[arch-general] Totem segfaulting on start on a fully up-to-date archlinux with testing enabled.
Hello. Got this error when I try to start totem today : [fred@fredo-arch ~]$ totem & [1] 7902 [fred@fredo-arch ~]$ (totem:7902): Grilo-WARNING **: [plugin-registry] grl-plugin-registry.c:623: Failed to open module: '/usr/lib/grilo-0.1/libgrlupnp.so' [1]+ Segmentation fault totem I verified, and of course, I have "/usr/lib/grilo-0.1/libgrlupnp.so" installed. Any idea or tip ? Thanks ! -- Frederic Bezies fredbezies@gmail.com
Le samedi 16 juin 2012 à 09:57 +0200, fredbezies a écrit :
Hello.
Got this error when I try to start totem today :
[fred@fredo-arch ~]$ totem & [1] 7902 [fred@fredo-arch ~]$ (totem:7902): Grilo-WARNING **: [plugin-registry] grl-plugin-registry.c:623: Failed to open module: '/usr/lib/grilo-0.1/libgrlupnp.so'
[1]+ Segmentation fault totem
I verified, and of course, I have "/usr/lib/grilo-0.1/libgrlupnp.so" installed. Any idea or tip ? Thanks !
I got that error from time to time when I run pacman -Syu and it updates the nvidia driver. So my best bet would be that you just reboot to see if it changes anything or not.
2012/6/16 solsTiCe d'Hiver
Le samedi 16 juin 2012 à 09:57 +0200, fredbezies a écrit :
Hello.
Got this error when I try to start totem today :
[fred@fredo-arch ~]$ totem & [1] 7902 [fred@fredo-arch ~]$ (totem:7902): Grilo-WARNING **: [plugin-registry] grl-plugin-registry.c:623: Failed to open module: '/usr/lib/grilo-0.1/libgrlupnp.so'
[1]+ Segmentation fault totem
I verified, and of course, I have "/usr/lib/grilo-0.1/libgrlupnp.so" installed. Any idea or tip ? Thanks !
I got that error from time to time when I run pacman -Syu and it updates the nvidia driver. So my best bet would be that you just reboot to see if it changes anything or not.
Guessed it right; Too bad it is sometimes too "windows-like" with nvidia driver updates :| -- Frederic Bezies fredbezies@gmail.com
Hello, ** fredbezies [2012-06-16 11:26:13 +0200]:
2012/6/16 solsTiCe d'Hiver
Le samedi 16 juin 2012 à 09:57 +0200, fredbezies a écrit :
Hello.
Got this error when I try to start totem today :
[fred@fredo-arch ~]$ totem & [1] 7902 [fred@fredo-arch ~]$ (totem:7902): Grilo-WARNING **: [plugin-registry] grl-plugin-registry.c:623: Failed to open module: '/usr/lib/grilo-0.1/libgrlupnp.so'
[1]+ Segmentation fault totem
I verified, and of course, I have "/usr/lib/grilo-0.1/libgrlupnp.so" installed. Any idea or tip ? Thanks !
I got that error from time to time when I run pacman -Syu and it updates the nvidia driver. So my best bet would be that you just reboot to see if it changes anything or not.
Guessed it right; Too bad it is sometimes too "windows-like" with nvidia driver updates :|
Too bad that whose who don't known the basics try to make such deep conclusion. The last time I really need to reboot after graphics driver update was when I updated kernel too (actually I updated kernel several times without rebooting before I updated the NVIDA driver and reboot). IMHO, the minimal steps to use new NVIDIA driver are 1. shutdown X system; 2. logout as usual user and login as root; 3. unload nvidia (kernel) module; 4. load new nvidia (kernel) module; 5. start X system. --- WBR, Vladimir Lomov -- Well, the handwriting is on the floor. -- Joe E. Lewis
2012/6/16 Vladimir Lomov
Hello, ** fredbezies [2012-06-16 11:26:13 +0200]:
2012/6/16 solsTiCe d'Hiver
Le samedi 16 juin 2012 à 09:57 +0200, fredbezies a écrit :
Hello.
Got this error when I try to start totem today :
[fred@fredo-arch ~]$ totem & [1] 7902 [fred@fredo-arch ~]$ (totem:7902): Grilo-WARNING **: [plugin-registry] grl-plugin-registry.c:623: Failed to open module: '/usr/lib/grilo-0.1/libgrlupnp.so'
[1]+ Segmentation fault totem
I verified, and of course, I have "/usr/lib/grilo-0.1/libgrlupnp.so" installed. Any idea or tip ? Thanks !
