Package Details: nvidia-xrun-git 0.3.79.gaf3b734-1

Git Clone URL: https://aur.archlinux.org/nvidia-xrun-git.git (read-only, click to copy)
Package Base: nvidia-xrun-git
Description: Script to run dedicated X server with discrete nvidia graphics (git version)
Upstream URL: https://github.com/Witko/nvidia-xrun
Licenses: GPL
Conflicts: nvidia-xrun
Provides: nvidia-xrun, nvidia-xrun-git
Submitter: robus
Maintainer: robus (Witko)
Last Packager: Witko
Votes: 19
Popularity: 0.012798
First Submitted: 2016-10-20 20:37 (UTC)
Last Updated: 2019-05-07 10:32 (UTC)

Latest Comments

seniorkiller commented on 2022-06-20 02:55 (UTC)

somehow i got an error with this tee /sys/bus/pci/devices/0000:00:01.0/power/control no such file or dir

i followed the wiki but still it doesnt work. i ever use it before and it work fine then i give optimus manager a try

falling back to xrun because optimus manager doenst support ly dm and i cant get this going again

ive tried to change the /etc/default/nvidia-xrun controler bus but i found 4 PCI express in lspci, just used the first one and it didnt work.

i some how manage to get it to the black screen but it instanly stopped(the server terminate by it self)

topcat01aur commented on 2022-04-29 20:24 (UTC)

@edoantonioco yes I use it on multiple machines. Works perfectly. Make sure to follow my recommendations and double check ALL folders for stray configuration files.

edoantonioco commented on 2022-01-12 21:42 (UTC)

Is this working properly? Since its flagged out of date

topcat01aur commented on 2021-11-22 00:22 (UTC)

There seems to a race condition between nvidia-xrun-pm.service and systemd-moudules-load.service which loads the nvidia modules. If the latter runs first, nvidia-xrun-pm will hang (actually the tee command) during device removal. If on the other hand nvidia-xrun-pm runs first then it will succeed, and later the modules will fail to load with an error (which is harmless but ugly).

The solution is to blacklist the nvidia modules as already described in https://wiki.archlinux.org/title/nvidia-xrun#NVIDIA_GPU_fails_to_switch_off_or_is_set_to_be_default. In this case the subsequent failed module load attempt is also avoided. This of course does not affect running nvidia-xrun later.

I understand all this is known but probably still worth a mention.

radiomike commented on 2021-11-15 12:33 (UTC) (edited on 2021-11-15 12:35 (UTC) by radiomike)

Another solution, as I wasn't able to get others working.

In the serverlayout section of /etc/X11/nvidia-xorg.conf change Screen 1 "nvidia" to Screen "nvidia"

Also placing 10-nvidia-drm-outputclass.conf from https://wiki.archlinux.org/title/NVIDIA_Optimus#Use_NVIDIA_graphics_only in /etc/X11/nvidia-xorg.conf.d seems to help with some small issues I came accross.

aiyion commented on 2021-11-14 16:19 (UTC)

Crap. It worked. May somebody explain, what we just did?

aiyion commented on 2021-11-14 16:16 (UTC) (edited on 2021-11-14 16:19 (UTC) by aiyion)

This feels wrong -.-' Will try. Isn't there some sort of config checker, we can throw at this? I mean, sure I'll delete that folder; but by all means; your contents possibly aren't mine? Blindly replacing and deleting stuff is not a proper solution...

Lander commented on 2021-11-14 13:14 (UTC)

@aiyion Got the same issue, fixed it by deleting the contents of the /etc/X11/nvidia-xorg.conf.d directory

aiyion commented on 2021-11-14 12:54 (UTC)

xorg does indeed crash, but replacing the contents hangs the system on my end.

HariSeldon commented on 2021-11-12 18:28 (UTC)

topcat01aur You're a gentleman and a scholar, thanks!

topcat01aur commented on 2021-11-11 21:43 (UTC)

As of xorg-server 21.1.1-2 the configuration in /etc/X11/nvidia-xorg.conf causes the X server to crash. This can be fixed by replacing the contents of this file with those suggested in https://wiki.archlinux.org/title/NVIDIA_Optimus#Use_NVIDIA_graphics_only for the file /etc/X11/xorg.conf.d/10-nvidia-drm-outputclass.conf.

gesh commented on 2021-10-12 20:13 (UTC)

Just noticed the permissions put on /etc/X11/{nvidia-xorg.conf.d,xinit/nvidia-xinitrc.d} are 555, any reason for this? This differs from the non-nvidia versions of these directories.

rsa commented on 2020-11-22 00:12 (UTC)

Please install the service to /usr/lib/systemd/system instead of /etc/systemd/system

Also, if you make it depend on NVIDIA-MODULE instead of nvidia, you will fix the issue reported by @MartinX3

MartinX3 commented on 2020-09-27 21:15 (UTC)

If you run the LTS kernel, it wants to force you to install the packages "nvidia", instead "nvidia-lts" and "kernel" instead "kernel-lts". Please fix it.

freshmaster313 commented on 2020-06-13 16:16 (UTC)

For all those asking for 390xx support, you just have to clone and edit the PKGBUILD and change nvidia to nvidia-390xx on the depends array and then build the pakage the arch way.

