Description of problem: dkms failed to build the nvidia340 driver for kernel-desktop-4.7.0-0.rc7.1.mga6-1-1.mga6 This occurred after a system update that included the mentioned kernel and a reboot. A subsequent manually started dkms build was successful and after a update-alternatives, ldconfig and reboot the system is now using the nvidia340 driver without issues. Please see commands and outputs below for details. Version-Release number of selected component (if applicable): # rpm -qa | grep -Ei 'kernel|nvidia|dkms' | sort dkms-2.0.19-37.mga6 dkms-minimal-2.0.19-37.mga6 dkms-nvidia340-340.96-37.mga6.nonfree kernel-desktop-4.6.3-1.mga6-1-1.mga6 kernel-desktop-4.7.0-0.rc7.1.mga6-1-1.mga6 kernel-desktop-devel-4.6.3-1.mga6-1-1.mga6 kernel-desktop-devel-4.7.0-0.rc7.1.mga6-1-1.mga6 kernel-desktop-devel-latest-4.7.0-0.rc7.1.mga6 kernel-desktop-latest-4.7.0-0.rc7.1.mga6 kernel-firmware-20160608-1.mga6 kernel-userspace-headers-4.7.0-0.rc7.1.mga6 vboxadditions-kernel-4.6.3-desktop-1.mga6-5.0.22-5.mga6 vboxadditions-kernel-4.7.0-desktop-0.rc7.1.mga6-5.0.24-6.mga6 vboxadditions-kernel-desktop-latest-5.0.24-6.mga6 x11-driver-video-nvidia340-340.96-37.mga6.nonfree How reproducible: Only tested once. Some commands outputs: # dkms status nvidia340, 340.96-37.mga6.nonfree, 4.6.3-desktop-1.mga6, x86_64: installed vboxadditions, 5.0.22-2.mga6, 4.6.3-desktop-1.mga6, x86_64: installed-binary from 4.6.3-desktop-1.mga6 vboxadditions, 5.0.24-2.mga6, 4.7.0-desktop-0.rc7.1.mga6, x86_64: installed-binary from 4.7.0-desktop-0.rc7.1.mga6 # systemctl status dkms-autorebuild.service รข dkms-autorebuild.service - run dkms_autoinstaller on every boot to rebuild dkms modules for newly booted kernels Loaded: loaded (/usr/lib/systemd/system/dkms-autorebuild.service; enabled; vendor preset: enabled) Active: active (exited) since Ter 2016-07-12 11:29:30 WEST; 5min ago Process: 939 ExecStart=/usr/sbin/dkms-autorebuild.sh (code=exited, status=0/SUCCESS) Main PID: 939 (code=exited, status=0/SUCCESS) CGroup: /system.slice/dkms-autorebuild.service Jul 12 11:28:36 marte systemd[1]: Starting run dkms_autoinstaller on every boot to rebuild dkms modules for newly booted kernels... Jul 12 11:28:38 marte dkms-autorebuild.sh[939]: nvidia340 (340.96-37.mga6.nonfree): Installing module. Jul 12 11:28:38 marte dkms-autorebuild.sh[939]: dkms build -m nvidia340 -v 340.96-37.mga6.nonfree -k 4.7.0-desktop-0.rc7.1.mga6 -a x86_64 -q --no-clean-kernel Jul 12 11:29:30 marte dkms-autorebuild.sh[939]: (bad exit status: 10) Jul 12 11:29:30 marte dkms-autorebuild.sh[939]: Build failed. Installation skipped. Jul 12 11:29:30 marte systemd[1]: Started run dkms_autoinstaller on every boot to rebuild dkms modules for newly booted kernels. # journalctl -b -u dkms-autorebuild.service -- Logs begin at Sex 2016-05-06 16:32:09 WEST, end at Ter 2016-07-12 11:36:22 WEST. -- Jul 12 11:28:36 marte systemd[1]: Starting run dkms_autoinstaller on every boot to rebuild dkms modules for newly booted kernels... Jul 12 11:28:38 marte dkms-autorebuild.sh[939]: nvidia340 (340.96-37.mga6.nonfree): Installing module. Jul 12 11:28:38 marte dkms-autorebuild.sh[939]: dkms build -m nvidia340 -v 340.96-37.mga6.nonfree -k 4.7.0-desktop-0.rc7.1.mga6 -a x86_64 -q --no-clean-kernel Jul 12 11:29:30 marte dkms-autorebuild.sh[939]: (bad exit status: 10) Jul 12 11:29:30 marte dkms-autorebuild.sh[939]: Build failed. Installation skipped. Jul 12 11:29:30 marte systemd[1]: Started run dkms_autoinstaller on every boot to rebuild dkms modules for newly booted kernels. # dkms build -m nvidia340 -v 340.96-37.mga6.nonfree -k 4.7.0-desktop-0.rc7.1.mga6 Preparing kernel 4.7.0-desktop-0.rc7.1.mga6 for module build: (This is not compiling a kernel, just preparing kernel symbols) Storing current .config to be restored when complete Running Generic preparation routine make mrproper..... using /proc/config.gz make oldconfig..... make prepare.... Building module: cleaning build area.... 