[opensuse] Is this a Firefox bug?
Since the ... maybe upgrade yo kde4.10.4 or something else in the regular updates, but before the system upgrade 12.2->12.3 .. Sometimes .. only sometimes ... When I click on a URL in Thunderbird to go to that like in Firefox there is a small pause, a freeze while the pointer won't move and then the screen turns to hash. At that point the machine is dead-locked up. Can't hot-key to another screen/activity, can't telnet or ssh into it, it doesn't respond to pings. So I wonder, is this a problem with -- Firefox -- KDE4 -- X or the video driver. MozillaFirefox-21.0-1.18.1.i586 MozillaFirefox-branding-openSUSE-21-2.5.1.i586 libkde4-4.10.4-10.2.i586 kdebase4-openSUSE-12.3-10.41.9.i586 kdebase4-runtime-branding-openSUSE-12.3-4.2.i586 kdebase4-workspace-4.10.4-14.1.i586 kdebase4-nsplugin-4.10.4-341.1.i586 kdebase4-openSUSE-12.3-4.2.i586 kde4-filesystem-4.10.4-218.1.i586 kdelibs4-core-4.10.4-10.2.i586 mozilla-kde4-integration-0.6.4-2.1.1.i586 kdebase4-runtime-4.10.4-431.1.i586 kdebase4-session-4.10.4-60.1.noarch kdebase4-openSUSE-12.3-4.2.i586 kdebase4-openSUSE-12.3-4.2.i586 kdebase4-openSUSE-lang-12.3-4.2.noarch kdelibs4-4.10.4-10.2.i586 kdebase4-workspace-liboxygenstyle-4.10.4-14.1.i586 kdebase4-workspace-branding-openSUSE-12.3-4.2.i586 xorg-x11-server-7.6_1.13.2-1.5.1.i586 xorg-x11-driver-video-nouveau-1.0.6-2.1.1.i586 # lspci | grep VGA 00:0d.0 VGA compatible controller: NVIDIA Corporation C61 [GeForce 6100 nForce 405] (rev a2) I see that "over 500 bugs" were fixed in the upgrade from .21 to .22 of firefox. I can't see that "obviously" this is one of them. -- How long did the whining go on when KDE2 went on KDE3? The only universal constant is change. If a species can not adapt it goes extinct. That's the law of the universe, adapt or die. -- Billie Walsh, May 18 2013 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2013-06-29 07:36 (GMT-0400) Anton Aylward composed:
Since the ... maybe upgrade yo kde4.10.4 or something else in the regular updates, but before the system upgrade 12.2->12.3 ..
Sometimes .. only sometimes ...
When I click on a URL in Thunderbird to go to that like in Firefox there is a small pause, a freeze while the pointer won't move and then the screen turns to hash. At that point the machine is dead-locked up. Can't hot-key to another screen/activity, can't telnet or ssh into it, it doesn't respond to pings.
So I wonder, is this a problem with -- Firefox
You might be able to tell if build problem with openSUSE's FF or deps by running mozilla.org's static bz2. Same for TB. Also try no-remote option for starting both TB and FF. Since I run multiple builds at once more often than not, I keep it set always in ENV: $ set | grep MOZ MOZ_NO_REMOTE=1
-- KDE4 -- X or the video driver.
It's still possible to force use of nv driver instead of nouveau, at least with G84 and older. RAM failure maybe, run memtest a few passes? -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sat, 29 Jun 2013 07:36:17 -0400 Anton Aylward wrote:
Since the ... maybe upgrade yo kde4.10.4 or something else in the regular updates, but before the system upgrade 12.2->12.3 ..
Sometimes .. only sometimes ...
When I click on a URL in Thunderbird to go to that like in Firefox there is a small pause, a freeze while the pointer won't move and then the screen turns to hash. At that point the machine is dead-locked up. Can't hot-key to another screen/activity, can't telnet or ssh into it, it doesn't respond to pings.