Lander commented on 2020-04-07 06:24 (UTC) (edited on 2020-04-07 06:24 (UTC) by Lander)

Error while updating nvidia-dkms:

error: failed to prepare transaction (could not satisfy dependencies) :: installing nvidia-dkms (440.64-10) breaks dependency 'nvidia' required by nvidia-xrun-git

riazufila commented on 2020-02-15 15:04 (UTC) (edited on 2020-02-16 04:04 (UTC) by riazufila)

Does nvidia-xrun-git not work with nvidia-390xx-dkms? Installing nvidia-xrun-git warns of replacing nvidia-390-dkms with nvidia-dkms. Is this intended?

jpegxguy commented on 2019-09-12 01:10 (UTC)

The install script sounds cool. Either would work I guess. By the way the systemd service should be installed in /usr/lib/systemd/system.

Ashark commented on 2019-07-18 03:45 (UTC)

I suggest adding some way of keeping user changes of configs. Currently if you make some changes (for example I need to adjust pci ids), then after removing a package and then reinstalling it, I need to make adjustments again. I am making changes in: etc/X11/nvidia-xorg.conf (just fix nvidia gpu pci id) etc/default/nvidia-xrun (fix nvidia gpu and pci bridge ids) I am not sure (at the time of writing this) if it is intended to edit them directly or specifying correct ids in some overriding file (at least there are such examples in project's Readme and in ArchWiki).

To keep user changes it is possible to use backup array in PKGBUILD. Or I can think about another way of solving a problem: adding an .install script to the package that will automatically detect nvidia pci id and pci bridge id and then change it after installation.

Any thoughts?

Ashark commented on 2019-07-16 20:37 (UTC)

2019-05-07 Updated to latest + added install of the pm service

michelesr commented on 2019-05-05 22:07 (UTC)

Can you please add nvidia-xrun-pm.service to the package?

install -Dm 644 nvidia-xrun-pm.service "$pkgdir/etc/systemd/system/nvidia-xrun-pm.service"

Ernie commented on 2019-01-26 15:01 (UTC)

/usr/bin/nvidia-xrun sources /etc/default/nvidia-xrun. It should be included in PKGBUILD install -Dm 644 config/nvidia-xrun "$pkgdir/etc/default/nvidia-xrun"

Moreover in /usr/bin/nvidia-xrun for UNLOADING MODULES the command should be rmmod OR modprobe -r

echo "Unloading module ${module}" execute "sudo rmmod ${module}"

robus commented on 2019-01-26 10:19 (UTC) (edited on 2019-01-26 10:22 (UTC) by robus)

@Witko, I finally found the button that gives maintainer permissions. Now you have those.

PedroHLC commented on 2019-01-18 21:16 (UTC)

https://github.com/Witko/nvidia-xrun/pull/90 Seems like bbswitch is no longer needed.

Witko commented on 2018-09-24 18:20 (UTC)

Could I also be maintainer of this one? Ive asked for it quite some time ago. When i want to test new stuff this is a nice way for me to do it. But as im not the maintainer i cant change it when i need to. I also dont want to create a competitive package.

eschwartz commented on 2018-01-23 16:31 (UTC)

That fork has several problems.

1) It only contains updated .spec files for Fedora packaging.

2) It has been merged upstream, because upstream is not, in fact, dead.

timofonic commented on 2017-12-28 07:34 (UTC)

This fork seems more updated... https://github.com/AxelSilverdew/nvidia-xrun

Why bbswitch?

Witko commented on 2017-12-08 21:23 (UTC)

Hi Robus, could you add me as maintainer so i can update the latest changes? Thx!

setzer22 commented on 2017-10-07 11:29 (UTC)

It is definitely possible. I am using this with nvidia-dkms and bbswitch-dkms in my system without problems.

resetko commented on 2017-08-08 21:21 (UTC)

is it possible to use this module with nvidia-dkms? maybe need to list it optional dependency replacement for nvidia?