'make' -j4 SYSSRC=/lib/modules/4.7.0-desktop-0.rc7.1.mga6/build module; 'make' -j4 SYSSRC=/lib/modules/4.7.0-desktop-0.rc7.1.mga6/build -C uvm module KBUILD_EXTMOD=/var/lib/dkms/nvidia340/340.96-37.mga6.nonfree/build/uvm............ cleaning build area.... cleaning kernel tree (make mrproper)..... DKMS: build Completed. # dkms install -m nvidia340 -v 340.96-37.mga6.nonfree -k 4.7.0-desktop-0.rc7.1.mga6 Running module version sanity check. nvidia340.ko.xz: - Installation - Installing to /lib/modules/4.7.0-desktop-0.rc7.1.mga6/dkms/drivers/char/drm/ nvidia-uvm.ko.xz: - Installation - Installing to /lib/modules/4.7.0-desktop-0.rc7.1.mga6/dkms/drivers/char/drm/ depmod.............. DKMS: install Completed. # dkms status nvidia340, 340.96-37.mga6.nonfree, 4.6.3-desktop-1.mga6, x86_64: installed nvidia340, 340.96-37.mga6.nonfree, 4.7.0-desktop-0.rc7.1.mga6, x86_64: installed vboxadditions, 5.0.22-2.mga6, 4.6.3-desktop-1.mga6, x86_64: installed-binary from 4.6.3-desktop-1.mga6 vboxadditions, 5.0.24-2.mga6, 4.7.0-desktop-0.rc7.1.mga6, x86_64: installed-binary from 4.7.0-desktop-0.rc7.1.mga6
Assignee: bugsquad => tmb
I have a similar situation with 64-bit Mga6-RC classic install on real h/w (non-EFI non-GPT) desktop with nVidia GTS250 and the 340 driver. Appeared to go through all the motions of installing the proprietary driver but it finished up with 'nouveau'. Then - after a kernel update yesterday - on reboot there was the dreaded "...problem configuring the graphics display". Tried using the low-level drakconf/graphics to fix that, but although I eventually got it to offer the proprietary driver, it didn't happen, so I'm dead in the water for the moment. You say above: "A subsequent manually started dkms build was successful and after a update- alternatives, ldconfig and reboot the system is now using the nvidia340 driver without issues. Please see commands and outputs below for details." - but from all the following detail I am unable to figure out how to do that to give it a try...
CC: (none) => maurice
P.S. Booting was hanging after displaying: "nvidia 340 (340.96-38.mga6.nonfree): Already installed on this kernel. Checking for new hardware" Today I tried XFdrake again (which before had offered the proprietary driver, which the above message says *is* installed, but wasn't), but that process seemed to be a no-op, but today I fiddled around some more with XFdrake and eventually got to a list of offered settings, which - when OK'd - brought up the promising message "You need to reboot to reflect the changes". After reboot, I now have the 'nvidia 340.96' driver in action! BUT - and this may be a totally separate problem - the Folder desktop display of icons is NOT honouring the 'Icons Locked' and Widgets locked' settings, but instead the icons are arranged in rows along the top of the screen.
Mass-reassigning all bugs with "kernel" in the summary that are still assigned to tmb (or wrongly assigned to someone with "tmb" in his e-mail address) to the kernel packagers group, but without adding "kernel" to the SRPM field. Please reassign if needed, or add kernel to the SRPM field if this is correct.
Assignee: tmb => kernel
(In reply to Marja van Waes from comment #3) > Mass-reassigning all bugs with "kernel" in the summary that are still > assigned to tmb (or wrongly assigned to someone with "tmb" in his e-mail > address) to the kernel packagers group, but without adding "kernel" to the > SRPM field. > > Please reassign if needed, or add kernel to the SRPM field if this is > correct. For this bug, the driver maintainers group might be more appropriate. CC'ing them and letting them decide. Btw, Maurice, is this bug still valid?
