After a 42.2 upgrade from 42.1 that went smoothly the startup fails with the error on a text screen…
sddm[1166]:segfault at 0 ip(long number) sp(long number) erro 4 in libQt5core.so.6.1
…and after a reboot…
sddm[1210] segfault at 780 ip(long number) sp(long number) error 14 in sddm
I didn’t think the ip and sp numbers were significant to a solution. If needed I will copy them. I can still open text consoles and log in but startx fails.
Thanks in advance
what hardware and which repo’s?
GeForce 6100 nForce 430
# | Alias | Name | Enabled | GPG Check | Refresh--+-----------------------------------+-----------------------------------+---------+-----------+--------
1 | Jacobs | PACKMAN-Jacobs | No | ---- | No
2 | download.nvidia.com-leap | nVidia Graphics Drivers | No | ---- | No
3 | http-download.nvidia.com-511e98f1 | nVidia Graphics Drivers | No | ---- | Yes
4 | http-opensuse-guide.org-6b0c11f1 | libdvdcss repository | No | ---- | No
5 | http-packman.inode.at-6dad5200 | Packman Repository | No | ---- | Yes
6 | repo-non-oss | openSUSE-Leap-42.2-Non-Oss | Yes | (r ) Yes | Yes
7 | repo-oss | openSUSE-Leap-42.2-Oss | Yes | (r ) Yes | Yes
8 | repo-update | openSUSE-Leap-42.2-Update | Yes | (r ) Yes | Yes
9 | repo-update-non-oss | openSUSE-Leap-42.2-Update-Non-Oss | Yes | (r ) Yes | Yes
And here is some of the Xorg.log, and I wonder why it’s trying to load an Nvidia module. Doesn’t the update force a switch to nouveau? Or do I need to rerun the Nvidia *run file?
364.115] (==) AIGLX enabled 364.115] (II) LoadModule: "nvidia"
364.116] (WW) Warning, couldn't open module nvidia
364.116] (II) UnloadModule: "nvidia"
364.116] (II) Unloading nvidia
364.116] (EE) Failed to load module "nvidia" (module does not exist, 0)
364.116] (EE) No drivers available.
364.116] (EE)
Fatal server error:
364.116] (EE) no screens found(EE)
364.116] (EE)
Please consult the The X.Org Foundation support
at http://wiki.x.org
for help.
364.116] (EE) Please also check the log file at "/home/ion/.local/share/xorg/Xorg.0.log" for additional information.
364.116] (EE)
you have the nvidia repo but it’s disabled did you install the driver and then disable the repo?
that’s a bad idea because after every kernel update we get a driver update re-enable the nvidia repo and try to do an update (you can do it in text mode)
zypper mr -e -rk 2
zypper ref
zypper up
also post your repo’s with the -d switch so we can see the url’s as you seam to have 2 nvidia repos and reoo #2 has autorefresh disabled this is a bad idea as you won’t get updates
zypper lr -d
did you install nvidia the hard way?
in that case you need to reinstall it after every kernel update
if you are nor sure if you have installed any nvidia packages try running this
zypper se -si nvidia
that searches for installed only packages with nvidia in their name
I was working with instructions in How to upgrade from openSUSE Leap 42.1 to openSUSE Leap 42.2 | 2DayGeek as they were the clearest I could find. It instructed me to disable all repos except the four for 42.2 which I did. I assumed that the upgrade would default to the nouveau driver so it made sense to disable the Nvidia repos.
And yes I installed the Nvidia driver the hard way as it has to be an earlier version than current. And yes I disabled the repo as updating to a newer version does not work.
Running zypper se -si nvidia produced…
Loading repository data...
Reading installed packages...
No matching items found.
Are you suggesting I should just re-run the Nvidia *.run file again. But if, as the above zypper reports there is no Nvidia item installed what is the driver it is trying to use? Not nouveau?
Sorry it’s late here and I have to keep rebooting to a different drive to connect with the forum. It’s time consuming. I will answer the other questions tomorrow.
you need to reinstall the driver the hard way
just re-run the run file
I haven’t seen that site but the repo way is much simpler to maintain, we get updates automatically whenever the kernel is updated the nvidia driver is updated, one of the reasons I stay with LEAP is the fact that TW does not have an nvidia repo and I don’t want to mess with the run file
Yes but the updated Nvidia drivers do NOT work with my card. Not with 13.2 not with 42.1. That’s why I do it the hard way.I’ll run the *run file tomorrow and report.
