[solved]plasmashell crash caused by plasma-nm when connecting to a VPN

whites11 · 2016-07-24 08:53:28

Hello everyone
since last friday I am experiencing a reproducible crash of plasmashell somehow caused by plasma-nm.
Please note that I use plasma-nm widget many times a day to deal with VPN connections for working purposes, so I am sure that this issue wasn't present before friday morning in my system.

My situation:
up-to-date arch linux with:

plasma-workspace 5.7.2-1
qt5-base 5.7.0-1
plasma-nm-5.7.2-1 (but tried all versions down to 5.6.5, all the same problem)

how to reproduce the crash:

Just open the networkmanager widget from the panel and click "connect" on a VPN network connection OR use the nmcli/nm-tui tool to connect to a network.
This happens for connections that require to input a password or that have a password saved (in other words with or without the password prompt).
As a sidenote,the VPN connection get established correctly.
Of course the applet must be attached to the panel to get the crash.

the STDOUT of plasmashell when the crash happens (any hint on how to make it more verbose?):

QDBusObjectPath: invalid path ""
KCrash: Attempting to start /usr/bin/plasmashell from kdeinit
sock_file=/run/user/1000/kdeinit5__0
KCrash: Application 'plasmashell' crashing...
KCrash: Attempting to start /usr/lib/drkonqi from kdeinit
sock_file=/run/user/1000/kdeinit5__0
QSocketNotifier: Invalid socket 8 and type 'Read', disabling...
QSocketNotifier: Invalid socket 17 and type 'Read', disabling...

I tried to reset plasma to the default settings as well as downgrading a few core packages that gave issues in the past with plasma in my system (xf86-video-intel).

I have no idea which component really causes this crash, which didn't happen at all till last thursday.
This is a log of the packages I upgraded last thursday afternoon:

[2016-07-21 16:27] [PACMAN] Running 'pacman -Syu'
[2016-07-21 16:27] [PACMAN] synchronizing package lists
[2016-07-21 16:27] [PACMAN] starting full system upgrade
[2016-07-21 16:27] [ALPM] transaction started
[2016-07-21 16:27] [ALPM] upgraded libsystemd (230-5 -> 230-7)
[2016-07-21 16:27] [ALPM] upgraded php (7.0.8-1 -> 7.0.9-1)
[2016-07-21 16:27] [ALPM] upgraded php-imap (7.0.8-1 -> 7.0.9-1)
[2016-07-21 16:27] [ALPM] upgraded php-intl (7.0.8-1 -> 7.0.9-1)
[2016-07-21 16:27] [ALPM] upgraded php-mcrypt (7.0.8-1 -> 7.0.9-1)
[2016-07-21 16:27] [ALPM] upgraded php-pgsql (7.0.8-1 -> 7.0.9-1)
[2016-07-21 16:27] [ALPM] upgraded python-setuptools (1:24.0.3-1 -> 1:24.2.0-1)
[2016-07-21 16:27] [ALPM] upgraded python2-setuptools (1:24.0.3-1 -> 1:24.2.0-1)
[2016-07-21 16:27] [ALPM] upgraded systemd (230-5 -> 230-7)
[2016-07-21 16:27] [ALPM] upgraded systemd-sysvcompat (230-5 -> 230-7)
[2016-07-21 16:27] [ALPM] upgraded xorg-server-common (1.18.3-3 -> 1.18.4-1)
[2016-07-21 16:27] [ALPM] upgraded xorg-server (1.18.3-3 -> 1.18.4-1)
[2016-07-21 16:27] [ALPM] transaction completed
[2016-07-21 16:27] [ALPM] running 'man-db.hook'...

Any hint on how to troubleshoot will be really appreciated.

Last edited by whites11 (2016-07-28 09:32:16)

arojas · 2016-07-24 09:01:23

Console output is useless for debugging crashes, please post a full backtrace

whites11 · 2016-07-24 09:49:41

Could you point me to a page that explains how to get a backtrace?

ewaller · 2016-07-24 15:38:17

whites11 wrote:

Could you point me to a page that explains how to get a backtrace?

Try this: Use coredumpctl to get a list of recent crashes on your system. If plasma-nm is the last thing in the list, then just run coredumpctl gdb
If it isn't, but plasma-nm is in the list someplace, then do coredumpctl gdb plasma-nm

If all goes well, you will find yourself in gdb. Enter the command bt

whites11 · 2016-07-25 10:46:04

thanks ewaller for the response.
with coredumpctl I can see that all the recent crashes in my system happened to plasmashell itself.
I've run coredumpctl gdb as suggested but I think I am without the debugging symbols

$ coredumpctl gdb
           PID: 3037 (plasmashell)
           UID: 1000 (whites)
           GID: 100 (users)
        Signal: 11 (SEGV)
     Timestamp: lun 2016-07-25 09:05:50 CEST (3h 37min ago)
  Command Line: /usr/bin/plasmashell --shut-up
    Executable: /usr/bin/plasmashell
 Control Group: /user.slice/user-1000.slice/session-c2.scope
          Unit: session-c2.scope
         Slice: user-1000.slice
       Session: c2
     Owner UID: 1000 (whites)
       Boot ID: a9d393454c334abbbc38526fe46a6eed
    Machine ID: 33cac937ea1a406db83de0e81f0d58af
      Hostname: whitesbook
      Coredump: /var/lib/systemd/coredump/core.plasmashell.1000.a9d393454c334abbbc38526fe46a6eed.3037.1469430350000000000000.lz4
       Message: Process 3037 (plasmashell) of user 1000 dumped core.
                
