Bug #545397 “gnome-screensaver-gl-helper crashed with SIGSEGV in...” : Bugs : nvidia-graphics-drivers package : Ubuntu

Revision history for this message

Tkachenko Bogdan (bogus-bogus) wrote on 2010-03-23:

#1

Dependencies.txt Edit (4.7 KiB, text/plain; charset="utf-8")
Disassembly.txt Edit (596 bytes, text/plain; charset="utf-8")
GconfGnomeLockdown.txt Edit (212 bytes, text/plain; charset="utf-8")
GconfGnomePowerManager.txt Edit (1.9 KiB, text/plain; charset="utf-8")
GconfGnomeScreensaver.txt Edit (414 bytes, text/plain; charset="utf-8")
GconfGnomeSession.txt Edit (238 bytes, text/plain; charset="utf-8")
ProcMaps.txt Edit (21.1 KiB, text/plain; charset="utf-8")
ProcStatus.txt Edit (784 bytes, text/plain; charset="utf-8")
Registers.txt Edit (470 bytes, text/plain; charset="utf-8")
Stacktrace.txt Edit (438 bytes, text/plain; charset="utf-8")
ThreadStacktrace.txt Edit (465 bytes, text/plain; charset="utf-8")
XorgConf.txt Edit (213 bytes, text/plain; charset="utf-8")
XorgLog.txt Edit (28.3 KiB, text/plain; charset="utf-8")
XsessionErrors.txt Edit (1.9 KiB, text/plain; charset="utf-8")

visibility:

private → public

Pedro Villavicencio (pedro) on 2010-03-24

Changed in gnome-screensaver (Ubuntu):
importance:	Undecided → Medium

Revision history for this message

Apport retracing service (apport) wrote on 2010-03-30:

#2

StacktraceTop:
?? () from /usr/lib/nvidia-current/libGL.so.1
__run_exit_handlers (status=-1216857360, listp=0x1de0324,
*__GI_exit (status=0) at exit.c:100
__libc_start_main (main=0x8048cd0 <main>, argc=1,
_start ()

Revision history for this message

Apport retracing service (apport) wrote on 2010-03-30: Stacktrace.txt

#3

Stacktrace.txt Edit (966 bytes, text/plain)

Revision history for this message

Apport retracing service (apport) wrote on 2010-03-30: ThreadStacktrace.txt

#4

ThreadStacktrace.txt Edit (994 bytes, text/plain)

tags:

removed: need-i386-retrace

Revision history for this message

PJO (lexicographer) wrote on 2010-04-03:

#5

Got a crash when trying to load Nvidia drivers for a Gigabyte motherboard with 630i chipset. Worked on original 10.04 beta install but not now on a reinstallation. Tried both drivers offered

2010-04-03 15:42:56,513 DEBUG: Instantiated Handler subclass __builtin__.NvidiaDriverCurrent from name NvidiaDriverCurrent
2010-04-03 15:42:56,524 DEBUG: NVIDIA accelerated graphics driver availability undetermined, adding to pool
2010-04-03 15:42:56,525 DEBUG: Could not instantiate Handler subclass __builtin__.NvidiaDriverBase from name NvidiaDriverBase
Traceback (most recent call last):
File "/usr/lib/python2.6/dist-packages/jockey/detection.py", line 914, in get_handlers
inst = obj(backend)
TypeError: __init__() takes exactly 3 arguments (2 given)
2010-04-03 15:42:56,536 WARNING: modinfo for module nvidia_173 failed: ERROR: modinfo: could not find module nvidia_173

Revision history for this message

PJO (lexicographer) wrote on 2010-04-03:

#6

jockey.log Edit (104.9 KiB, text/plain)

NOTE ERRONEOUS # OF ARGUMENTS ABOVE

and for the latest driver:

Traceback (most recent call last):
  File "/usr/lib/python2.6/dist-packages/jockey/detection.py", line 914, in get_handlers
    inst = obj(backend)
  File "/usr/share/jockey/handlers/fglrx.py", line 16, in __init__
    raise SystemError, 'does not currently work with X.org 1.7'
SystemError: does not currently work with X.org 1.7

jockey.log attached

Revision history for this message

PJO (lexicographer) wrote on 2010-04-03:

#7

Made some progress on this.

My initial install of 10.04 beta 1 -- the first Lucid version I tried -- wouldn't do 1920x1200 video with an ATEN CS1734B KVM switch in place. The switch, although it supports monitor resolutions well in excess of this, prevents the detection of some monitor parameters and the display defaults to a lower resolution. This seems to have been the cause of some problems, for me anyway.

Here http://ohioloco.ubuntuforums.org/showthread.php?t=1435151 there's a reference to the driver being unable to

> identify the correct EDID info,

which reminded me of this.

First time around I ended up installing with a directly connected monitor (hardware drivers installed ok) and then getting xorg.conf configured to use 1920x1200 to eliminate black borders caused by the system seeing a lower resolution display. It was because I was having difficulty getting1920x1200 that I did a clean install with the monitor directly attached.