So I wonder, is this a problem with -- Firefox -- KDE4 -- X or the video driver.
MozillaFirefox-21.0-1.18.1.i586 MozillaFirefox-branding-openSUSE-21-2.5.1.i586
libkde4-4.10.4-10.2.i586 kdebase4-openSUSE-12.3-10.41.9.i586 kdebase4-runtime-branding-openSUSE-12.3-4.2.i586 kdebase4-workspace-4.10.4-14.1.i586 kdebase4-nsplugin-4.10.4-341.1.i586 kdebase4-openSUSE-12.3-4.2.i586 kde4-filesystem-4.10.4-218.1.i586 kdelibs4-core-4.10.4-10.2.i586 mozilla-kde4-integration-0.6.4-2.1.1.i586 kdebase4-runtime-4.10.4-431.1.i586 kdebase4-session-4.10.4-60.1.noarch kdebase4-openSUSE-12.3-4.2.i586 kdebase4-openSUSE-12.3-4.2.i586 kdebase4-openSUSE-lang-12.3-4.2.noarch kdelibs4-4.10.4-10.2.i586 kdebase4-workspace-liboxygenstyle-4.10.4-14.1.i586 kdebase4-workspace-branding-openSUSE-12.3-4.2.i586
xorg-x11-server-7.6_1.13.2-1.5.1.i586 xorg-x11-driver-video-nouveau-1.0.6-2.1.1.i586
# lspci | grep VGA 00:0d.0 VGA compatible controller: NVIDIA Corporation C61 [GeForce 6100 nForce 405] (rev a2)
I see that "over 500 bugs" were fixed in the upgrade from .21 to .22 of firefox. I can't see that "obviously" this is one of them.
Hi Anton, I like Felix's answer, too, but in my 27 years experience with this crazy stuff any virtually instantaneous 'dead lock' like you've just described is almost certainly happening at the hardware, firmware, kernel, modules/drivers level. The pointer freeze coincident with the 'hashed' screen, in my mind, almost nails it down to the graphics hardware, the driver module and/or the configuration (a lot less likely.) If possible, I'd roll nouveau back to the last known 'reliably working' (in that system) release or temporarily install the proprietary counterpart, just to see if the problem 'evaporates.' In parallel, I'd be pursuing diagnostic / stress testing of the graphics adapter and it's memory as well as the mainboard and it's memory. Are you monitoring CPU, GPU and other critical system temperatures? Could be an invisible correlation there -- it's summer where I am. Is the system on protected / conditioned power? Any electrical storms recently? :-) Good luck! Carl -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Carl Hartung said the following on 06/29/2013 08:58 AM:
I like Felix's answer, too, but in my 27 years experience with this crazy stuff any virtually instantaneous 'dead lock' like you've just described is almost certainly happening at the hardware, firmware, kernel, modules/drivers level. The pointer freeze coincident with the 'hashed' screen, in my mind, almost nails it down to the graphics hardware, the driver module and/or the configuration (a lot less likely.)
I'm inclined to agree. Applications live in user space and if they crash they crash in user space. To kill the system this thoroughly something has to happen in kernel space, and that would most likely be a driver. I had an 'empty' xorg.conf and the autoconfig listed in the log probing for 8 itens including the frame buffer :-) It saw this was a nvidia and include both drivers.
If possible, I'd roll nouveau back to the last known 'reliably working' (in that system) release or temporarily install the proprietary counterpart, just to see if the problem 'evaporates.' In parallel, I'd be pursuing diagnostic / stress testing of the graphics adapter and it's memory as well as the mainboard and it's memory.