I got that error from time to time when I run pacman -Syu and it updates the nvidia driver. So my best bet would be that you just reboot to see if it changes anything or not.
Guessed it right; Too bad it is sometimes too "windows-like" with nvidia driver updates :|
Too bad that whose who don't known the basics try to make such deep conclusion.
Don't know the basics ? Just using linux since christmas 2004...
The last time I really need to reboot after graphics driver update was when I updated kernel too (actually I updated kernel several times without rebooting before I updated the NVIDA driver and reboot).
Rebooting is sometimes the cleanest and the better way to be sure modules are correctly loaded.
IMHO, the minimal steps to use new NVIDIA driver are 1. shutdown X system; 2. logout as usual user and login as root; 3. unload nvidia (kernel) module; 4. load new nvidia (kernel) module; 5. start X system.
Well, it could be a solution, but a reboot is quicker and less hard to set up.
--- WBR, Vladimir Lomov
-- Well, the handwriting is on the floor. -- Joe E. Lewis
-- Frederic Bezies fredbezies@gmail.com
** fredbezies [2012-06-16 13:02:44 +0200]:
2012/6/16 Vladimir Lomov
Hello, ** fredbezies [2012-06-16 11:26:13 +0200]:
2012/6/16 solsTiCe d'Hiver
Le samedi 16 juin 2012 à 09:57 +0200, fredbezies a écrit :
Hello.
Got this error when I try to start totem today :
[fred@fredo-arch ~]$ totem & [1] 7902 [fred@fredo-arch ~]$ (totem:7902): Grilo-WARNING **: [plugin-registry] grl-plugin-registry.c:623: Failed to open module: '/usr/lib/grilo-0.1/libgrlupnp.so'
[1]+ Segmentation fault totem
I verified, and of course, I have "/usr/lib/grilo-0.1/libgrlupnp.so" installed. Any idea or tip ? Thanks !
I got that error from time to time when I run pacman -Syu and it updates the nvidia driver. So my best bet would be that you just reboot to see if it changes anything or not.
Guessed it right; Too bad it is sometimes too "windows-like" with nvidia driver updates :|
Too bad that whose who don't known the basics try to make such deep conclusion.
Don't know the basics ? Just using linux since christmas 2004...
I'm sorry, if you think I insult you. But I wouldn't mess the basics knownledge with experience. You have to experiment and try to know something new, always.
The last time I really need to reboot after graphics driver update was when I updated kernel too (actually I updated kernel several times without rebooting before I updated the NVIDA driver and reboot).
Rebooting is sometimes the cleanest and the better way to be sure modules are correctly loaded.
If restrict the subject only to kernel and its modules then yes, it is, IMHO, the only sane way to deal with modules.
IMHO, the minimal steps to use new NVIDIA driver are 1. shutdown X system; 2. logout as usual user and login as root; 3. unload nvidia (kernel) module; 4. load new nvidia (kernel) module; 5. start X system.
Well, it could be a solution, but a reboot is quicker and less hard to set up.
See my other message in the thread. P.S. I don't like to reboot my desktop without serious reasone (it runs like a server and, it is a secret, I'm using rtorrent on it :). --- WBR, Vladimir Lomov -- Cleaning your house while your kids are still growing is like shoveling the walk before it stops snowing. -- Phyllis Diller There is no need to do any housework at all. After the first four years the dirt doesn't get any worse. -- Quentin Crisp
2012/6/16 Vladimir Lomov
[...]
Don't know the basics ? Just using linux since christmas 2004...
I'm sorry, if you think I insult you. But I wouldn't mess the basics
No problems at all.
knownledge with experience. You have to experiment and try to know something new, always.
I do agree. But I think it is cleaner to reboot your linux if some modules are modified, and of course if kernel is upgraded.
The last time I really need to reboot after graphics driver update was when I updated kernel too (actually I updated kernel several times without rebooting before I updated the NVIDA driver and reboot).
Rebooting is sometimes the cleanest and the better way to be sure modules are correctly loaded.
If restrict the subject only to kernel and its modules then yes, it is, IMHO, the only sane way to deal with modules.
Even if I am a "long time" linux user, I try to reboot only when it is needed. But for nvidia (or radeon) drivers and udev changes, reboot is the simplest and cleanest way to act ;)
IMHO, the minimal steps to use new NVIDIA driver are 1. shutdown X system; 2. logout as usual user and login as root; 3. unload nvidia (kernel) module; 4. load new nvidia (kernel) module; 5. start X system.
Well, it could be a solution, but a reboot is quicker and less hard to set up.
