[Bug 1036342] New: [20170424]: gnome-session segfault in libgtk-3.so.0.2200.11
http://bugzilla.opensuse.org/show_bug.cgi?id=1036342 Bug ID: 1036342 Summary: [20170424]: gnome-session segfault in libgtk-3.so.0.2200.11 Classification: openSUSE Product: openSUSE Tumbleweed Version: Current Hardware: Other OS: Other Status: NEW Severity: Normal Priority: P5 - None Component: GNOME Assignee: bnc-team-gnome@forge.provo.novell.com Reporter: markos.chandras@suse.com QA Contact: qa-bugs@suse.de Found By: --- Blocker: --- Hi, Ever since I updated to 20170424 I can't get gnome session to start. gdm is stuck in an infinite loop and I can see the following issue on dmesg [ 209.298800] gnome-session-f[2668]: segfault at 0 ip 00007fb96a6e78a9 sp 00007ffc3a8734b0 error 4 in libgtk-3.so.0.2200.11[7fb96a3f8000+708000] [ 256.323964] gnome-session-f[3378]: segfault at 0 ip 00007f7a6bfe08a9 sp 00007ffd9c426de0 error 4 in libgtk-3.so.0.2200.11[7f7a6bcf1000+708000] I can't tell if the previous snapshot worked for me because I haven't update in a while. Please let me know if you need more information from me -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1036342 http://bugzilla.opensuse.org/show_bug.cgi?id=1036342#c1 Vasilis Liaskovitis <vliaskovitis@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |vliaskovitis@suse.com --- Comment #1 from Vasilis Liaskovitis <vliaskovitis@suse.com> --- Can you share the gnome-session coredump backtrace? If coredumpctl is enabled, "coredumpctl list" should show the core dumps generated, and "coredumpctl gdb PID" should be able to give you the backtrace. Some debuginfo packages may need to be installed for symbol info. Is wayland enabled (i think it should be on by default in recent TW) or is this an Xorg session? Is gnome-session-wayland installed? I can't reproduce this on an updated TW. But I think I have seen a gnome-session fail/retry loop in the past when doing "systemctl restart display-manager.service". But restarting the display-manager has other problems currently for me (gnome-shell segfaults (bsc#1036432), or input-device problems /screen tearing) -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1036342 http://bugzilla.opensuse.org/show_bug.cgi?id=1036342#c2 Dominique Leuenberger <dimstar@opensuse.org> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |dimstar@opensuse.org, | |markos.chandras@suse.com Flags| |needinfo?(markos.chandras@s | |use.com) --- Comment #2 from Dominique Leuenberger <dimstar@opensuse.org> --- (In reply to Markos Chandras from comment #0)
Hi,
Ever since I updated to 20170424 I can't get gnome session to start. gdm is stuck in an infinite loop and I can see the following issue on dmesg
[ 209.298800] gnome-session-f[2668]: segfault at 0 ip 00007fb96a6e78a9 sp 00007ffc3a8734b0 error 4 in libgtk-3.so.0.2200.11[7fb96a3f8000+708000] [ 256.323964] gnome-session-f[3378]: segfault at 0 ip 00007f7a6bfe08a9 sp 00007ffd9c426de0 error 4 in libgtk-3.so.0.2200.11[7f7a6bcf1000+708000]
I can't tell if the previous snapshot worked for me because I haven't update in a while. Please let me know if you need more information from me
That seems not to be gnome-session crashing, but actually gnome-session-failed, which means even if it would not crash, you'd not be happy: you would simply see the 'Something has gone wrong" screen. If you are running in a wayland session, then it is known that gnome-session-failed does not work (if gnome-shell crashed/failed to start, there is no wayland compositor and hence nothing gtk3 could paint on) So in short: the crash you see is 'not nice', but is not the root of you not having a GNOME session running. Please get more information from your journal to actually trace the root cause -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1036342 http://bugzilla.opensuse.org/show_bug.cgi?id=1036342#c3 --- Comment #3 from Dominique Leuenberger <dimstar@opensuse.org> --- (In reply to Dominique Leuenberger from comment #2)