I'm not sure about 'rollback' without rolling back all of xorg. I did try putting 'nv' in 50-device but that errored saying [ 39.635] (EE) NV: The PCI device 0x10de03d1 (GeForce 6100 nForce 405) at 00@00:13:0 has a kernel module claiming it. [ 39.635] (EE) NV: This driver cannot operate until it has been unloaded. I ran lsmod and the nouveau driver was there dealing with various other things. I'm not sure how. Nothing in /etc except this :::::::::::::: modprobe.d/50-nvidia.conf :::::::::::::: options nvidia NVreg_DeviceFileUID=0 NVreg_DeviceFileGID=33 NVreg_DeviceFileMode=0660 So I'm trying "nouveau" in the xorg.conf.d/50-device
Are you monitoring CPU, GPU and other critical system temperatures? Could be an invisible correlation there -- it's summer where I am. Is the system on protected / conditioned power? Any electrical storms recently? :-)
I have other machines here; this is the only one doing this; this is the only one with a nvidia. I've had thermal shutdown on other machines and it says so at the time and appears in the logs. This isn't thermal shutdown. -- How long did the whining go on when KDE2 went on KDE3? The only universal constant is change. If a species can not adapt it goes extinct. That's the law of the universe, adapt or die. -- Billie Walsh, May 18 2013 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On Sat, 29 Jun 2013 10:23:19 -0400 Anton Aylward wrote: <snipped>
I'm not sure about 'rollback' without rolling back all of xorg.
Yeah, that's a hassle just for 'process of elimination' testing.
I did try putting 'nv' in 50-device but that errored saying
[ 39.635] (EE) NV: The PCI device 0x10de03d1 (GeForce 6100 nForce 405) at 00@00:13:0 has a kernel module claiming it. [ 39.635] (EE) NV: This driver cannot operate until it has been unloaded.
I ran lsmod and the nouveau driver was there dealing with various other things. I'm not sure how. Nothing in /etc except this
:::::::::::::: modprobe.d/50-nvidia.conf :::::::::::::: options nvidia NVreg_DeviceFileUID=0 NVreg_DeviceFileGID=33 NVreg_DeviceFileMode=0660
So I'm trying "nouveau" in the xorg.conf.d/50-device
Good luck on this!
I have other machines here; this is the only one doing this; this is the only one with a nvidia.
Point taken.
I've had thermal shutdown on other machines and it says so at the time and appears in the logs. This isn't thermal shutdown.
Of course this was conjecture, but I was thinking more along the lines of marginal, even damaged hardware that might pass POST when switched on and operate properly while cool(er) but not follow built-in protocols above some threshold temperature, hence the reference to diagnostic / stress testing at the hardware level. Again, good luck! Carl -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2013-06-29 10:23 (GMT-0400) Anton Aylward composed:
I had an 'empty' xorg.conf and the autoconfig listed in the log probing for 8 itens including the frame buffer :-) It saw this was a nvidia and include both drivers.
I'm not sure about 'rollback' without rolling back all of xorg.
I did try putting 'nv' in 50-device but that errored saying
[ 39.635] (EE) NV: The PCI device 0x10de03d1 (GeForce 6100 nForce 405) at 00@00:13:0 has a kernel module claiming it. [ 39.635] (EE) NV: This driver cannot operate until it has been unloaded.
Lines containing NOUVEAU(0): mean it's the driver in actual use. If you can't make them go away you won't get NV to work.
I ran lsmod and the nouveau driver was there dealing with various other things. I'm not sure how.
KMS does it. I spent most of more than two days last week trying to define a nouveau bug using a G84. Most of my systems use anything but NVidia video, so my nouveau experience is least of all the X drivers. Avoiding loading nouveau requires at the least putting nomodeset or nouveau.modeset=0 on cmdline, but may also require an initrd built with NO_KMS_IN_INITRD="yes" in /etc/sysconfig/kernel. Whether I tried modprobe -r nouveau at any point I don't remember, but may have been a workaround for needing a rebuilt initrd. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
El 29/06/13 07:36, Anton Aylward escribió: , a freeze while the pointer won't move and then the
screen turns to hash. At that point the machine is dead-locked up. Can't hot-key to another screen/activity, can't telnet or ssh into it, it doesn't respond to pings.
