Bug 22256

Advisory, added to svn:

type: bugfix
subject: Updated nvidia-current packages fixes support for kernel 4.14
src:
  6:
   nonfree:
     - nvidia-current-384.98-2.mga6.nonfree
description: |
  This update fixes the nvidia-current driver so it builds with 4.14
  series kernels.
references:
 - https://bugs.mageia.org/show_bug.cgi?id=22256

Keywords: (none) => advisory
Blocks: (none) => 22166

Upgraded the nvidia driver from 340.104 and ran XFdrake before rebooting.

Rebooted smoothly to the Mate desktop.

System:    Host: vega Kernel: 4.14.9-desktop-1.mga6 x86_64 (64 bit)
Graphics:  Card-1: Intel Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics Controller
           Card-2: NVIDIA GK104 [GeForce GTX 770]
           Display Server: Mageia X.org 119.5 drivers: nvidia,v4l,intel
           Resolution: 2560x1440, 1024x768
           GLX Renderer: GeForce GTX 770/PCIe/SSE2
           GLX Version: 4.5.0 NVIDIA 384.98

Checked graphics performance with glmark2 - about 14% slower than with the older driver after the kernel update on this machine.

CC: (none) => tarazed25

I've added ldetect-lst-0.3.7.4-1.mga6 to this bugreport as earlier update missed some nvidia-current pci ids

tested on x86_64 with current 4.9.56-1 kernel, the 4.14.10 currently in testing and a rebuilt 4.14.11-3 from cauldron.

And since this update is not a version update that would add/remove/change any features, instead is a simple rebuild (to add a missing include a kernel header file in the dkms package source) of the 384.98 drvier we pushed as

http://advisories.mageia.org/MGAA-2017-0116.html

it cant really break anything...

Whiteboard: (none) => MGA6-64-OK

Wondering... The performance drop in comment 2 - might be because of new kernel if that kernel have patches against spectre / meltdown ?

CC: (none) => fri

(In reply to Morgan Leijström from comment #5)
> Wondering... The performance drop in comment 2 - might be because of new
> kernel if that kernel have patches against spectre / meltdown ?

Nope, its an upstream issue, and nVidia is aware of it and working on fixing it  but no ETA so far...

I've resurrected an old box that uses nvidia current and can confirm the OK for mga6-64

With a bit of luck, I should be able to get mga6-32 installed on it later today

CC: (none) => jim

On mga6-32 plasma

Packages installed cleanly:
- dkms-nvidia-current-384.98-2.mga6.nonfree.i586
- ldetect-lst-0.3.7.4-1.mga6.i586
- nvidia-current-doc-html-384.98-2.mga6.nonfree.i586
- x11-driver-video-nvidia-current-384.98-2.mga6.nonfree.i586

No regressions. Seems to be working fine.
OK for mga6-32

$ uname -r
4.9.56-server-1.mga6

# dkms status -m nvidia-current
nvidia-current, 384.98-2.mga6.nonfree, 4.9.56-server-1.mga6, i586: installed 

Machine:   Device: desktop Mobo: ECS model: GeForce7050M-M v: 1.0
           BIOS: American Megatrends v: 080014 date: 09/12/2008
CPU:       Quad core AMD Phenom 9500 (-MCP-) cache: 2048 KB 
           clock speeds: max: 2200 MHz 1: 2200 MHz 2: 2200 MHz 3: 2200 MHz 4: 
           2200 MHz
Graphics:  Card: NVIDIA GF108 [GeForce GT 630]
           Display Server: Mageia X.org 119.5 drivers: nvidia,v4l
           Resolution: 1440x900@59.89hz
           GLX Renderer: GeForce GT 630/PCIe/SSE2/3DNOW!
           GLX Version: 4.5.0 NVIDIA 384.98

This update is now validated and can be pushed to updates

Whiteboard: MGA6-64-OK => MGA6-64-OK MGA6-32-OK
Keywords: (none) => validated_update
CC: (none) => sysadmin-bugs

An update for this issue has been pushed to the Mageia Updates repository.

https://advisories.mageia.org/MGAA-2018-0008.html

Resolution: (none) => FIXED
Status: NEW => RESOLVED