my system is running opensuse leap 42.1 with plasma 5.7.3:
Laptop=Dell latitude E6510, RAM=8Gb, GPU=GT218 NVS 3100M, CPU=i7 Q 720 @ 1.60GHz
upgraded 13.2 to leap 42.1, using wolfi repos. I’m running KDE 4.14.18 and plasma 5.7.3, KDE frameworks 5.24.0, KDE applications 16.04.3, Kernel 4.1.27-27-default
After the upgrade to plasma 5.7.3 (and all the other updates included nvidia), nvidia drivers stops to work.
these was the installed nvidia drivers http://paste.opensuse.org/view/download/93111505
so I uninstalled them but yast proposed me to install G04 version, but my graphic board seems to be not supported by that driver so I refused the G04 http://paste.opensuse.org/view/download/18319280
but I had to taboo to have the possibility to refuse the proposed installation.
What I have to do to have back my working nvidia drivers??
Manythanks, ciao, pier
That’s not true.
As mentioned in the other thread too, the kernel module gets compiled for the running kernel when you install the package.
Also, the (official) kernel updates in a released openSUSE version are compatible, so the (old) driver module should still work after such an update.
Symlinks are created automatically when a kernel update is installed, so that the modules are found on every installed kernel version…
@pier:
Yes, it seems your card is not supported by G04.
If YaST wants to automatically install it anyway, the G04 package maybe incorrectly specifies that it does support your card (or maybe it really does? Then the list at nvidia’s homepage would be wrong… )
OTOH, it may also be caused by package dependencies. Nothing should depend on the nvidia driver itself, but you may have 32bit applications installed that require libGL1, so the package management tries to fulfill that dependency (and chooses nvidia-glG04 instead of Mesa-libGL1-32bit).
If the latter is the case, you should get out of this loop by manually selecting Mesa-GL1-32bit (and the other Mesa-libXXX-32bit packages), or remove/disable the nvidia repo before uninstalling the driver packages.
What never worked for you?
The nvidia driver packages in general, or that the kmp was compiled on installation?
IIRC that got changed 2 years ago, for all versions of the driver. Before that the mvidia kmp packages did contain pre-compiled kernel modules.
But note that the module is only compiled when you install the nvidia kmp package, not when you install a kernel update.
So installing an “incompatible” kernel will break the driver and makes a reinstallation necessary, of the kmp package at least.
Btw, the current G04 packages even contain a patch for kernel 4.7 already, the standard nvidia driver doesn’t support that yet…
The official nvidia packages are maintained in OBS btw, if you want to take a look yourself.
Hi
The nvidia driver packages and kmp, I’ve only decided to add the GT8800 since Leap came out, since I’ve always done it with the run file…just carried on…
Patches can be added to the run file via --apply-patch option then it creates a custom run file to install, so it’s still usable. Haven’t got around to tumbleweed on this machine (probably never will…).
Thanks for all the info All the rest of my systems have intel or AMD running the radeon driver.
I used them successfully with a GeForce4 years ago.
But of course that old driver (96.43.xx) doesn’t work since a while any more since quite a while because it is not maintained/updated by nvidia any more since 2011… It only can be used with Xorg 1.10 or older.
I replaced that GeForce with a Radeon in 2009.
The nvidia packages should work with older cards as well though, provided they are supported by the corresponding driver version of course.
(G02 goes back to GeForce 6xxx).
In any case, there’s really no difference between the packaged driver and the .run installer regarding support for cards, the driver is a closed-source binary blob anyway (that’s the reason why the packages cannot be built or published on OBS, the projects do not contain the driver itself, only the packaging files).
Thanks for all the info All the rest of my systems have intel or AMD running the radeon driver.
pla@suseST-pla:~> su
Password:
suseST-pla:/home/pla # mkinitrd
Creating initrd: /boot/initrd-4.1.27-24-default
Executing: /usr/bin/dracut --logfile /var/log/YaST2/mkinitrd.log --force /boot/initrd-4.1.27-24-default 4.1.27-24-default
*** Including module: bash ***
*** Including module: warpclock ***
*** Including module: i18n ***
*** Including module: ifcfg ***
*** Including module: drm ***
*** Including module: plymouth ***
*** Including module: kernel-modules ***
Omitting driver i2o_scsi
*** Including module: resume ***
*** Including module: rootfs-block ***
*** Including module: terminfo ***
*** Including module: udev-rules ***
Skipping udev rule: 91-permissions.rules
Skipping udev rule: 80-drivers-modprobe.rules
*** Including module: biosdevname ***
*** Including module: haveged ***
*** Including module: systemd ***
*** Including module: usrmount ***
*** Including module: base ***
*** Including module: fs-lib ***
*** Including module: shutdown ***
*** Including module: suse ***
*** Including modules done ***
*** Installing kernel module dependencies and firmware ***
*** Installing kernel module dependencies and firmware done ***
*** Resolving executable dependencies ***
*** Resolving executable dependencies done***
*** Hardlinking files ***
*** Hardlinking files done ***
*** Stripping files ***
*** Stripping files done ***
*** Generating early-microcode cpio image ***
*** Constructing GenuineIntel.bin ****
*** Store current command line parameters ***
Stored kernel commandline:
resume=UUID=3f1cbf5f-82cd-4714-b0cc-24c48a775be2
root=UUID=167d3751-360f-4d2e-b342-27f62b8b16b7 rootflags=rw,relatime,data=ordered rootfstype=ext3
*** Creating image file ***
*** Creating image file done ***
Some kernel modules could not be included
This is not necessarily an error:
swap
ext3
pciehp
Creating initrd: /boot/initrd-4.1.27-27-default
Executing: /usr/bin/dracut --logfile /var/log/YaST2/mkinitrd.log --force /boot/initrd-4.1.27-27-default 4.1.27-27-default
*** Including module: bash ***
*** Including module: warpclock ***
*** Including module: i18n ***
*** Including module: ifcfg ***
*** Including module: drm ***
*** Including module: plymouth ***
*** Including module: kernel-modules ***
Omitting driver i2o_scsi
*** Including module: resume ***
*** Including module: rootfs-block ***
*** Including module: terminfo ***
*** Including module: udev-rules ***
Skipping udev rule: 91-permissions.rules
Skipping udev rule: 80-drivers-modprobe.rules
*** Including module: biosdevname ***
*** Including module: haveged ***
*** Including module: systemd ***
*** Including module: usrmount ***
*** Including module: base ***
*** Including module: fs-lib ***
*** Including module: shutdown ***
*** Including module: suse ***
*** Including modules done ***
*** Installing kernel module dependencies and firmware ***
*** Installing kernel module dependencies and firmware done ***
*** Resolving executable dependencies ***
*** Resolving executable dependencies done***
*** Hardlinking files ***
*** Hardlinking files done ***
*** Stripping files ***
*** Stripping files done ***
*** Generating early-microcode cpio image ***
*** Constructing GenuineIntel.bin ****
*** Store current command line parameters ***
Stored kernel commandline:
resume=UUID=3f1cbf5f-82cd-4714-b0cc-24c48a775be2
root=UUID=167d3751-360f-4d2e-b342-27f62b8b16b7 rootflags=rw,relatime,data=ordered rootfstype=ext3
*** Creating image file ***
*** Creating image file done ***
Some kernel modules could not be included
This is not necessarily an error:
swap
ext3
pciehp
Update bootloader...
suseST-pla:/home/pla #
and then reinstalled nvidia drivers G03 but it doesn’t worked, I uninstalled nvidia, but in this case yast didn’t suggest me to install G04 version, a lot of warning during uninstall, I don’t know where find the log, but now it works with nouveau.
manythanks, ciao, pier
P.S. to uninstall nvidia I had to boot in recovery mode, so I noted that plasma5 cannot works booting leap in recovery mode, I had to use KDE4, so, KDE4 remain very useful
You need to run mkinitrd after you reinstalled the nvidia drivers.
The point is to make sure the nvidia kernel module is in the initrd.
I uninstalled nvidia, but in this case yast didn’t suggest me to install G04 version
Probably because you now have the Mesa 32bit libraries installed.
a lot of warning during uninstall, I don’t know where find the log, but now it works with nouveau.
When being uninstalled, the packages run mkinitrd again, so the “warnings” are probably related to that and “normal”.
P.S. to uninstall nvidia I had to boot in recovery mode, so I noted that plasma5 cannot works booting leap in recovery mode, I had to use KDE4, so, KDE4 remain very useful
No, Plasma5 will not work in recovery mode when nvidia is installed, because nvidia breaks Mesa’s OpenGL support including the software renderer.
KDE4 can work without OpenGL.
Actually it is Qt5’s QML that requires OpenGL (not Plasma5). It is possible to switch back to a Qt internal software renderer but that doesn’t happen automatically (yet, maybe this will be improved in the future), you’d need to set an environment variable.
But IceWM should be installed by default anyway, and be even more robust to graphics driver problems than even KDE4…
There’s no point having in KDE4 just to be able to uninstall the nvidia driver in case of problems.
I reinstalled nvidia drivers and then ran mkinitrd but doesn’t solved, I had to came back to nouveau.
IceWM mm, I tried it, thanks, it works fine, a bit rudimantal compared with kde4 but enough good to uninstall nvidia
So, when nvidia drivers are not installed does plasma5 works in recovery mode??
PS: Do you maybe have virtualbox-guest-kmp-default installed?
There has been a problem with this package in Leap that broke the installation of nvidia’s kernel module…
If you have it, uninstall it and reinstalling the nvidia driver should work. You only need it in a guest anyway, not on the host.
I’m not sure whether this has been fixed already, or whether it maybe is again a problem after the recent kernel update.
hoping before september 