in QApplication::sendEvent () from /usr/lib/libqt-mt.so.3
#12 0xb6eb69ee in QApplication::sendPostedEvents ()
   from /usr/lib/libqt-mt.so.3
#13 0xb6eb6af6 in QApplication::sendPostedEvents ()
   from /usr/lib/libqt-mt.so.3
#14 0xb6e5a385 in QEventLoop::processEvents () from /usr/lib/libqt-mt.so.3
#15 0xb6ed0d30 in QEventLoop::enterLoop () from /usr/lib/libqt-mt.so.3
#16 0xb6ed0a2e in QEventLoop::exec () from /usr/lib/libqt-mt.so.3
#17 0xb6eb755f in QApplication::exec () from /usr/lib/libqt-mt.so.3
#18 0x08058913 in ?? ()
#19 0xbfa3eb90 in ?? ()
#20 0x08071838 in ?? ()
#21 0x00000001 in ?? ()
#22 0x00000000 in ?? ()

Comment 3 dukat 2008-04-13 17:03:25 UTC

Reference to bug 1 (Seems to be a problem with newer Palm devices)

https://bugs.kde.org/show_bug.cgi?id=152681

Comment 4 Dom 2008-11-28 00:24:09 UTC

*** This bug has been confirmed by popular vote. ***

Comment 5 Jason 'vanRijn' Kasper 2009-01-19 06:14:14 UTC

Yeah, #1 is a dupe of the mentioned bug.

And try as I might, I cannot for the life of me figure out why libusb ("usb:" devices) causes kpilotDaemon to crash. We do EXACTLY the same thing with it as we do for standard devices and it just fails hard. One possible workaround would be to not poll, but to instead sit on it forever, but that isn't very nice of us either. Anyway, I'd love some help with this, because I've beat my head over it and just can't seem to get anywhere.

Comment 6 Felipe Figueiredo 2009-01-19 15:44:38 UTC

Jason,

I believe I experience the same bug over a wifi (net:) connection to my Palm TX. Is it possible the libusb is not the root cause, or perhaps even not the only cause of this behaviour? 

How can I provide useful debug information?

Comment 7 Jason 'vanRijn' Kasper 2009-01-19 19:46:18 UTC

Hi Felipe!! Yes, it's entirely possible, and I would REALLY appreciate your help in testing and debugging this!! You're going to either get packages that provide debugging information symbols (kdepim-kpilot-debuginfo or something), or compile your own version of KPilot with debugging symbols. Then run "gdb --args /path/to/your/kpilotDaemon --debug=9 --nofork 2>&1 | tee ~/kpilotDaemon.debug", then type "r" to get it to run, get it to crash, and then type "thread apply bt" and hit enter in gdb. that should tell us exactly where the crash is happening. Please create an attachment with that and I'll have a look.

Thank you!! =:)

Comment 8 Robert Marmorstein 2009-01-20 08:03:05 UTC

Created attachment 30444 [details]
Log file

Comment 9 Robert Marmorstein 2009-01-20 08:04:13 UTC

Created attachment 30445 [details]
Backtrace

Comment 10 Robert Marmorstein 2009-01-20 08:04:48 UTC

I think I am running into the same problem.  It seems to be an assertion failure in the Contacts Conduit.

Here are my stats:

Palm LifeDrive connecting using "net:any" (although I also tried usb: and /dev/pilot to no avail).

KDE: 4.3 trunk (4.3 svn 20090120)
QT: 4.4.3 (actually, qt-copy I think)
pilot-link: 0.12.3
KPilot: 5.2.0-beta1 (latest trunk from subversion)

I will attach a backtrace and a KPilot Log File.

Comment 11 Bertjan Broeksema 2009-01-27 10:06:49 UTC

Hi Robert,

I don't think that you're running in the same problem. Looking at your trace it looks more like a bug in the contacts conduit. A record got copied to the palm but after the copy the record on the pc and the palm are not equal. Are you able to reproduce this exact error and provide me with the details of the contact on your pc that triggers this assert?

Comment 12 Robert Marmorstein 2009-01-27 14:53:35 UTC

After playing around for awhile, my bug went away (now I just have the problem with Akonadi not syncing Kolab subfolders).  Sorry not to be more help!

Comment 13 Christophe Marin 2009-12-11 17:13:49 UTC

Unfortunately, KPilot is unmaintained due to lack of manpower and will not be part of KDE SC 4.4.

This bug reports will now be closed.