["Help" request] Strange and serious software'ish problem with Maemo's call/messaging system (?) - maemo.org

Estel | # 1 | 2015-01-23, 03:15 | Report

I got quite a serious, software-side (cross-checked) with my main N900 - it's quite complicated thing that I've never seen before, so, please, bear with me.

It started when, one day, device turned off due to discharged battery, while my son was gaming on it. Device was connected to WiFi at the time, and SIP account was registered (why it may be important, later on). fsck at the boot time found some minor problems, but none if it seemed serious (some thumbnail or temporary files of browser were damaged, some got connected to lost+found). FSCK ended up successfully, and marked filesystem as clean.

Hoever, after boot, strange things manifested themselves. I'll list them in points, cause there is plethora of them (and all of them seems inter-connected, like coming on from the same issue):

1. I can open phone-ui or adressbook, but I can't make any calls, despite having GSM/3G connection working all-right. When i try to call any (normal, cellular) number, either from dialpad, recent calls or addressbook, a "account not in use" message appears.

2. On random intervals (between 30 seconds and 2 minutes, at most), notification bar about "Phone application shut down, internal error" appears. If I have phone-ui open at the time, it closes itself.

3. If someone calls me, call prompt doesn't appear, and no ringtone is played. I know that phone is "processing" the incoming call, due to sudden slowness and characteristic cellular "noise" in speakers. Also, immediately when I'm called, message about phone application shutting down (internal error) from point #2 appears.

The other side sees it as if I've rejected call instantly. Immediately afterwards, I get a Maemo notification about failing to pick up a call.

4. I can receive SMS all right, but I can't send them - if I try to open conversations (either from menu or addressbook), I get the "phone application closed, internal error" message and nothing happens. If I try to reply after receiving and reading message, I can hit the "send" button, but it just becomes grayed out. There is no "processing" circle, or any other reaction.

5. I can receive e-mails, but can't send them - modest open fine, but just like with SMS, if I hit "send", it gets grayed out and nothing else happens (message doesn't even "close" itself, it just stay open, with "send" grayed out, infinitely).

6. I don't see "green circle" in status bar menu indicating that I'm connected to SIP proxy account, not I have any accounts visible, when I try to edit them. Hoever, I AM logged in to my SIP proxy - I'm visible from the proxy's side, and I can get calls from SIP (ending just like normal calls - like immediately rejected, no call prompt, and Maemo's message about missed call).

7. If I try to add some SIP account, I can fill the details and hit "add" button, but it just get grayed out, and nothing happens.
---

Dmesg doesn't show anything unusual. Only interesting lines in syslog (after I try to do anything with calling/messaging) is:

Code:

BloodRose rtcom-call-ui[5086]: GLIB MESSAGE default - Plugin registered: CALL.
BloodRose rtcom-call-ui[5086]: GLIB MESSAGE default - Plugin registered: CHAT.
BloodRose rtcom-call-ui[5086]: GLIB MESSAGE default - Plugin registered: SMS.

BloodRose rtcom-call-ui[5613]: GLIB WARNING ** libosso-abook - account_manager_ready_cb: Activation of org.freedesktop.Telepathy.AccountManager timed out
BloodRose rtcom-call-ui[5613]: GLIB CRITICAL ** default - Failed to get accounts: Activation of org.freedesktop.Telepathy.AccountManager timed out

BloodRose maemo-launcher[1364]: child (pid=5072) terminated due to exit()=1 
BloodRose maemo-launcher[1364]: no child 5073 found in the kindergarten. 
BloodRose maemo-launcher[1364]: child (pid=5073) terminated due to exit()=0 

BloodRose rtcom-call-ui[5617]: GLIB WARNING ** libosso-abook - account_manager_ready_cb: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.
BloodRose rtcom-call-ui[5617]: GLIB CRITICAL ** default - Failed to get accounts: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.