If you are running in a wayland session, then it is known that gnome-session-failed does not work (if gnome-shell crashed/failed to start, there is no wayland compositor and hence nothing gtk3 could paint on)
For reference, this would be https://bugzilla.gnome.org/show_bug.cgi?id=775337 -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1036342 http://bugzilla.opensuse.org/show_bug.cgi?id=1036342#c4 Markos Chandras <markos.chandras@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(markos.chandras@s | |use.com) | --- Comment #4 from Markos Chandras <markos.chandras@suse.com> --- (In reply to Dominique Leuenberger from comment #3)
(In reply to Dominique Leuenberger from comment #2)
If you are running in a wayland session, then it is known that gnome-session-failed does not work (if gnome-shell crashed/failed to start, there is no wayland compositor and hence nothing gtk3 could paint on)
For reference, this would be https://bugzilla.gnome.org/show_bug.cgi?id=775337
OK thank you for the link. However I am not using wayland. I now use sddm as display-manager and select Gnome on Xorg there and it still fails. In the coredumpctl list I see 2 entries with identical timestamps. One is for gnome-session-failed and the other is for Xorg. I guess what matters the most here is the Xorg one so here it is #0 0x00007f996bae64c7 in x86_64_fallback_frame_state (context=0x7ffd3a329180, context=0x7ffd3a329180, fs=0x7ffd3a329270) at ./md-unwind-support.h:58 #1 0x00007f996bae64c7 in uw_frame_state_for (context=context@entry=0x7ffd3a329180, fs=fs@entry=0x7ffd3a329270) at ../../../libgcc/unwind-dw2.c:1257 #2 0x00007f996bae8078 in _Unwind_Backtrace (trace=0x7f996dd17490 <backtrace_helper>, trace_argument=0x7ffd3a329430) at ../../../libgcc/unwind.inc:290 #3 0x00007f996dd17612 in __GI___backtrace (array=array@entry=0x7ffd3a329460, size=size@entry=64) at ../sysdeps/x86_64/backtrace.c:110 #4 0x000000000059b71a in xorg_backtrace () at backtrace.c:125 #5 0x000000000059f4c9 in OsSigHandler (signo=11, sip=0x7ffd3a329830, unused=<optimized out>) at osinit.c:136 #6 0x00007f996dfd4230 in <signal handler called> () at /lib64/libpthread.so.0 #7 0x0000000000000024 in () #8 0x00007f996f27de3a in drmHandleEvent (fd=24, evctx=0x17e13b8) at xf86drmMode.c:923 #9 0x000000000059ff81 in ospoll_wait (ospoll=0x179bd00, timeout=<optimized out>) at ospoll.c:412 #10 0x0000000000598eec in WaitForSomething (are_ready=<optimized out>) at WaitFor.c:226 #11 0x000000000043a71a in Dispatch () at dispatch.c:422 #12 0x000000000043e968 in dix_main (argc=11, argv=0x7ffd3a32af38, envp=<optimized out>) at main.c:287 #13 0x00007f996dc3d541 in __libc_start_main (main=0x428610 <main>, argc=11, argv=0x7ffd3a32af38, init=<optimized out>, fini=<optimized out>, rtld_fini=<optimized out>, stack_end=0x7ffd3a32af28) at ../csu/libc-start.c:295 #14 0x000000000042864a in _start () at ../sysdeps/x86_64/start.S:120 Here are the related journal entries Apr 27 09:09:17 aegean sddm[2284]: Running: /usr/bin/X -nolisten tcp -auth /run/sddm/{6cae7b1a-6d96-4e50-b6a9-84a8fbb42d2d} -background none -noreset -displayfd 19 vt7 Apr 27 09:09:17 aegean gnome-session-binary[30730]: WARNING: App 'org.gnome.SettingsDaemon.A11yKeyboard.desktop' exited with code 1 Apr 27 09:09:17 aegean gnome-session-binary[30730]: WARNING: App 'org.gnome.SettingsDaemon.Clipboard.desktop' exited with code 1 Apr 27 09:09:17 aegean gnome-session-binary[30730]: WARNING: App 'org.gnome.SettingsDaemon.Power.desktop' exited with code 1 Apr 