Keywords: (none) => NEEDINFOCC: (none) => drivers, marja11
Since the Aug.13 classic ISO's, I no longer have this problem, so am now happy bunny. :-)
Thx Maurice :-) @ PC_LX Is this bug still valid for you (with nvidia340-340.96-38.mga6.nonfree, kernel-4.7.2-1.mga6, etc.)?
My cauldron installation has the most up-to-date kernel, dkms and nvidia packages installed and working but I don't remember if the newer nvidia packages were installed without issues or if I had to help a bit. I'm currently trying to make another "Mageia 5" to "Mageia Cauldron" upgrade but some other packages (unrelated to this issue) are causing problems. If I can get these other issues sorted I'll check if the newer nvidia packages install correctly and update this bug.
My personal experience with 340 was not good; there were multiple issues. I recommend moving to the next driver (346 I think) and the problems (including DKMS) should be gone.
CC: (none) => jiml
(In reply to PC LX from comment #7) > If I > can get these other issues sorted I'll check if the newer nvidia packages > install correctly and update this bug. Thx :-) (In reply to jim l from comment #8) > My personal experience with 340 was not good; there were multiple issues. I > recommend moving to the next driver (346 I think) and the problems > (including DKMS) should be gone. Thx for the feedback, Jim. Reassigning to the driver maintainers group, then :-)
CC: drivers => kernelAssignee: kernel => drivers
No can do... nvidia 346 drops some hw support compared to 340 series. If nvidia-current does not support that hw, I guess a nvidia346 is needed with matching ldetect-lst updates for specific hw. And any dkms "issues" is only a question about possible missing patches to fix "broken builds"...
CC: (none) => tmb
This thread describes some of my tribulations with the 340 driver. I did not realize, until the end of the thread, that the problem WAS the 340 driver. There are multiple regressions in that driver; if hardware support is an issue, I suggest you roll back to the 334 driver. https://forums.mageia.org/en/viewtopic.php?f=7&t=8839&p=53814&hilit=virtual+consoles&sid=9c597027b32b6064421e058695f2b592#p53814 Alternatively, the hardware support could be added back into the 346 driver, but this would be a non-trivial exercise. Point is that the 340 driver is basically just broken.
I've done another upgrade from Mageia 5 to cauldron and the dkms built and installed the nvidia driver 340.96 correctly. The driver is also working correctly. # dkms status nvidia340, 340.96-38.mga6.nonfree, 4.4.16-desktop-1.mga5, x86_64: installed nvidia340, 340.96-38.mga6.nonfree, 4.7.3-desktop-1.mga6, x86_64: installed virtualbox, 5.1.4-1.mga6, 4.4.16-desktop-1.mga5, x86_64: installed virtualbox, 5.1.4-1.mga6, 4.7.3-desktop-1.mga6, x86_64: installed
CC: kernel => (none)Assignee: drivers => kernel
Have done more upgrades from Mageia 5 to cauldron and the dkms driver packages are installing correctly so I'm marking this bug as fixed. # rpm -qa | grep -Ei 'kernel|dkms|nvidia' | sort dkms-2.0.19-37.mga6 dkms-minimal-2.0.19-37.mga6 dkms-nvidia340-340.98-1.mga6.nonfree dkms-virtualbox-5.1.8-2.mga6 kernel-desktop-4.8.7-1.mga6-1-1.mga6 kernel-desktop-4.8.8-1.mga6-1-1.mga6 kernel-desktop-4.8.9-1.mga6-1-1.mga6 kernel-desktop-devel-4.8.7-1.mga6-1-1.mga6 kernel-desktop-devel-4.8.8-1.mga6-1-1.mga6 kernel-desktop-devel-4.8.9-1.mga6-1-1.mga6 kernel-desktop-devel-latest-4.8.9-1.mga6 kernel-desktop-latest-4.8.9-1.mga6 kernel-firmware-20160914-1.mga6 kernel-firmware-nonfree-20160914-2.mga6.nonfree kernel-userspace-headers-4.8.9-1.mga6 nvidia340-cuda-opencl-340.98-1.mga6.nonfree nvidia340-devel-340.98-1.mga6.nonfree x11-driver-video-nvidia340-340.98-1.mga6.nonfree
Status: NEW => RESOLVEDResolution: (none) => FIXED