See my other message in the thread.
P.S. I don't like to reboot my desktop without serious reasone (it runs like a server and, it is a secret, I'm using rtorrent on it :).
I understand. I only have a "normal" desktop computer. And rtorrent ? Well :D
--- WBR, Vladimir Lomov
-- Cleaning your house while your kids are still growing is like shoveling the walk before it stops snowing. -- Phyllis Diller
There is no need to do any housework at all. After the first four years the dirt doesn't get any worse. -- Quentin Crisp
-- Frederic Bezies fredbezies@gmail.com
Le samedi 16 juin 2012 à 19:36 +0900, Vladimir Lomov a écrit :
IMHO, the minimal steps to use new NVIDIA driver are 1. shutdown X system; 2. logout as usual user and login as root; 3. unload nvidia (kernel) module; 4. load new nvidia (kernel) module; 5. start X system.
Are you sure you don't need to run some udev related command like udevadm ? Because last time I tried something like you said it didn't work
Hello, ** solsTiCe d'Hiver [2012-06-16 13:24:34 +0200]:
Le samedi 16 juin 2012 à 19:36 +0900, Vladimir Lomov a écrit :
IMHO, the minimal steps to use new NVIDIA driver are 1. shutdown X system; 2. logout as usual user and login as root; 3. unload nvidia (kernel) module; 4. load new nvidia (kernel) module; 5. start X system.
Are you sure you don't need to run some udev related command like udevadm ?
Because last time I tried something like you said it didn't work
Don't know you environment and hardware, but I just tried: on my desktop: Archlinux x86_64 (AMD), kernel 3.4.2-2, systemd; last update brought nvidia 259.59 when I was running openbox. I installed update, logout (exited Openbox) which should be bring me to KDM but I got a terminal. I logged into terminal as a user, run $ sudo systemctl stop kdm.service $ sudo modprobe -r nvidia $ sudo modprobe nvidia $ sudo systemctl kdm.service so now all works fine.
From /var/log/Xorg.0.log: $ grep 'NVIDIA GLX' /var/log/Xorg.0.log [ 47278.388] (II) NVIDIA GLX Module 295.59 Wed Jun 6 21:40:44 PDT 2012
--- WBR, Vladimir Lomov -- Cinemuck, n.: The combination of popcorn, soda, and melted chocolate which covers the floors of movie theaters. -- Rich Hall, "Sniglets"
Le samedi 16 juin 2012 à 22:50 +0900, Vladimir Lomov a écrit :
on my desktop: Archlinux x86_64 (AMD), kernel 3.4.2-2, systemd; last update brought nvidia 259.59 when I was running openbox. I installed update, logout (exited Openbox) which should be bring me to KDM but I got a terminal. I logged into terminal as a user, run $ sudo systemctl stop kdm.service $ sudo modprobe -r nvidia $ sudo modprobe nvidia $ sudo systemctl kdm.service so now all works fine.
From /var/log/Xorg.0.log: $ grep 'NVIDIA GLX' /var/log/Xorg.0.log [ 47278.388] (II) NVIDIA GLX Module 295.59 Wed Jun 6 21:40:44 PDT 2012
Well, I should have been more explicit. I logout from gnome then alt+F1, login as root (yeah!) $ rc.d stop gdm $ modprobe -r nvidia this one failed. I don't why. I don't use systemd and nvidia not anymore so this is irrelevant, at least to me now.
On Sat, 2012-06-16 at 18:33 +0200, solsTiCe d'Hiver wrote:
I don't use systemd and nvidia not anymore
You still use an NVIDIA graphics? If so, which graphics do you use and do you use the nv or nouveau driver? I ask because I'm uncertain what to do, for keeping my system usable for my needs. I'm only interested in experiences with gfxcards and drivers, I can't contribute something useful. Regards, Ralf
Le samedi 16 juin 2012 à 19:03 +0200, Ralf Mardorf a écrit :
On Sat, 2012-06-16 at 18:33 +0200, solsTiCe d'Hiver wrote:
I don't use systemd and nvidia not anymore
You still use an NVIDIA graphics? If so, which graphics do you use and do you use the nv or nouveau driver?
I have a geforce 8300 GS. Nvidia driver since 295.xx began giving random 30s freeze when software attempt to use openGL one way or the other, I guess; like launching firefox or lauching gnome. I don't know what is the problem/bug. I have almost no hope that this will ever be fixed and I can't downgrade because X was updated. But nouveau drivers are perfectly fine except I can't play some game anymore. But I never played that much.
On Sat, 2012-06-16 at 19:50 +0200, solsTiCe d'Hiver wrote:
I have a geforce 8300 GS. Nvidia driver since 295.xx began giving random 30s freeze when software attempt to use openGL one way or the other, I guess; like launching firefox or lauching gnome. I don't know what is the problem/bug.
Firefox 13? My Ubuntu and Arch do use similar versions of software, most of the times when something evil happens, is when using Firefox. It's equal to the trouble caused by hardware that will brake soon, but if it would be related to damaged hardware, it also should happen for completely outdated Linux installs on my machine, but those installs seem to be still stable. Arch nvidia is 295.53 and Ubuntu nvidia is 295.40.
I have almost no hope that this will ever be fixed and I can't downgrade because X was updated.
But nouveau drivers are perfectly fine except I can't play some game anymore. But I never played that much.
I suspect that nouveau still isn't an option for me. Ubuntu doesn't support the nv driver anymore and on Arch I'm only using it for the kernel-rt, for the regular kernel I'm using the proprietary driver, since I sometimes wish to have 3D acceleration. I don't know if Arch is stable with the kernel-rt and the nv driver, because I couldn't start X on Arch when running kernel-rt's < 3.4 for the last month. FWIW, the "regular" Arch kernel is still 3.3.x. I'll test current nouveau on Arch and current Ubuntu ASAP, however I already know that for my Debian stable, nouveau still is a PITA. If it should be related to the nvidia driver version, I don't need packages, if X should work with 17x or whatever, I'm willing to install it manually. Thank you for the information, Ralf
I just tried the nvidia 173.xx version (nvidia-173xx from AUR). It's working fine (almost). I never tried it, thinking it was for old hardware. But it's not. I guess, just an old branch/version with no support for the latest/newest shiny hardware/cg
On Sat, 2012-06-16 at 20:50 +0200, solsTiCe d'Hiver wrote:
I just tried the nvidia 173.xx version (nvidia-173xx from AUR). It's working fine (almost).
I never tried it, thinking it was for old hardware. But it's not. I guess, just an old branch/version with no support for the latest/newest shiny hardware/cg
It wasn't ironically. 173, I wasn't sure if it was ..3 (still had 3 in mind), was my preferred version. I was and I'm serious! I didn't run Arch, but Suse, Debian and Ubuntu and a "Debian and Ubuntu derivative", 64Studio with 17x, while there already was 2x available, IIRC. I'm not hunting for latest software versions, I'm hunting for softwares that in combination with other software and my available hardware, fit to my needs and I freeze a stable Linux that fit to my needs, as long, as possible. Hardware could irreparable brake, happened to me, or a bonanza allows me to switch from a semi-pro to a pro-audio card, happened to me ... thus an upgrade might be needed, happened to me. ;) I prefer a rolling release to other releases, but I'm willing to freeze a rolling release on my machine, as soon as I'm satisfied, as long as possible. Apologize, broken English, but I'm sure you understand ;). Regards, Ralf
Hello, ** solsTiCe d'Hiver [2012-06-16 18:33:23 +0200]:
Le samedi 16 juin 2012 à 22:50 +0900, Vladimir Lomov a écrit :
on my desktop: Archlinux x86_64 (AMD), kernel 3.4.2-2, systemd; last update brought nvidia 259.59 when I was running openbox. I installed update, logout (exited Openbox) which should be bring me to KDM but I got a terminal. I logged into terminal as a user, run $ sudo systemctl stop kdm.service $ sudo modprobe -r nvidia $ sudo modprobe nvidia $ sudo systemctl kdm.service so now all works fine.
From /var/log/Xorg.0.log: $ grep 'NVIDIA GLX' /var/log/Xorg.0.log [ 47278.388] (II) NVIDIA GLX Module 295.59 Wed Jun 6 21:40:44 PDT 2012
Well, I should have been more explicit.
I logout from gnome then alt+F1, login as root (yeah!) $ rc.d stop gdm $ modprobe -r nvidia this one failed. I don't why. I don't use systemd and nvidia not anymore so this is irrelevant, at least to me now. And I had to tell the whole story too, before `sudo modprobe -r nvidia' I did $ lsmod | grep nvidia to see what modules may use nvidia. If there are such modules I have to unload them first and nvidia after.
Of course all this will work if modules can be unloaded, sometimes `modprobe' "don't" work, then you have to reboot. --- WBR, Vladimir Lomov -- FORCE YOURSELF TO RELAX!
participants (4)
-
fredbezies
-
Ralf Mardorf
-
solsTiCe d'Hiver
-
Vladimir Lomov