27 09:09:17 aegean gnome-session-binary[30730]: WARNING: App 'org.gnome.SettingsDaemon.MediaKeys.desktop' exited with code 1 Apr 27 09:09:17 aegean gnome-session-binary[30730]: WARNING: App 'org.gnome.SettingsDaemon.Keyboard.desktop' exited with code 1 Apr 27 09:09:17 aegean gnome-session-binary[30730]: WARNING: App 'org.gnome.SettingsDaemon.PrintNotifications.desktop' exited with code 1 Apr 27 09:09:17 aegean polkitd[1744]: Unregistered Authentication Agent for unix-session:5 (system bus name :1.263, object path /org/freedesktop/PolicyKit1/AuthenticationAgent, locale en_US.UTF-8) (disconnected from bus) Apr 27 09:09:17 aegean gnome-session[30730]: gnome-session-binary[30730]: WARNING: App 'org.gnome.Shell.desktop' exited with code 1 Apr 27 09:09:17 aegean gnome-session-binary[30730]: WARNING: App 'org.gnome.Shell.desktop' exited with code 1 Apr 27 09:09:17 aegean systemd-coredump[31110]: Process 30703 (X) of user 0 dumped core. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1036342 http://bugzilla.opensuse.org/show_bug.cgi?id=1036342#c5 Dominique Leuenberger <dimstar@opensuse.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Component|GNOME |X.Org Assignee|bnc-team-gnome@forge.provo. |xorg-maintainer-bugs@forge. |novell.com |provo.novell.com QA Contact|qa-bugs@suse.de |xorg-maintainer-bugs@forge. | |provo.novell.com --- Comment #5 from Dominique Leuenberger <dimstar@opensuse.org> --- (In reply to Markos Chandras from comment #4)
OK thank you for the link. However I am not using wayland. I now use sddm as display-manager and select Gnome on Xorg there and it still fails. In the coredumpctl list I see 2 entries with identical timestamps. One is for gnome-session-failed and the other is for Xorg. I guess what matters the most here is the Xorg one so here it is
Right - with X crashing away, there is no way gnome-session-failed could paint the fail screen there. Assigning component X.org -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1036342 http://bugzilla.opensuse.org/show_bug.cgi?id=1036342#c6 Stefan Dirsch <sndirsch@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags| |needinfo?(markos.chandras@s | |use.com) --- Comment #6 from Stefan Dirsch <sndirsch@suse.com> --- Ok. First things first. Please attach /var/log/Xorg.0.log or wherever sddm writes the X log into. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1036342 Stefan Dirsch <sndirsch@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Priority|P5 - None |P3 - Medium Status|NEW |CONFIRMED -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1036342 http://bugzilla.opensuse.org/show_bug.cgi?id=1036342#c7 Markos Chandras <markos.chandras@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(markos.chandras@s | |use.com) | --- Comment #7 from Markos Chandras <markos.chandras@suse.com> --- Created attachment 722939 --> http://bugzilla.opensuse.org/attachment.cgi?id=722939&action=edit xorg log This is the log generated after the crash -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1036342 http://bugzilla.opensuse.org/show_bug.cgi?id=1036342#c8 --- Comment #8 from Stefan Dirsch <sndirsch@suse.com> --- Hmm. nouveau driver segfaults apparently. Truncated backtrace. Not much we can do about nouveau. I suggest uninstalling Mesa-dri-nouveau package (fallback to software rendering). -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1036342 http://bugzilla.opensuse.org/show_bug.cgi?id=1036342#c9 --- Comment #9 from Markos Chandras <markos.chandras@suse.com> --- (In reply to Stefan Dirsch from comment #8)
Hmm. nouveau driver segfaults apparently. Truncated backtrace. Not much we can do about nouveau. I suggest uninstalling Mesa-dri-nouveau package (fallback to software rendering).
