Description of problem: Tried to log in to a gnome-wayland session from gdm. Version-Release number of selected component: gnome-shell-3.15.91-1.fc22 Additional info: reporter: libreport-2.4.0 backtrace_rating: 4 cmdline: /usr/bin/gnome-shell --wayland --display-server crash_function: output_free executable: /usr/bin/gnome-shell kernel: 4.0.0-0.rc1.git0.1.fc22.x86_64 runlevel: N 5 type: CCpp uid: 1006 Truncated backtrace: Thread no. 1 (10 frames) #0 output_free at winsys/cogl-winsys-egl-kms.c:728 #1 _cogl_winsys_egl_display_destroy at winsys/cogl-winsys-egl-kms.c:765 #2 _cogl_winsys_display_destroy at winsys/cogl-winsys-egl.c:441 #3 _cogl_winsys_display_setup at winsys/cogl-winsys-egl.c:484 #4 cogl_display_setup at ./cogl-display.c:154 #5 cogl_renderer_check_onscreen_template at ./cogl-renderer.c:369 #6 clutter_backend_real_create_context at /lib64/libclutter-1.0.so.0 #7 _clutter_feature_init at /lib64/libclutter-1.0.so.0 #8 clutter_init_real at /lib64/libclutter-1.0.so.0 #9 post_parse_hook at /lib64/libclutter-1.0.so.0 Potential duplicate: bug 1015677
Created attachment 999982 [details] File: backtrace
Created attachment 999983 [details] File: cgroup
Created attachment 999984 [details] File: core_backtrace
Created attachment 999985 [details] File: dso_list
Created attachment 999986 [details] File: environ
Created attachment 999987 [details] File: limits
Created attachment 999988 [details] File: maps
Created attachment 999989 [details] File: open_fds
Created attachment 999990 [details] File: proc_pid_status
Created attachment 999991 [details] File: var_log_messages
Same here on Thinkpad t440p with intel 4600 gpu
Same system. The system was updated from F21. Now it is unusable.
I have found on google+ an workaround that let gnome via X11 work. Simply yum remove xorg-x11-drv-libinput
Tested today again. Same problem, same workaround.
Solved by last mutter update.
FWIW, this crash pattern is reliably reproducible here when the computer (a HTPC in the making, ZOTAC Nano C321, Celeron 2961Y) starts with the TV (attached to the HDMI port) off. The same combo works fine if the TV is on on startup or after a gdm service restart.
Another user experienced a similar problem: Happens on any attempt to login from GDM to gnome on Wayland session. Standard gnome session works fine. reporter: libreport-2.6.1 backtrace_rating: 4 cmdline: /usr/bin/gnome-shell --wayland --display-server crash_function: output_free executable: /usr/bin/gnome-shell global_pid: 10050 kernel: 4.0.7-300.fc22.x86_64 package: gnome-shell-3.16.3-1.fc22 reason: gnome-shell killed by SIGSEGV runlevel: N 5 type: CCpp uid: 1000
Cannot reproduce the problem, but I think that this could solve the segfault: https://bugzilla.gnome.org/show_bug.cgi?id=752406 Anyway, if we solve the segfault the gnome won't start either, because according to backtrace something went wrong - it is either that platform display initialization failed or creating cogl context failed. But we'll get some nice error message and will know where to look next.
I'm willing to test your patch if there are instructions on how to compile it and install on my laptop.
Thanks for help, Here are rpms for F22 built with this patch: https://kojipkgs.fedoraproject.org/work/tasks/2179/10412179/
I dowloaded cogl-1.20.0-3.fc22.x86_64.rpm but dnf said nothing to do. I already seem to have that version. Maybe already in updates-testing?
Yes, this rpm is the same version as the one in your system. You must use reinstall command: $ sudo dnf reinstall path_to_rpm
I did that and logged out and then selected gnome on wayland. Looks like there is still a problem as it returns to the gdm login screen. Logging in to the standard gnome session abrt is displaying the following: GNOME Shell quit unexpectedly The application encountered a problem and could not continue. When I try to report the problem it fails as follows: --- Running report_uReport --- Unable to open './pkg_epoch': No such file or directory. reporter-ureport failed with exit code 1 ('report_uReport' exited with 1) Then I click "upload for analysis: --- Running report_EmergencyAnalysis --- Compressing data Sending /var/tmp/ccpp-2015-07-21-21:42:53-1445.new.tar.gz to https://retrace.fedoraproject.org/faf/dumpdirs/new/ Successfully sent /var/tmp/ccpp-2015-07-21-21:42:53-1445.new.tar.gz to https://retrace.fedoraproject.org/faf/dumpdirs/new/ <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN"> <title>Redirecting...</title> <h1>Redirecting...</h1> So maybe abrt isn't working properly?