Once I had the xorg.conf file figured out (or borrowed to be precise; see blog posts) I didn't think I'd need to unplug from the switch again. I thought I could just reuse that xorg.conf file. Wrong! And furthermore, I couldn't install the drivers until after I'd done a system update (not the case last time as far as I remember). Doing it the other way around led to a crash, and this time the drivers were installed automatically after the system update (nice).

http://wombatdiet.net/2010/03/25/first-impressions-ubuntu-10-04/
http://wombatdiet.net/2010/03/19/fedora-12-and-the-aten-cs1734b/

I haven't been able to get xorg.conf adjusted for use with Fedora 12 yet (with a Samsung 2443NW); the Ubuntu config doesn't work.

ATEN have previously blamed my monitor (a different one) for Ubuntu not getting the resolution right, but I now know and I've sent them xorg log files made with and without the switch in place, that what's detected is very different with the switch in use. Until today however, that didn't cause any crashes.

Bottom line is, I think, not to use a KVM switch during installation, just in case.

Message for Canonical: I would gladly give preference to Ubuntu certified hardware.

Made some progress on this.

My initial install of 10.04 beta 1 -- the first Lucid version I tried -- wouldn't do 1920x1200 video with an ATEN CS1734B KVM switch in place. The switch, although it supports monitor resolutions well in excess of this, prevents the detection of some monitor parameters and the display defaults to a lower resolution. This seems to have been the cause of some problems, for me anyway.

Here http://ohioloco.ubuntuforums.org/showthread.php?t=1435151 there's a reference to the driver being unable to

> identify the correct EDID info,

which reminded me of this.

First time around I ended up installing with a directly connected monitor (hardware drivers installed ok) and then getting xorg.conf configured to use 1920x1200 to eliminate black borders caused by the system seeing a lower resolution display. It was because I was having difficulty getting1920x1200 that I did a clean install with the monitor directly attached.

Once I had the xorg.conf file figured out (or borrowed to be precise; see blog posts) I didn't think I'd need to unplug from the switch again. I thought I could just reuse that xorg.conf file. Wrong! And furthermore, I couldn't install the drivers until after I'd done a system update (not the case last time as far as I remember). Doing it the other way around led to a crash, and this time the drivers were installed automatically after the system update (nice).

http://wombatdiet.net/2010/03/25/first-impressions-ubuntu-10-04/
http://wombatdiet.net/2010/03/19/fedora-12-and-the-aten-cs1734b/

I haven't been able to get xorg.conf adjusted for use with Fedora 12 yet (with a Samsung 2443NW); the Ubuntu config doesn't work.

ATEN have previously blamed my monitor (a different one) for Ubuntu not getting the resolution right, but I now know and I've sent them xorg log files made with and without the switch in place, that what's detected is very different with the switch in use. Until today however, that didn't cause any crashes.

Bottom line is, I think, not to use a KVM switch during installation, just in case.

Message for Canonical: I would gladly give preference to Ubuntu certified hardware.

Revision history for this message

Michael T (michaeltandy) wrote on 2010-06-11:

#8

Crash report log file Edit (8.7 MiB, text/plain)

I have encountered a crash with symptoms matching this one, while running from the 10.04 32-bit live CD.

After starting the live CD, a message displayed suggesting I could install the proprietary nvidia drivers, which I agreed to, selecting the recommended (latest) version - but when it asked me to reboot I realised that while using a live-cd (presumably without persistent storage) rebooting would not in fact cause those drivers to be loaded. Instead, I left the computer running without having rebooted.

I left the computer running in this state; when I returned to it later, apport informed me of a crash. I submitted the report and thought no more about it.

Later, I left the computer again, and returned to find the same error message displayed. I have attached the apport report.

The computer I was doing this on was an NEC WA2510 with specification including 2 AMD Opteron 2218 2.6GHz processors and an nVidia Quadro FX 560 video card running two CRT monitors.

Pedro Villavicencio (pedro) on 2010-06-23

affects:

gnome-screensaver (Ubuntu) → nvidia-graphics-drivers (Ubuntu)

Revision history for this message

dino99 (9d9) wrote on 2015-06-26:

#9

That version is no more supported; please open a new bug report if the actual archive found version also has the same issue.

Changed in nvidia-graphics-drivers (Ubuntu):
status:	New → Invalid

Ubuntu
nvidia-graphics-drivers package

gnome-screensaver-gl-helper crashed with SIGSEGV in exit()

Bug Description

Other bug subscribers

Bug attachments

Remote bug watches

Ubuntunvidia-graphics-drivers package

gnome-screensaver-gl-helper crashed with SIGSEGV in exit()

Bug Description

Other bug subscribers

Bug attachments

Remote bug watches

Ubuntu
nvidia-graphics-drivers package