                Stack trace of thread 3037:
                #0  0x00007fc721c3e159 n/a (/usr/lib/qt/qml/org/kde/plasma/networkmanagement/libplasmanm_qmlplugins.so)

GNU gdb (GDB) 7.11.1
Copyright (C) 2016 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-pc-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>.
Find the GDB manual and other documentation resources online at:
<http://www.gnu.org/software/gdb/documentation/>.
For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from /usr/bin/plasmashell...(no debugging symbols found)...done.
BFD: Warning: /var/tmp/coredump-1CuLcH is truncated: expected core file size >= 2368049152, found: 2147483648.
[New LWP 3037]
[New LWP 3044]
[New LWP 3040]
[New LWP 3042]
[New LWP 3043]
[New LWP 3045]
[New LWP 3046]
[New LWP 3050]
Cannot access memory at address 0x7fc7e7f9f0e8
Cannot access memory at address 0x7fc7e7f9f0e0
Failed to read a valid object file image from memory.
Core was generated by `/usr/bin/plasmashell --shut-up'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0  0x00007fc721c3e159 in ?? ()
[Current thread is 1 (LWP 3037)]
(gdb)

As long as I will be able to I will try and rebuild the package including the debugging symbols

arojas · 2016-07-25 12:08:41

You missed the last step:

ewaller wrote:

If all goes well, you will find yourself in gdb. Enter the command bt

whites11 · 2016-07-25 13:26:39

I have rebuilt plasma-workspace and plasma-nm with debugging symbols enabled.
anyway I must have done something wrong 'cause it doesnt' give me a backtrace:

$ coredumpctl gdb
           PID: 19230 (plasmashell)
           UID: 1000 (whites)
           GID: 100 (users)
        Signal: 11 (SEGV)
     Timestamp: lun 2016-07-25 15:23:17 CEST (31s ago)
  Command Line: /usr/bin/plasmashell
    Executable: /usr/bin/plasmashell
 Control Group: /user.slice/user-1000.slice/session-c2.scope
          Unit: session-c2.scope
         Slice: user-1000.slice
       Session: c2
     Owner UID: 1000 (whites)
       Boot ID: a9d393454c334abbbc38526fe46a6eed
    Machine ID: 33cac937ea1a406db83de0e81f0d58af
      Hostname: whitesbook
      Coredump: /var/lib/systemd/coredump/core.plasmashell.1000.a9d393454c334abbbc38526fe46a6eed.19230.1469452997000000000000.lz4
       Message: Process 19230 (plasmashell) of user 1000 dumped core.
                
                Stack trace of thread 19230:
                #0  0x00007fe55ce10179 operator() (/usr/lib/qt/qml/org/kde/plasma/networkmanagement/libplasmanm_qmlplugins.so)

GNU gdb (GDB) 7.11.1
Copyright (C) 2016 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-pc-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>.
Find the GDB manual and other documentation resources online at:
<http://www.gnu.org/software/gdb/documentation/>.
For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from /usr/bin/plasmashell...done.
BFD: Warning: /var/tmp/coredump-yGVvtP is truncated: expected core file size >= 2350624768, found: 2147483648.
[New LWP 19230]
[New LWP 19235]
[New LWP 19233]
[New LWP 19236]
[New LWP 19237]
[New LWP 19238]
[New LWP 19239]
Cannot access memory at address 0x7fe61bb470e8
Cannot access memory at address 0x7fe61bb470e0
Failed to read a valid object file image from memory.
Core was generated by `/usr/bin/plasmashell'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0  0x00007fe55ce10179 in ?? ()
[Current thread is 1 (LWP 19230)]
(gdb) bt
#0  0x00007fe55ce10179 in ?? ()
Backtrace stopped: Cannot access memory at address 0x7ffe81ea8b30
(gdb)

any suggestions?

whites11 · 2016-07-27 14:12:31

I had to install the QT debug symbols, I opened a bug on kde bugs: https://bugs.kde.org/show_bug.cgi?id=366164

thanks for the help

whites11 · 2016-07-28 09:33:10

For anybody coming here, there is a bug in plasma-nm

see the bug linked in comment #8 for details

Arch Linux

#1 2016-07-24 08:53:28

[solved]plasmashell crash caused by plasma-nm when connecting to a VPN

#2 2016-07-24 09:01:23

Re: [solved]plasmashell crash caused by plasma-nm when connecting to a VPN

#3 2016-07-24 09:49:41

Re: [solved]plasmashell crash caused by plasma-nm when connecting to a VPN

#4 2016-07-24 15:38:17

Re: [solved]plasmashell crash caused by plasma-nm when connecting to a VPN

#5 2016-07-25 10:46:04

Re: [solved]plasmashell crash caused by plasma-nm when connecting to a VPN

#6 2016-07-25 12:08:41

Re: [solved]plasmashell crash caused by plasma-nm when connecting to a VPN

#7 2016-07-25 13:26:39

Re: [solved]plasmashell crash caused by plasma-nm when connecting to a VPN

#8 2016-07-27 14:12:31

Re: [solved]plasmashell crash caused by plasma-nm when connecting to a VPN

#9 2016-07-28 09:33:10

Re: [solved]plasmashell crash caused by plasma-nm when connecting to a VPN

Board footer