Yeah removing that packages brought back my gnome session. Thanks. Not sure if it helps or not, but plasma5 works fine with that package installed so this feels like a gnome issue to me :/ -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1036342 Stefan Dirsch <sndirsch@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Summary|[20170424]: gnome-session |[20170424]: nouveau DRI |segfault in |driver: gnome-session |libgtk-3.so.0.2200.11 |segfault in | |libgtk-3.so.0.2200.11 -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1036342 http://bugzilla.opensuse.org/show_bug.cgi?id=1036342#c10 Mark Draheim <rickscafe.casablanca@gmx.net> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |rickscafe.casablanca@gmx.ne | |t --- Comment #10 from Mark Draheim <rickscafe.casablanca@gmx.net> --- (In reply to Markos Chandras from comment #9)
Not sure if it helps or not, but plasma5 works fine with that package installed so this feels like a gnome issue to me :/
and I almost thought that your problem is essentially the same as in https://bugzilla.opensuse.org/show_bug.cgi?id=1017478 ie nouveau 1.0.14-2.1 and X dumping core early. In your initial report you said that gdm loops. How then did you get a plasma5 session up? The reason I ask is that the bug I referenced points a finger at sddm/QT/Plasma but if you see the same problem with Gnome/GTK, then we can narrow it down to Xorg/nouveau. Which I think is highly likely given that similar reports from nouveau users pop up in the forums almost daily now. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1036342 http://bugzilla.opensuse.org/show_bug.cgi?id=1036342#c11 --- Comment #11 from Markos Chandras <markos.chandras@suse.com> --- (In reply to Mark Draheim from comment #10)
(In reply to Markos Chandras from comment #9)
Not sure if it helps or not, but plasma5 works fine with that package installed so this feels like a gnome issue to me :/
and I almost thought that your problem is essentially the same as in
https://bugzilla.opensuse.org/show_bug.cgi?id=1017478
ie nouveau 1.0.14-2.1 and X dumping core early.
In your initial report you said that gdm loops. How then did you get a plasma5 session up?
I simply switched to sddm to try things out :) I am unable to start a gnome session from sddm for the reasons mentioned above but at least sddm does not loop like gdm does which allows me to select the plasma5 session instead. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1036342 http://bugzilla.opensuse.org/show_bug.cgi?id=1036342#c12 --- Comment #12 from Mark Draheim <rickscafe.casablanca@gmx.net> --- (In reply to Markos Chandras from comment #11)
I simply switched to sddm to try things out :)
interesting, maybe I should switch to gdm then because here sddm loops ;-) But I will try a local rebuild of nouveau package first. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1036342 http://bugzilla.opensuse.org/show_bug.cgi?id=1036342#c13 --- Comment #13 from Mark Draheim <rickscafe.casablanca@gmx.net> --- @Markos: Can you update to current TW's nouveau 1.0.15 and Mesa-dri-nouveau? My problem from bsc#1017478 was fixed by the new packages. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1036342 http://bugzilla.opensuse.org/show_bug.cgi?id=1036342#c14 Markos Chandras <markos.chandras@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|CONFIRMED |RESOLVED Resolution|--- |FIXED --- Comment #14 from Markos Chandras <markos.chandras@suse.com> --- (In reply to Mark Draheim from comment #13)
@Markos: Can you update to current TW's nouveau 1.0.15 and Mesa-dri-nouveau? My problem from bsc#1017478 was fixed by the new packages.
yeah seems like today's update fixed everything again so I am closing this bug -- You are receiving this mail because: You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠novell.com