From ArchWiki
Revision as of 15:34, 11 April 2017 by Lahwaacz (talk | contribs) (Automatic re-compilation of the NVIDIA module with every update of any kernel: remove closed discussion)
Jump to: navigation, search


NVoption Online Version - great tool to make tv-out easy and fast

[I'm using gmplayer with gl and twinview] [1] —This unsigned comment is by Suw (talk) 01:23, 23 March 2006‎. Please sign your posts with ~~~~!

A user found this useful to get TV-Out working with an old Geforce 2 MX: nvtvAUR -- Alad (talk) 13:56, 3 March 2016 (UTC)

'/dev/nvidia0' Input/Output error... suggested fixes

Can anyone verify that the BIOS related suggestions work and are not coincidentally set (either automatically when changing the IRQ or turning off ACPI) while troubleshooting? I have found little information that confirms any of the suggestions would work. The file permissions thing seems to be completely unfounded and never works (as noted in the article) that I could find. It would probably be a good idea if we cleaned out items that have not been verified to work. For my setup I was having this error and none of the items in the wiki nor the many file permission search results worked. -- click, them so hard 19:16, 4 March 2012 (EST)

I've added an Accuracy template, please next time add it yourself so that discussions like this are more visible. -- Kynikos 05:40, 6 March 2012 (EST)

Run a test

There is confusing paragraph saying You can run a test to see if the Xorg server will function correctly without a configuration file.. IMHO, it should be clarified what kind of test the author has in mind, an exact command would be helpful. Currently, this suggestion is simply confusing, especially to less experienced users. --Mloskot (talk) 19:52, 26 November 2012 (UTC)


Several of the commands which are suggested to be run with nvidia-xconfig (such as nvidia-xconfig --twinview) don't work with the current nvidia packages in the repository. I just went through setting mine up so I intend to clean up the ones that I can from my experience. Some don't seem to have a 1:1 replacement (there is a --dynamic-twinview argument now; is that the same as --twinview was?). Teh (talk) 13:10, 20 June 2013 (UTC)

While it shouldn't be necessary to use xconfig to get nvidia working on X, creating a 20-nvidia.conf file is an integral step to fix screen tearing for people that suffer from that issue. Perhaps this section should have this explained and include a link to the nvidia troubleshooting article which contains the section about fixing screen tearing? --TheChickenMan (talk) 21:54, 22 December 2016 (UTC)

Troubleshooting sections link to the main documentation, not vice versa. Otherwise there would be no point in having separate troubleshooting sections. -- Lahwaacz (talk) 22:03, 22 December 2016 (UTC)

Accuracy of driver selection

A thread on the forums revealed that nvidia is more conservative in their suggestion than needed and checking their website tells you to use an older driver (nvidia-340xx) while the latest 'nvidia' one will do just fine. Gusar suggested the removal of the second point in the 'Installing' section. Do we want to get rid of this potentially useful, although somewhat confusing info? -- Karol (talk) 02:52, 15 November 2014 (UTC)

Well, I've rewritten the second point to make it clearer, I hope I got it correctly. However steps 2 and 3 there should probably merged into one, because they deal with the same problem, i.e. finding the correct driver to install. Maybe we could use a table? (just brainstorming). -- Kynikos (talk) 03:47, 16 November 2014 (UTC)

Pacman Hook

Is there some way to alter this so that the hook only runs on transactions where mkinitcpio hasn't already been called by an upgrade to the kernel? Having it run twice doesn't seem like the most efficient solution. It might also be worth mentioning that a second hook file will be needed if also running the linux-lts kernel. TheChickenMan (talk) 02:41, 18 June 2016 (UTC)

If other packages run mkinitcpio from the .install script (like linux), you can't detect this from the hook. -- Lahwaacz (talk) 09:55, 18 June 2016 (UTC)
The hook from a kernel upgrade won't necessarily work because it depends on the order that the packages were upgraded. I just upgraded linux and nvidia and got this during the first mkinitcpio:
  -> Running build hook: [fsck]
==> ERROR: module not found: `nvidia'
==> ERROR: module not found: `nvidia_modeset'
==> ERROR: module not found: `nvidia_uvm'
==> ERROR: module not found: `nvidia_drm'
==> Generating module dependencies
because the nvidia package had not yet been upgraded. Only the second mkinitcpio fixes it. Really the proper fix would be to have a single hook that runs when linux or nvidia is upgraded. But that would involves changing how linux is packaged, and I'm not sure if there's some important reason why it does it in the .install script.
Silverhammermba (talk) 15:09, 13 August 2016 (UTC)
As of linux 4.8.8-2, mkinitcpio is now running as a pacman hook (source). It now upgrade the nvidia package before running mkinicpio, solving the problem.
Terence (talk) 16:18, 17 November 2016 (UTC)
Does that hook fire when nvidia is upgraded and the kernel is not? Don't you need to run mkinitcpio in that case as well? Silverhammermba (talk) 04:42, 19 November 2016 (UTC)
It does it for me at least after changing the hook to work with nvidia-dkms Terence (talk) 16:05, 2 December 2016 (UTC)