it’s the same driver
what card do you have as the issues wore not the driver but X
my old 5200 worked on 13.1 but not 13.2 the reason wasn’t the driver or kernel but the X server was updated
if your card is that old I’d suggest you install openbox and openbox-kde then use plasma5 with openbox not with kwin5
it will be faster and you can use noveaou and won’t mess with the propitiatory driver
http://software.opensuse.org/package/openbox
http://software.opensuse.org/package/openbox-kde
there is a bug with http://software.opensuse.org/ it doesn’t show the packages from the 42.2 oss repo someone should report it at bugzilla I’m too lazy to do it
my point was even if the package is not shown for 42.2 it’s there and you can install it with yast or zypper
there is a bug with http://software.opensuse.org/ it doesn’t show the packages from the 42.2 oss repo someone should report it at bugzilla I’m too lazy to do it
[QUOTE=I_A;2802770]you need to reinstall the driver the hard way
just re-run the run file[/QUOTE$]
I blacklisted nouveau, ran mkinitrd and rebooted. I executed the *run file and it errored out with “Unable to load the kernel module 'nvidia.ko” with many possibilities for the cause. I’m really sorry I started this upgrade as I now have a useless box. Well, text only.
Which Nvidia driver?
304.132?
Try the 304.131:
http://us.download.nvidia.com/XFree86/Linux-x86_64/304.131/NVIDIA-Linux-x86_64-304.131.run
Are you saying that the current version in the SDB additional package repository 304.132-9 is the same as 304.131 which is the one that works with my GeForce 6100 nForce 430? Because it didn’t work when I first installed 42.1. Which is why I did the Nvidia hard way install.
if your card is that old I’d suggest you install openbox and openbox-kde then use plasma5 with openbox not with kwin5
it will be faster and you can use noveaou and won’t mess with the propitiatory driver
openSUSE Software
openSUSE Software
I’m not keen on switching to yet another method I am not familiar with. If that’s the only solution I think I’ll move over to another distro that I have tested. It would be a shame since I’ve been with this one since 9.1.
That’s the one I just ran (304.131) and it failed.
I had a similar situation recently, with similar messages as far as I can remember.
I had the nvidia driver installed the hard way, which I do usually. When the normal login became impossible, I thought that I have missed a kernel update, and, hence, had to reinstall the nvidia driver.
But I checked the version; the kernel was not updated.
I reinstalled the nvidia driver anyway, and it solved the problem.
Try a few versions of the nvidia driver, attempying ones that are 0.5 - 1 year apart from each other, going back five years if you can. Don’t forget to install packages kernel-source, kernel-syms, kernel-devel, though some of these may be excessive. After installation, issuing a
mkinitrd
may or may not be required, but do it anyway.
I made sure that kernel-source, kernel-syms and kernel-devel and re-ran the 304.131 run file. It’s the one I know works with my card. At least in 13.2 and 42.1. It errored out. Then I blacklisted the nvidia driver removed the nouveau blacklist and ran mkinitrd, and rebooted. My Xorg.0.log file reports it’s still trying to load the nvidia module which it can’t find. I’ve already wasted many hours on an update that has caused nothing but trouble. Fortunately all my data is backed up. If I don’t reach a solution soon I am going to re-install 42.1 and all the associated software I use. Here goes another day’s work.
Thanks to all for the help.
I am guessing that the upgrade from 42.1 and several attempts of installing the nvidia driver have left barring traces, like links to opengl files.
I suggest perform a fresh install, perhaps on another partition or hard drive.
Then install the nvidia driver the hard way, beginning with the latest version.
if your card is not supported by the rpm driver it will not be supported by the run driver
I told you to use noveau (not to bklacklist it) with openbox if you want a stable plasma5
what card?
as far as I know the Geforce 5 was dropped with 13.2 when X was updated (I have an old celeron with 5200 running 13.1) and the old 6-9 series are still supported with the G02 driver, newer cards are supported with the G03 and G04 driver
if you have an old card do not use the nvidia driver but noveau, as noveau does not work well with kwin5 for plasma 5 use you need openbox-kde
if you have a supported card use the rpm driver from the nvidia repo and do not manually blacklist noveau that’s done automatically when needed
I took your advice. I installed 42.2 from a DVD (not an upgrade from 42.1 like the last attempt) to be sure everything was cleaned out. I installed openbox and openbox-kde and rebooted. At the login I selected openbox-kde and I got kde running on a 640 x 480 screen that is not changeable in the Desktop Settings. How do I now change the display resolution?
I changed the resolution in grub2 via YaST and it established it as the default so when KDE started it was at the correct resolution. Thanks to all for their help.
afaik any pc that can run LEAP ie a 64bit machine has a supported nvidia graphic card, the Geforce 5 series that was dropped in 13.2 was the last agp card nvidia made, I am guessing but I’m pretty sure your pc has a supported graphic card and you could use nvidia’s drivers if nothing else for accelerated video decoding
what’s the output of
lspci | grep VGA
or
hwinfo --gfxcard