![](https://seccdn.libravatar.org/avatar/77cb4da5f72bc176182dcc33f03a18f3.jpg?s=120&d=mm&r=g)
On 2018-02-05 14:23, Peter Suetterlin wrote:
Carlos E. R. wrote:
zypper lr --details
24 | download.nvidia.com-leap | nVidia Graphics Drivers | Yes | (r ) Yes | Yes | 99 | rpm-md | http://download.nvidia.com/opensuse/leap/42.3
Repo is correct and there is no newer rpm there, says YaST.
Looks fine indeed.
Telcontar:/lib/modules # uname -a Linux Telcontar 4.4.104-39-default #1 SMP Thu Jan 4 08:11:03 UTC 2018 (7db1912) x86_64 x86_64 x86_64 GNU/Linux
Telcontar:/lib/modules # l 4.4.104-39-default/weak-updates/updates/ ls: cannot access '4.4.104-39-default/weak-updates/updates/': No such file or directory
Then it can't load the module indeed :o
I copy the tree from /lib/modules/4.4.104-18.44-default, and now I have:
Telcontar:/lib/modules # l 4.4.104-39-default/weak-updates/updates/ total 8 drwxr-xr-x 2 root root 4096 Feb 4 20:16 ./ drwxr-xr-x 3 root root 4096 Jan 11 01:27 ../ lrwxrwxrwx 1 root root 47 Feb 5 13:03 nvidia.ko -> /lib/modules/4.4.76-1-default/updates/nvidia.ko lrwxrwxrwx 1 root root 51 Feb 5 13:03 nvidia-uvm.ko -> /lib/modules/4.4.76-1-default/updates/nvidia-uvm.ko Telcontar:/lib/modules #
But still:
Telcontar:~ # modprobe nvidia
Missing the depmod.... Or 'insmod /path/to/nvidia.ko'
I tried insmod, it failed:
Telcontar:/lib/modules/4.4.104-18.44-default # insmod /lib/modules/4.4.76-1-default/updates/nvidia.ko insmod: ERROR: could not insert module /lib/modules/4.4.76-1-default/updates/nvidia.ko: Invalid module format
OK, that's the reason the weak-update links are not there: They are only created if the module loads without errors, which it obviously doesn't. So something's wrong there. Best bet would indeed be to recompile (just) the kernel module yourself against the running kernel... Does the nvidia repo have a bug report address?
The reason is that yesterday night I reinstalled everything from the OSS repo: YaST, repo view, all in this list, update unconditionally. That solved the hidden problem; but later I should have reinstalled the nvidia rpms, and I forgot, it was 5 AM. The OSS dup deleted the links and nvidia needed modules, I guess.
Telcontar:/lib/modules/4.4.104-18.44-default #
Telcontar:/lib/modules/4.4.104-18.44-default # rpm -qf /lib/modules/4.4.76-1-default/updates/nvidia.ko nvidia-gfxG03-kmp-default-340.104_k4.4.76_1-11.1.x86_64
Just a stupid question: Does 'zypper se -s nvidia-gfxG03-kmp-default' report it as coming from the nvidia repo, or as System Package?
I can still check, I have a backup of that info, I think. Right after the upgrade, I had: Wed Aug 09 2017 Mon Jul 24 2017 nvidia-gfxG03-kmp-default 340.102_k4.4.27_2-18.1 x86_64 obs://build.suse.de/Proprietary:X11:D rivers (none) == Proprietary:X11:Drivers / openSUSE_Leap_42.2 (none) It is the output from this command: rpm -q -a --queryformat "%{INSTALLTIME}\t%{INSTALLTIME:day} \ %{BUILDTIME:day} %-30{NAME}\t%15{VERSION}-%-7{RELEASE}\t%{arch} \ %25{VENDOR}%25{PACKAGER}\n" | sort | cut --fields="2-" | less -S Notice it was the one from 42.2 - the upgrade failed to update it :-/ Another YaST run removed it, because I find not the string "nvidia" in the corresponding saved query. I can't believe that, so maybe I did a mistake in the query (wrong point in time?) "zypper se -si" did include it: i+ | nvidia-computeG03 | package | 340.104-11.1 | x86_64 | nVidia Graphics Drivers i+ | nvidia-gfxG03-kmp-default | package | 340.104_k4.4.76_1-11.1 | x86_64 | nVidia Graphics Drivers i+ | nvidia-glG03 | package | 340.104-11.1 | x86_64 | nVidia Graphics Drivers i+ | nvidia-uvm-gfxG03-kmp-default | package | 340.104_k4.4.76_1-11.1 | x86_64 | nVidia Graphics Drivers But I forget what the "i+" means :-? -- Cheers / Saludos, Carlos E. R. (from 42.3 x86_64 "Malachite" (Minas Tirith))