Could you please attach output of journalctl after the crash? Optionally, you could switch to tty and try running this: $ gnome-session --session=gnome-wayland 2>&1 > gnome-session.log and $ strace gnome-shell --wayland --display-server 2>&1 > gnome-shell.strace (if you have strace installed) and attach the gnome-session.log and gnome-shell.strace Thanks
sorry, the commands should be $ gnome-shell --wayland --display-server 2>gnome-shell.log $ strace gnome-shell --wayland --display-server 2>gnome-shell.strace actually, the one with strace is enough
Created attachment 1054824 [details] strace output requested strace
The gnome-shell.log output was just: (gnome-shell:5492): Clutter-CRITICAL **: Unable to initialize Clutter: No currently active connector found (gnome-shell:5492): mutter-ERROR **: Unable to initialize Clutter.
Nice, now without the segfault we know what caused the gnome-shell to give up. According to the strace it tried to connect to only one connector and it could fail either here: http://cgit.freedesktop.org/mesa/drm/tree/xf86drmMode.c?id=2.4.61#n561 or here: https://git.gnome.org/browse/cogl/tree/cogl/winsys/cogl-winsys-egl-kms.c#n389
Could that be because my laptop has hybrid graphics. I believe it has Intel and Radeon but only the Intel driver is loaded? glxinfo | grep renderer GLX_MESA_multithread_makecurrent, GLX_MESA_query_renderer, GLX_MESA_multithread_makecurrent, GLX_MESA_query_renderer, OpenGL renderer string: Mesa DRI Intel(R) Haswell Mobile sudo lshw -C display *-display description: VGA compatible controller product: 4th Gen Core Processor Integrated Graphics Controller vendor: Intel Corporation physical id: 2 bus info: pci@0000:00:02.0 version: 06 width: 64 bits clock: 33MHz capabilities: msi pm vga_controller bus_master cap_list rom configuration: driver=i915 latency=0 resources: irq:31 memory:f5800000-f5bfffff memory:d0000000-dfffffff ioport:f000(size=64) I don't know how to interpret the strace but I did see some references to radeon in there?
> I don't know how to interpret the strace but I did see some references to > radeon in there? Yep, it looks like egl is using radeon card (which is card0). You can try export DRI_PRIME=1 (or DRI_PRIME=pci-0000_00_02_0 ) and then try to run gnome session. Anyway, it looks like mesa's bug now, so I filled a bug there. There are people that know much more about it than me :) https://bugs.freedesktop.org/show_bug.cgi?id=91474
The patch fixing segfault was pushed to cogl master. Looking into the code reveals that it is not probably mesa's bug, but mutter's. It always uses /dev/dri/card0, which is the radeon one in this case. Then it makes clutter and subsequently cogl to use this card. https://git.gnome.org/browse/mutter/tree/src/backends/native/meta-launcher.c#n270
https://bugzilla.gnome.org/show_bug.cgi?id=753434
Patches that should fix this were pushed to master 79f755bf0fa
Yes. I have been following the other thread on freedesktop.org. Thanks for your efforts on this. Even getting involved in another project you weren't all that familiar with. I presume this will arrive in fedora updates testing at some stage or will I have to wait for fedora 23?
In the mean time I upgraded to fedora 23 (pre beta) which still had the problem with selecting the wrong graphics card. However a new bunch of f23 updates arrived today including mutter, and various other gnome related stuff and I am happy to report that the problem appears to solved. I am posting this from a gnome on wayland session (f23 beta +). Well done and thanks to those involved in getting this sorted out.
*** Bug 1134133 has been marked as a duplicate of this bug. ***
Fedora 22 changed to end-of-life (EOL) status on 2016-07-19. Fedora 22 is no longer maintained, which means that it will not receive any further security or bug fix updates. As a result we are closing this bug. If you can reproduce this bug against a currently maintained version of Fedora please feel free to reopen this bug against that version. If you are unable to reopen this bug, please file a new report against the current release. If you experience problems, please add a comment to this bug. Thank you for reporting this bug and we are sorry it could not be fixed.