| View previous topic :: View next topic |
| Author |
Message |
John R. Graham
Administrator
Joined: 08 Mar 2005
Posts: 10645
Location: Somewhere over Atlanta, Georgia
|
 Posted: Mon Sep 02, 2024 9:29 pm Post subject: Can't Get a Core Dump on a Background Process |
 |
|
After an apparently trouble free upgrade to Plasma 6 on my main workstation, I am encountering a problem during startup of Plasma. I'm using display-manager configured for sddm. I get to the graphical greeter and appear, from the logs, to log in successfully, but startplasma-x11 (part of kde-plasma/plasma-workspace), invoked by sddm, is segfaulting. The greeter subsequently restarts.
| /var/log/messages: | Sep 1 21:22:18 terra sddm-helper[9819]: gkr-pam: unable to locate daemon control file
Sep 1 21:22:18 terra sddm-helper[9819]: gkr-pam: stashed password to try later in open session
Sep 1 21:22:18 terra sddm-helper[9819]: pam_kwallet5(sddm:auth): pam_kwallet5: pam_sm_authenticate
Sep 1 21:22:18 terra sddm-helper[9819]: pam_kwallet5(sddm:setcred): pam_kwallet5: pam_sm_setcred
Sep 1 21:22:18 terra sddm-helper[9819]: pam_unix(sddm:session): session opened for user jgraham(uid=1000) by jgraham(uid=0)
Sep 1 21:22:18 terra elogind-daemon[3889]: New session 6 of user jgraham.
Sep 1 21:22:18 terra sddm-helper[9819]: gkr-pam: gnome-keyring-daemon started properly and unlocked keyring
Sep 1 21:22:18 terra sddm-helper[9819]: pam_kwallet5(sddm:session): pam_kwallet5: pam_sm_open_session
Sep 1 21:22:18 terra sddm-helper[9833]: pam_kwallet5: final socket path: /run/user/1000/kwallet5.socket
Sep 1 21:22:19 terra sddm-helper[9797]: pam_unix(sddm-greeter:session): session closed for user sddm
Sep 1 21:22:20 terra elogind-daemon[3889]: Removed session c4.
Sep 1 21:22:20 terra dbus-daemon[9846]: [session uid=1000 pid=9844 pidfd=4] Activating service name='org.a11y.Bus' requested by ':1.1' (uid=1000 pid=9838 comm="/usr/bin/nvidia-settings --load-config-only" label="kernel")
Sep 1 21:22:20 terra dbus-daemon[9846]: [session uid=1000 pid=9844 pidfd=4] Successfully activated service 'org.a11y.Bus'
Sep 1 21:22:20 terra kernel: startplasma-x11[9836]: segfault at 7ffab39d3780 ip 00007ffab30c591b sp 00007ffdf8e1ac80 error 6 in libKF6ConfigCore.so.6.5.0[7ffab3099000+4e000] likely on CPU 10 (core 4, socket 0)
Sep 1 21:22:20 terra kernel: Code: 04 25 28 00 00 00 48 89 84 24 c8 00 00 00 31 c0 c4 e1 f9 7e c8 48 89 54 24 60 c5 f9 7f 0c 24 c5 f9 7f 4c 24 50 48 85 c0 74 03 <f0> ff 00 e8 3d 0f 00 00 48 83 7c 24 60 00 49 89 c7 75 09 45 85 f6
Sep 1 21:22:20 terra sddm-helper[9819]: pam_unix(sddm:session): session closed for user jgraham
Sep 1 21:22:20 terra sddm-helper[9819]: pam_kwallet5(sddm:session): pam_kwallet5: pam_sm_close_session
Sep 1 21:22:20 terra sddm-helper[9819]: pam_kwallet5(sddm:setcred): pam_kwallet5: pam_sm_setcred
Sep 1 21:22:20 terra elogind-daemon[3889]: Removed session 6.
Sep 1 21:22:22 terra sddm-helper[9877]: pam_unix(sddm-greeter:session): session opened for user sddm(uid=219) by root(uid=0)
Sep 1 21:22:22 terra elogind-daemon[3889]: Existing logind session ID 3 used by new audit session, ignoring.
Sep 1 21:22:22 terra elogind-daemon[3889]: New session c5 of user sddm. |
Command line configuration to enable core dumps do not appear to result in one for startplasma-x11. For reference, | Code: | terra ~ # ulimit -c
unlimited
terra ~ # cat /proc/sys/kernel/core_pattern
/var/cores/core.%e.%p |
If I appear to be on the right track, could someone provide me pointers on how to get a core dump out of startplasma-x11?. If I could get that, I could poke around with gdb and likely see why it's happening.
I'll also mention that this may not be a Plasma problem, but a "John" problem. I have two other systems on which the upgrade was flawless.
Edit: Changed title and moved to Portage and Programming. I am not actually concerned at the moment with Plasma, per se.
- John
_________________
I can confirm that I have received between 0 and 499 National Security Letters.
Last edited by John R. Graham on Wed Sep 04, 2024 1:16 pm; edited 3 times in total |
|
| Back to top |
|
 |
acmondor
n00b
Joined: 08 Aug 2014
Posts: 60
Location: Canadian Prairies
|
| Posted: Tue Sep 03, 2024 2:09 pm Post subject: |
|
|
When I need to capture core dumps from apps not started at the command line I usuallly just put the following in /etc/security/limits.conf:
| Code: | | * soft core unlimited |
The default is:
Hope this helps. |
|
| Back to top |
|
|
asturm
Developer
Joined: 05 Apr 2007
Posts: 9122
|
| Posted: Tue Sep 03, 2024 4:11 pm Post subject: |
|
|
| That's very odd. Did you already try removing ~/.cache dir? |
|
| Back to top |
|
|
John R. Graham
Administrator
Joined: 08 Mar 2005
Posts: 10645
Location: Somewhere over Atlanta, Georgia
|
| Posted: Tue Sep 03, 2024 6:53 pm Post subject: |
|
|
| acmondor wrote: | When I need to capture core dumps from apps not started at the command line I usuallly just put the following in /etc/security/limits.conf:
| Code: | | * soft core unlimited |
|
Thanks! I didn't know that that was one of PAM's jobs. So I did that and I still didn't get a core dump out of startplasma-x11, even when running sddm from the command line. To prove I wasn't missing something basic, I did write the simplest little C program to force a segfault and got a core dump from that, so the basic setup is working.
| asturm wrote: | | That's very odd. Did you already try removing ~/.cache dir? |
Hadn't. Tried just now. Alas, it didn't help.
- John
_________________
I can confirm that I have received between 0 and 499 National Security Letters. |
|
| Back to top |
|
|
John R. Graham
Administrator
Joined: 08 Mar 2005
Posts: 10645
Location: Somewhere over Atlanta, Georgia
|
| Posted: Wed Sep 04, 2024 2:23 pm Post subject: |
|
|
So... it turns out I was missing something basic. Note to self: make sure the directory where you have directed core dumps to be placed is writable to the user running the process. I have a core dump. Analyzing...
- John
_________________
I can confirm that I have received between 0 and 499 National Security Letters. |
|
| Back to top |
|
|
|
Portage & Programming