---

After trying to pinpoint the cause for a week, I must admit that I have no idea how to fix it. tried --purge reinstalling every package that I thought might have something to do with non-working functions:

Code:

apt-get install --reinstall --purge phonet-at phonet-utils csd-info libphinfo0 libtelcommon0 rtcom-accounts-plugin-sip rtcom-accounts-voip-support rtcom-messaging-ui rtcom-presence-ui rtcom-accounts-ui librtcom-accounts-widgets0 rtcom-call-ui librtcom-call-ui0 librtcom-eventlogger-ui1 librtcom-eventlogger-plugins1 librtcom-eventlogger1 librtcom-accounts-ui-client0 connui-conndlgs telepathy-ring libnice0 libconnui libconnui-cellular libconic0 libtelepathy-glib0 telepathy-sofiasip telepathy-stream-engine libtelepathy-farsight0 libtelcommon0

...to no avail. I tried nuking messages and events databases, too.

So, anyone got idea what could be related to issues that I'm experiencing? Maybe some other package that I could reinstall? There MUST be some file(s) that got damaged due to sudden shutdown resulting in this mess.

At this point, I'm close to do a painful (and first time ever, honestly) full-reflash with installing everything from scratch. But, before I do so, I would really like to try everything that I could to solve the mystery - in case such weirdo ever happens again. Thanks in advance for any help.

/Estel

// Edit
If it isn't obvious from the description - it is NOT hardware issue.

The Following 2 Users Say Thank You to Estel For This Useful Post:
reinob, Wikiwide

Mentalist Traceur | # 2 | 2015-01-23, 04:06 | Report

Do you have strace/ltrace/latrace installed? I would recommend installing those and running those programs that are malfunctioning through those tools, to see what libraries/system calls it makes, and probably see something about what/where it fails. (You can either launch those processes from the command line under {s/l/la}trace, or attach {s/l/la}trace to existing processes if you know their PIDs.) If the file corrupted is some low-end library, you may need to get statically compiled versions of the aforementioned to have them work right...

The Following 2 Users Say Thank You to Mentalist Traceur For This Useful Post:
Estel, Wikiwide

Estel | # 3 | 2015-01-23, 05:45 | Report

Thanks for the tip - here is strace output, from the start of rtcom-call-ui, to the moment it closes itself with "yellow bar" internal error message:
http://pastebin.aquilenet.fr/?01c721...gJt0XTkyaWuHI=

What caught my (ignorant, I'm using those tools for the very first time

) eye is the line:

Code:

access("/etc/ld.so.preload", R_OK)      = -1 ENOENT (No such file or directory)

Could it have something to do with it? Could someone who still have this file, check which packages it belongs to? The same ENOENT appears when invoking rtcom-messaging-ui.

latrace output (rtcom-call-ui):
http://pastebin.aquilenet.fr/?349a69...Svg4QkClW5kOo=

the interesting line might be:

Code:

 7472       __errno_location [/lib/libc.so.6]

...but without libc I guess I would have been experiencing more serious problems, right?

Again, same error with rtcom-messaging-ui.

ltrace output:

http://pastebin.aquilenet.fr/?94144d...40vOmJUXbhn3g=

Again, the interesting bit might be:

Code:

__errno_location() = 0x4001fd70

...but as the former two, I would be glad if someone knowledgeable would take look at whole log. My idea of what is "interesting" in the log might be not related to what is important.

Thanks again,
/Estel

The Following User Says Thank You to Estel For This Useful Post:
Wikiwide

Wikiwide | # 4 | 2015-01-23, 07:31 | Report

I have /etc/ld.so.cache and /etc/ld.so.conf, but no /etc/ld.so.preload

/lib/libc.so.6 is link to libc-2.5.so

Not read the logs, yet.

Best wishes.

