| Summary: | *gnome-control-center mouse* causes a segmentation fault (core dumped) | ||
|---|---|---|---|
| Product: | Mageia | Reporter: | Ben McMonagle <westel> |
| Component: | RPM Packages | Assignee: | David GEIGER <geiger.david68210> |
| Status: | NEW --- | QA Contact: | |
| Severity: | normal | ||
| Priority: | Normal | CC: | davidwhodgins, fri, mageia, tmb |
| Version: | Cauldron | Keywords: | IN_ERRATA9 |
| Target Milestone: | --- | ||
| Hardware: | i586 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Source RPM: | gstreamer??? | CVE: | |
| Status comment: | |||
|
Description
Ben McMonagle
2023-05-10 10:20:44 CEST
Could it be related to this bug https://gitlab.gnome.org/GNOME/gnome-control-center/-/issues/2459 ? CC:
(none) =>
geiger.david68210 If yes, is rygel package installed? David, thanks for jumping in on this. Inevitably assigning to you. I have just tried (x64) gnome-control-center under LxQt, and nothing happens! Will try later under Gnome. Source RPM:
(none) =>
gnome-control-center-44.1-1.mga9.src.rpm Lewis, this is specific to i586. no rygel packages installed also, it seems a multi DE install masks the issue. one of *those* failures. (In reply to David GEIGER from comment #1) > Could it be related to this bug > https://gitlab.gnome.org/GNOME/gnome-control-center/-/issues/2459 ? yes, this is, with except that report OP does not specify 32/64 bit. a later comment indicates 32bit So please install rygel to see what's happen, normally it should fix the segfault. # urpmi rygel To satisfy dependencies, the following packages are going to be installed: Package Version Release Arch (medium "Core Release (distrib1)") libges-gir1.0 1.22.0 1.mga9 i586 libges1.0_0 1.22.0 1.mga9 i586 libgupnp-dlna-gst2.0_4 0.12.0 4.mga9 i586 libgupnp-dlna-gstreamer 0.12.0 4.mga9 i586 librygel2.8_0 0.42.3 1.mga9 i586 rygel 0.42.3 1.mga9 i586 5.9MB of additional disk space will be used. 1.5MB of packages will be retrieved. Proceed with the installation of the 6 packages? (Y/n) --------------------------------------------------------------------------------- $ gnome-control-center mouse Segmentation fault (core dumped) relevant output from journalctl -b: May 12 15:43:17 localhost kernel: multiqueue1:src[4096]: segfault at 1 ip b7f3fb1b sp 959e6c18 error 4 in libgobject-2.0.so.0.7600.2[b7f2d000+33000] likely on CPU 0 (core 0, socket 0) May 12 15:43:17 localhost kernel: Code: 44 24 04 c7 40 04 00 00 00 00 c3 8d 74 26 00 8b 44 24 04 8b 40 04 c3 8d b4 26 00 00 00 00 90 53 8b 44 24 08 8b 58 04 8b 48 10 <0f> be 03 84 c0 74 1a 8d b6 00 00 00 00 89 ca 83 c3 01 c1 e2 05 29 I'm wondering if this is an issue of similar to sse2 support and compiler flags. What's the output of the flags line for one of the cpus listed in /proc/cpuinfo? Adding Thomas to the cc list. CC:
(none) =>
davidwhodgins, tmb (In reply to Dave Hodgins from comment #11) > I'm wondering if this is an issue of similar to sse2 support and compiler > flags. > > What's the output of the flags line for one of the cpus listed in > /proc/cpuinfo? > > Adding Thomas to the cc list. all cpus so far have been either x86_64 capable or Virtualbox 64bit cpu the issue is repeatable on real h/w or virtual h/w will post some cpu info later not sure if this is what (In reply to Dave Hodgins from comment #11) > I'm wondering if this is an issue of similar to sse2 support and compiler > flags. > > What's the output of the flags line for one of the cpus listed in > /proc/cpuinfo? > > Adding Thomas to the cc list. not sure if this is what you requested: lscpu Architecture: i686 CPU op-mode(s): 32-bit, 64-bit Address sizes: 36 bits physical, 48 bits virtual Byte Order: Little Endian CPU(s): 4 On-line CPU(s) list: 0-3 Vendor ID: GenuineIntel BIOS Vendor ID: Intel(R) Corporation Model name: Intel(R) Core(TM) i3 CPU M 350 @ 2.27GHz BIOS Model name: Intel(R) Core(TM) i3 CPU M 350 @ 2.27GHz CPU @ 2.2GHz BIOS CPU family: 206 CPU family: 6 Model: 37 Thread(s) per core: 2 Core(s) per socket: 2 Socket(s): 1 Stepping: 2 CPU(s) scaling MHz: 55% CPU max MHz: 2266.0000 CPU min MHz: 933.0000 BogoMIPS: 4521.51 Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mc a cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ht tm pbe nx rdtscp lm constant_tsc arch_perfmon pebs bts xtopology nonstop_tsc cpuid aperfmperf pni dtes64 monit or ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm sse4_1 sse4_ 2 popcnt lahf_lm pti ssbd ibrs ibpb stibp tpr_shadow vn mi flexpriority ept vpid dtherm arat flush_l1d The bug is in gstreamer. I downloaded one of the gnome-control-center animations: https://gitlab.gnome.org/Teams/Design/settings-mockups/-/blob/master/mouse-and-touchpad/assets/scroll-natural-l.webm and got the same segfault when executing this command: gst-play-1.0 scroll-natural-l.webm However this command: gst-launch-1.0 filesrc location=scroll-natural-l.webm ! matroskademux ! vp9dec ! autovideosink plays the animation without any problem, so it seems likely this is a bug in the base gstreamer package and not in the plugins. Source RPM:
gnome-control-center-44.1-1.mga9.src.rpm =>
gstreamer??? Looking at Martin's detective work (and where on earth did you dig up the second command?!): $ urpmq --requires-recursive gnome-control-center | grep gstreamer gstreamer1.0-libav gstreamer1.0-plugins-bad gstreamer1.0-plugins-base gstreamer1.0-plugins-good gstreamer1.0-plugins-ugly gstreamer1.0-soup gstreamer1.0-tools lib64gstreamer-plugins-base-gir1.0 lib64gstreamer-plugins-base1.0_0 lib64gstreamer1.0_0 lib64gupnp-dlna-gstreamer $ urpmq --requires-recursive mate-notification-daemon | grep gstreamer gstreamer1.0-tools lib64gstreamer-plugins-base1.0_0 lib64gstreamer1.0_0 No sign of the base gstreamer pkg, but the last three in common at least. ------------------------------------------------------------------------- DavidG wonders whether this bug is the same as bug 31928, but there are differences: * Ben thinks this bug is 32-bit only (comment 4, comment 7), Martin does not say. The newer bug is presumed x64. * Ben says this bug only happens with a Gnome-only installation (comment 6). Martin does not say. The newer bug is for Mate, not yet known whether that also has Gnome. But the desktops have much in common. I only see this bug in a 32-bit install. I see the same fault in a Mate-only install. Oddly the gst-launch command doesn't work there, but nor does it work in a GNOME on Xorg session. That looks to be a different problem, and may just be a gstreamer usage issue. On investigation, the gst-play-1.0 command comes from the gstreamer1.0-plugins-base package. Thanks for those clarifications: - 32-bit [so perhaps comment 11 re sse2 ?] - single desktop On a multi-desktop setup, x64, under Gnome, the fault does *not* happen. Same system, Mate, 'gnome-control-center' just exits immediately! See nothing in the journal. Keywords:
(none) =>
FOR_ERRATA9 still valid: ISO Name: Mageia-9-rc1-i586.iso DATE.txt: Mon Jul 3 06:29:42 PM CEST 2023 https://wiki.mageia.org/en/Mageia_9_Errata#GNOME Keywords:
FOR_ERRATA9 =>
IN_ERRATA9 valid for Mga10 cauldron |