Not a firefox bug, most likely a bug in the kernel. I hate to suggest this, but you might want to try the nvidia propietary driver instead of nouveau which is known to be a work in progress. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
It seems I'm not along with this problem ... http://www.linuxquestions.org/questions/slackware-14/firefox-nouveau-on-fres... As he says, no difference with safe-mode or -P https://bbs.archlinux.org/viewtopic.php?pid=1288218 and https://bugzilla.mozilla.org/show_bug.cgi?id=729817 So how do I tell if I have 2D or 3D? -- How long did the whining go on when KDE2 went on KDE3? The only universal constant is change. If a species can not adapt it goes extinct. That's the law of the universe, adapt or die. -- Billie Walsh, May 18 2013 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2013-07-01 11:20 (GMT-0400) Anton Aylward composed:
It seems I'm not along with this problem ...
http://www.linuxquestions.org/questions/slackware-14/firefox-nouveau-on-fres...
As he says, no difference with safe-mode or -P
So how do I tell if I have 2D or 3D?
I don't know if I know, but I'm pretty sure I know how to ensure that 2D is used. Put the following in a configfile in /etc/X11/ somewhere: Section "Extensions" Option "Composite" "Disable" EndSection Except for an occasional testing session, I use it always, usually in /etc/X11/xorg.conf.d/70-extensions.conf, as bling and I do not get along. Trouble with any Gecko crashing or trashing any of my systems is extraordinarily rare, even though I usually have 5 different ones open at once in the same X session. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
Felix Miata said the following on 07/01/2013 01:26 PM:
On 2013-07-01 11:20 (GMT-0400) Anton Aylward composed:
It seems I'm not along with this problem ...
http://www.linuxquestions.org/questions/slackware-14/firefox-nouveau-on-fres...
As he says, no difference with safe-mode or -P
So how do I tell if I have 2D or 3D?
I don't know if I know, but I'm pretty sure I know how to ensure that 2D is used. Put the following in a configfile in /etc/X11/ somewhere:
Section "Extensions" Option "Composite" "Disable" EndSection
Except for an occasional testing session, I use it always, usually in /etc/X11/xorg.conf.d/70-extensions.conf, as bling and I do not get along. Trouble with any Gecko crashing or trashing any of my systems is extraordinarily rare, even though I usually have 5 different ones open at once in the same X session.
After googling I found a couple of threads for an older version of Fedora and another for ArcOS that advised booting with "nouveau.noaccel=1" on the command line. This seems to work. I ran for most of yesterday without that problem. Now where to put it in the xorg.conf.d/ -- How long did the whining go on when KDE2 went on KDE3? The only universal constant is change. If a species can not adapt it goes extinct. That's the law of the universe, adapt or die. -- Billie Walsh, May 18 2013 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
On 2013-07-02 08:16 (GMT-0400) Anton Aylward composed:
Felix Miata composed:
On 2013-07-01 11:20 (GMT-0400) Anton Aylward composed:
So how do I tell if I have 2D or 3D?
I don't know if I know, but I'm pretty sure I know how to ensure that 2D is used. Put the following in a configfile in /etc/X11/ somewhere:
Section "Extensions" Option "Composite" "Disable" EndSection
Except for an occasional testing session, I use it always, usually in /etc/X11/xorg.conf.d/70-extensions.conf, as bling and I do not get along. Trouble with any Gecko crashing or trashing any of my systems is extraordinarily rare, even though I usually have 5 different ones open at once in the same X session.
After googling I found a couple of threads for an older version of Fedora and another for ArcOS that advised booting with "nouveau.noaccel=1" on the command line.
This seems to work. I ran for most of yesterday without that problem.
Now where to put it in the xorg.conf.d/
Give 50-device.conf a try: Section "Device" ... Option "NoAccel" "true" ... EndSection I've not used it with nouveau, but it did fix i845G 2 years ago. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org
participants (4)
-
Anton Aylward -
Carl Hartung -
Cristian Rodríguez -
Felix Miata