The Following User Says Thank You to Wikiwide For This Useful Post:
Estel

Mentalist Traceur | # 5 | 2015-01-23, 08:05 | Report

Originally Posted by Estel View Post

Thanks for the tip - here is strace output, from the start of rtcom-call-ui, to the moment it closes itself with "yellow bar" internal error message:
http://pastebin.aquilenet.fr/?01c721...gJt0XTkyaWuHI=

What caught my (ignorant, I'm using those tools for the very first time

) eye is the line:

Code:

access("/etc/ld.so.preload", R_OK)      = -1 ENOENT (No such file or directory)

This is harmless. /etc/ld.so.preload is a special priviledged file that tells the linker what shared objects to load/link with everything being executed, more or less. It has a function similar to the envar LD_PRELOAD, except since /etc/ is traditionally writable only by root, the linker trusts this file more than the LD_PRELOAD variable. Anyway, when it's not there, nothing bad happens, and execution continues normally.

Here's a longer but still quick-ish explanation:
https://minipli.wordpress.com/2009/0...ld-so-preload/

That's all I got for now, sadly I have only limited experience reading those traces, and at any rate I have to go for the evening.

The Following 2 Users Say Thank You to Mentalist Traceur For This Useful Post:
Estel, Wikiwide

reinob | # 6 | 2015-01-23, 08:59 | Report

@Estel,

That looks weird. My guess is that it's something related to the "org.freedesktop.Telepathy.AccountManager" dbus service.

To start with I'd ask you to post the output of "ps ax | grep dbus". I have:

Code:

dbus-daemon --system --nofork
dbus-launch --exit-with-session
dbus-daemon --fork --print-pid 5 --print-address 7 --session

plus a few mafw-dbus-wrapper instances (I'm writing this by hand so I didn't post full details).

You could also check the contents of /etc/dbus-1/system.d

Another guess (as I'm only guessing) would be to check/reinstall librtcom-accounts-*
Also check/reinstall libosso-dbus-conf

Good luck!

The Following 3 Users Say Thank You to reinob For This Useful Post:
Estel, Mentalist Traceur, Wikiwide

Estel | # 7 | 2015-01-23, 14:00 | Report

Thanks for all the input, guys.

ps ax | grep dbus output is:

Code:

BloodRose:~# ps ax | grep dbus
  786 messageb  3508 S <  /usr/bin/dbus-daemon --system --nofork
 1247 user      2616 S    dbus-launch --exit-with-session
 1273 user      3128 S <  /usr/bin/dbus-daemon --fork --print-pid 5 --print-address 7 --session
 1780 user     63284 S <  /usr/bin/mafw-dbus-wrapper mafw-gst-renderer
 1800 user      7668 S    /usr/bin/mafw-dbus-wrapper mafw-iradio-source
 1817 user     15528 S    /usr/bin/mafw-dbus-wrapper mafw-tracker-source
 1824 user      8968 S    /usr/bin/mafw-dbus-wrapper mafw-upnp-source
 8999 user     23672 S    {dbus-switchboar} /usr/bin/python2.5 /usr/bin/dbus-switchboard.py
16925 user      3128 S <  /usr/bin/dbus-daemon --fork --print-pid 5 --print-address 7 --session
16947 root      2616 S    grep dbus

My /etc/dbus-1/system.d content is:

Code:

BloodRose:~# ls -lh /etc/dbus-1/system.d/
-rw-r--r--    1 root     root         589 Feb 22  2010 alarmd.conf
-rw-r--r--    1 root     root         774 Jun 15  2010 bluetooth.conf
-rw-r--r--    1 root     root         290 Jun 21  2010 clockd.conf
-rw-r--r--    1 root     root         449 Mar  3  2010 com.nokia.SystemInfo.conf
-rw-r--r--    1 root     root         538 Feb  2  2010 csd-call.conf
-rw-r--r--    1 root     root         536 Oct 13  2009 csd-gprs.conf
-rw-r--r--    1 root     root         425 Jan 26  2010 csd-info.conf
-rw-r--r--    1 root     root         412 Dec 15  2009 csd-sms.conf
-rw-r--r--    1 root     root         411 Jan 27  2010 csd-ss.conf
-rw-r--r--    1 root     root         306 Jan 18  2010 csd.conf
-rw-r--r--    1 root     root         610 Feb 12  2010 dnsmasq.conf
-rw-r--r--    1 root     root         282 Jun 24  2010 eap.conf
-rw-r--r--    1 root     root         264 Jul  8  2010 fmtxd.conf
-rw-r--r--    1 root     root         419 Jun  8  2010 gconfd.conf
-rw-r--r--    1 root     root        2.5K Apr 16  2010 hal.conf
-rw-r--r--    1 root     root         609 Jan 18  2010 hald-addon-bme.conf
-rw-r--r--    1 root     root         246 May 16  2008 hulda.conf
-rw-r--r--    1 root     root         293 Jun 29  2010 icd-ui.conf
-rw-r--r--    1 root     root         266 Jun 17  2010 icd2.conf
-rw-r--r--    1 root     root         285 Jun  7  2010 iphbd.conf
-rw-r--r--    1 root     root         635 Jun 30  2010 isync-dbus.conf
-rw-r--r--    1 root     root         253 May 16  2008 ke-recv-test.conf
-rw-r--r--    1 root     root         248 May 16  2008 ke-recv.conf
-rw-r--r--    1 root     root         413 Aug  7  2009 libcsnet.conf
-rw-r--r--    1 root     root         549 May 28  2010 libosso.conf
-rw-r--r--    1 root     root         411 Dec  2  2009 libsim.conf
-rw-r--r--    1 root     root         413 May 12  2009 libsimpb.conf
-rw-r--r--    1 root     root         485 Jul 29  2010 mce.conf
-rw-r--r--    1 root     root        1.2K Feb 22  2010 ohm.conf
-rw-r--r--    1 root     root         288 Jun  4  2009 omacp.conf
-rw-r--r--    1 root     root         396 Jan 25  2010 sscd.conf
-rw-r--r--    1 root     root         305 Jun 15  2010 system_ui.conf
-rw-r--r--    1 root     root         477 Mar 22  2010 tone-generator.conf
-rw-r--r--    1 root     root         305 Dec 16  2009 wappushd_policy.conf
-rw-r--r--    1 root     root         375 Jul  5  2010 wlancond.conf

...and none of these files looks like it got damaged/replaced by something else (due to cloned inodes or whatsnot). No idea if something might be lacking (as in file deleted), though.

Trying to reinstall packages that you've suggested, now.

/Estel

The Following 2 Users Say Thank You to Estel For This Useful Post:
reinob, Wikiwide

Estel | # 8 | 2015-01-24, 03:53 | Report

Reinstalled those (and other dbus-related) packages, to no avail

It's worth to mention, that - with almost 100% probability - it must be something on /dev/mmcblk0p2 (home, opt) damaged, not a thing that resides on ubifs root. Having the latter filesystem damaged is highly unlikely/impossible.

Not that the above gives me any idea where to look next, though... Maybe device gets confused by not having SIP accounts set, while at the same time being connected to them? It's a shoot in the dark, but it COULD be trying to ask me if I want to send/call via GSM or VOIP, then get confused somehow... Where VoIP accounts data is stored, exactly? I could try purging it.

/Estel

The Following 2 Users Say Thank You to Estel For This Useful Post:
reinob, Wikiwide

Maxdamantus | # 9 | 2015-01-24, 09:18 | Report

Random thought: maybe one of the osso-abook files is corrupt. I suspect you can move/back up ~/.osso-abook and reboot to reset it .. I'd probably back up all of ~/ anyway. I've noticed similarly weird stuff happening with a corrupt conversations database (though that's sqlite3 while osso-abook appears to be Berkley DB).

The Following 3 Users Say Thank You to Maxdamantus For This Useful Post:
Estel, Wikiwide, xman

peterleinchen | # 10 | 2015-01-24, 09:58 | Report

Thought of a moment something like that.
But does not match your other probs.
Especially modest??? Which let me think of a more global FS prob? But, sorry, I have no idea

.

So maybe maxdamantus hint is better?

The Following User Says Thank You to peterleinchen For This Useful Post:
Wikiwide