The Latest NVIDIA Linux Driver Release.

1337_powerslacker · 12-06-2018, 07:13 PM

For those of you who own NVIDIA cards and are attempting to run the 4.20-rcX series of kernels, please be advised that a patch is required for the driver to work. (Patch is at bottom of page).

Happy Slacking!

cwizardone · 12-07-2018, 10:25 AM

415.22

Nvidia has just released their latest "Short Lived Branch" driver, version 415.22.

https://www.nvidia.com/Download/driv...x/141072/en-us

Release highlights:

Quote:

Fixed a regression that prevented applications based on the Unity engine from starting.
Added support for VK_EXT_transform_feedback.
Fixed a build failure, "unknown type name 'ipmi_user_t'", when building the NVIDIA kernel module for Linux kernel 4.20 release candidates.

USUARIONUEVO · 12-07-2018, 07:29 PM

Hi , anyone know how can i add "nvidia" to dkms ?

Slackbuilds no have this , and when kernel update need to rebuild manually.

Thanks!

1337_powerslacker · 12-07-2018, 08:23 PM

Quote:

Originally Posted by USUARIONUEVO

Hi , anyone know how can i add "nvidia" to dkms ?

Slackbuilds no have this , and when kernel update need to rebuild manually.

Thanks!

When you boot into console mode and execute the .run file, it will offer to register the driver with DKMS automatically. Select 'yes', and that's all you need to do.

USUARIONUEVO · 12-07-2018, 08:58 PM

Quote:

Originally Posted by 1337_powerslacker

When you boot into console mode and execute the .run file, it will offer to register the driver with DKMS automatically. Select 'yes', and that's all you need to do.

I use the slackbuild package , NO DKMS support.

But i understand , need to exec the run file , in a init3 (console) , and install again (i supossed only one time cause after this DKMS registered).

I prefer DKMS , support in the slackbuild , but at the moment i go run the file in console.

Thanks.

cwizardone · 12-12-2018, 10:10 AM

415.23

Nvidia has just released their latest "Short Lived Branch" driver, version 415.23.

https://www.nvidia.com/download/driv...x/141300/en-us

Release highlights:

Quote:

Fixed a build failure, "implicit declaration of function 'vm_insert_pfn'", when building the NVIDIA DRM kernel module for Linux kernel 4.20 release candidates.

kingbeowulf · 12-15-2018, 07:42 PM

Quote:

Originally Posted by USUARIONUEVO

Hi , anyone know how can i add "nvidia" to dkms ?

Slackbuilds no have this , and when kernel update need to rebuild manually.

Thanks!

I very much doubt I will add DKMS to the SBo nvidia slackbuild scripts. There are serveral nvidia kernel modules (nvidia-modeset, nvidia-drm, nvidia-uvm, nvidia) and the added dkms dependency and complexity is not worth the effort. After updating the kernel, you only need to re-run the nvidia-kernel script and upgradepkg. This is not that hard and very reliable.

After all, Slackware is not Ubuntu, etc or a rolling release like Arch.

cwizardone · 12-18-2018, 06:25 PM

415.25

Nvidia has just released their latest "Short Lived Branch" driver, version 415.25.

https://www.nvidia.com/Download/driv...x/141448/en-us

Release highlights:

Quote:

Added support for the following GPU:
TITAN RTX
Fixed a bug that could prevent display detection from working on displays connected to some notebook docking stations.
Fixed a bug which could cause VK_KHR_external_semaphore_fd operations to fail.

Question: What can one do with a $10,000.00US graphics card to make it worth the price?

dugan · 12-18-2018, 06:52 PM

Quote:

Originally Posted by cwizardone

Question: What can one do with a $10,000.00US graphics card to make it worth the price?

Use it to power a billboard.

https://youtu.be/INkeQXetdWs

kingbeowulf · 12-23-2018, 11:51 PM

Quote:

Originally Posted by RadicalDreamer

The last couple of NVIDIA driver updates seem to mess up Plasma 5 occasionally when starting using startx with multiple monitors. I boot up to a blank screen and have to use computer shortcuts to navigate Plasma 5 (alt+space for krunner, etc). I usually F1+Ctrl+Alt and startx again to get it to work correctly. I had the 3 screens off for a couple hours and it did it. I notice it pauses at a blank screen longer than it used to between switching accounts with SDDM. Any ideas?

I would use the appropriate device and screen entries in /etc/X11/xorg.conf.d to configure multiple monitors instead of letting X.org and KDE automagically configure. This seems to work ok for 1 monitor, but 2+ seems to be unstable, at least here, given how nvidia stomps on various X server and mesa components.

kingbeowulf · 12-23-2018, 11:56 PM

I sent over nvidia-410.78 LLB updates to Slackbuilds.org for those installing that way. The "beta" version is at linuxgalaxy.org. I think I fixed the libglx.so error. Working fine so far here.

As always, bug reports are welcome.

RadicalDreamer · 12-24-2018, 02:05 PM

Quote:

Originally Posted by kingbeowulf

I would use the appropriate device and screen entries in /etc/X11/xorg.conf.d to configure multiple monitors instead of letting X.org and KDE automagically configure. This seems to work ok for 1 monitor, but 2+ seems to be unstable, at least here, given how nvidia stomps on various X server and mesa components.

I used nvidia-settings to create my xorg.conf. I do not know how to do more than this in my xorg.conf:

Code:

Section "Screen"

    Identifier     "Screen0"
    Device         "Device0"
    Monitor        "Monitor0"
    DefaultDepth    24
    Option         "UseNvKmsCompositionPipeline" "false"
    Option         "Stereo" "0"
    Option         "nvidiaXineramaInfoOrder" "DFP-2"
    Option         "metamodes" "DVI-D-0: nvidia-auto-select +5760+480 {ForceCompositionPipeline=On}, DP-0: nvidia-auto-select +1920+0 {ForceCompositionPipeline=On}, DP-4: nvidia-auto-select +0+480 {ForceCompositionPipeline=On}"
    Option         "SLI" "Off"
    Option         "MultiGPU" "Off"
    Option         "BaseMosaic" "off"
    SubSection     "Display"
        Depth       24
    EndSubSection
EndSection

kingbeowulf · 12-25-2018, 02:57 AM

Well, for one thing you don't have one monitor - you have 3 - and using metamodes with nvidia-auto-select can confuse X and your DE, especially when powering down. On power up it autodetects wrong. Also in KDE/Plasma make sure their settings utility isn't trying to override xord.conf settings. Set up the monitor sections, then attach each to one or more screens. you also need a server layout section, for example,

Code:

Section "ServerLayout"
    Identifier     "Layout0"
    Screen      0  "Screen0" 0 0
    Screen      1  "Screen1" RightOf "Screen0"
    Option         "Xinerama" "0"
EndSection

I use 2 independent X screens on 2 GPUs so yours will be a bit different. Once you get ANY screen up, run the nvidia-settings gui (advanced) to configure size and position and then save that configuration to an xorg.conf file; it should save more details than you showed. With the proprietary Nvidia drivers, you need to use nvidia-xconfig and nvidia-settings to configure, not the DE tools or xrandr as these can be problematic.

I can probably dredge up some examples once I find them.

RadicalDreamer · 12-25-2018, 08:25 PM

Thank you for the information. I tried using different Screens before but it didn't work too well. I couldn't move my cursor from one screen to the next. I have one GPU. I'll try again with your suggestions.

cwizardone · 01-03-2019, 10:35 AM

410.93

Nvidia has released their latest "Long Lived Branch" driver, version 410.93.

https://www.nvidia.com/download/driv...x/141700/en-us

The Release Highlights,

Quote:

Added support for the following GPUs:
Quadro RTX 8000
Fixed a bug that could prevent display detection from working on displays connected to some notebook docking stations.
Fixed a bug which could cause VK_KHR_external_semaphore_fd operations to fail.
Fixed a build failure, "implicit declaration of function
'vm_insert_pfn'", when building the NVIDIA DRM kernel module for Linux
kernel 4.20 release candidates.
Fixed a build failure, "unknown type name 'ipmi_user_t'", when building the NVIDIA kernel module for Linux kernel 4.20 release candidates.
Fixed a bug that caused kwin OpenGL compositing to crash when launching certain OpenGL applications.
Added a new kernel module parameter, NVreg_RestrictProfilingToAdminUsers, to allow restricting the use of GPU performance counters to system administrators only.