I am having trouble installing nvidia drivers and I don't seem to know why. I am still trying to familiarize my self with UNIX as a whole more so than what I know and in particular with how slackware is operated and maintained. How do I enable the /extra directory, what is slackpkg with respect to the "repository"? Is it just a program to install programs on the slackware iso? How do I install this and what documentation to you suggest?
https://slackware.pkgs.org/14.2/slac...rch-1.txz.html
I know I must disable nouveau and I modified the respective folder with the text needed and made another one with the suggested name as in this article. https://docs.slackware.com/howtos:ha...aphics_drivers
But it is not disabled upon boot and the installer fails. I want to disable nouveau

Those who tell you that Slackware doesn't have a package manager either don't know what they're talking about or are being deliberately obtuse. Slackware uses pkgtools as a package manager. This is a very simple program that manually allows you to install or remove packages on your system. slackpkg, on the other hand, is a shell script for pulling packages off the official repo[s] and updating them using pkgtool.

In short, slackpkg allows you to update your official packages over a network.

There are other directories on the official iso/repo such as /extra which are not part of the standard installation but which are officially supported. You can download these and install them. Once downloaded you can use

make sure you are in the folder you downloaded the package to first.

https://mirrors.slackware.com/slackw...ckware64-14.2/

If in doubt, you can search for your required package through the packages list:

https://packages.slackware.com

Or by using

Look at

for more. 'man' means manual, it will tell you about slackpkg, what it's for and list the most common commands. Type the above command into your terminal to see what happens.

As an aside, typing

into your terminal will very often bring up a manual about that package.

EDIT: note that in Slackare it's possible to install two versions of a package alongside each other. installpkg will do this for you and is useful for if, e.g. you want to install more than one kernel. However, in many cases you can use upgradepkg instead to upgrade.

If you are unsure of whether you have a package installed or not, you can check by using

Using

will upgrade the package if it's already installed rather than installing a new version in parallel.

Okay. Would you suggest I use the slackbuild.com drivers or the ones provided by NVIDIA?

Also what does "noarch" mean and what does it entail? I tried googling it but found nothing.

I'm afraid someone else more au fait with Nvidia than I is going to have to answer that question. But as a rule I would always go for Slackbuild first since they have been tested and built specifically for the distro. As I say, I would wait till someone else answers that question though.

It means it's universal, i.e. not linked to specific architecture - 64bit, 32bit etc.

https://www.lifewire.com/what-is-noarch-package-2193808

I have had systems with Nvidia hardware for a long time. I have tried both methods of installing Nvidia drivers, and have settled on using Nvidia.run files. It works on all systems I have used it on.

I know its not the Slackware way, since the packages file will not be updated with the currently installed driver package. That is a disadvantage.

I found the .SBo nvidia packages difficult to figure out what version of Nvidia driver would be installed. If you go here -->https://www.nvidia.com/object/unix.html and look, you will see several things, drivers for different acxh's, and different versions. If you select any particular version and arch, you can find out what chips sets are supported. You can decide if you want the "long live" branch, or "short live" branch, or your card may work well with an older version, and no need to update. Your choice.

SBo does not give you as many choices. There are also more packages to install. Nvidia's blobs give you the whole thing in one .run file. For me, that is simpler.

Currently I have a current slackware system with an Nvidia card. Kernels change about once a week. You have to remove and re-install the Nvida driver any time you update the kernel. I keep a copy of the Nvidia driver on my system. It makes the update to the Nvidia driver one step, and one file, once the kernel is updated.

This is just my 2 cents worth, I'm sure there are others that have a different opinion. In the end, its up to you to choose your method of install and maintenance of your system.

It depends on how you want to administer your system.
If you have a setup with a single nVidia GPU and just want to use the proprietary driver, then the easiest solution is to use the nVidia .run as suggested by camorri.
If you want a setup where you can cleanly uninstall the proprietary driver, then use the SlackBuild. The nVidia .run proprietary install makes changes that cause problems if not reversed correctly. This forum is littered with threads on this issue.
It signifies that the package is not dependent on the architecture (32bit or 64bit) on which it is installed.

1 members found this post helpful.

Further to Allends post, I have had issues with using the Nvidia run files. I've never been the first to run in to a Nvidia created problem. I little duckduckgo ing, and I found solutions. So, this could be a concern for a newer Slackware user.

For this reason alone, I would suggest a newer user use the .SBo route. Tracking current is recommended for experienced users. Since I do, and I've been using Slackware a few years now, I find this faster. Your choice.

Well it looks like I will be the user that will have my system messed up due to improper driver installation. I wish to make a shell script and C program to fix this issue for future users. I will try the .SBo but it looks bleak.

I mean are older versions of slackware more.... rational?

I mean is there not any majority agreed upon automated scripts to preform these driver downloads , taking into consideration all systems are more or less the same at the genesis?

What legacy build do you suggest I use? I know I am still learning I apologize. I am going the slackbuild way.

So...

Are you attempting to run the 32 bit or 64 bit version of Slackware?

Are you attempting to run version 14.2 or -current?

Have you done a full install of whatever combination of 32/64 bit and 14.2/-current you are trying to run? (You may not bother to install the kdei packages; most of us would consider an installation with all packages other than that a full installation.)


For example, I am running Slackware64 14.2 (aka, version 14.2 and 64 bits). I am running the proprietary NVidia driver for X windows via the packages that I built using the stuff at slackbuilds.org. That works out to be the packages...I haven't installed the xf86-video-nouveau-blacklist package but have added to the bottom of the file /etc/modprobe.d/blacklist

I also do not have an /etc/X11/xorg.conf file. I do have a file named /etc/X11/xorg.conf.d/nvidia.conf (you can really name it anything you want) which contains...

...the important bits are the "Driver" and "BusID" lines. The "Driver" line would be the same for you. The BusID line may not be. I find the correct value by booting into runlevel 3 (which is text-only), log in as root and run the command . In my case, that returns (bold and underline added) That bold and underlined bit is the BusID in a slightly different format. I set my BusID value to the mapped value, save the file, and then run the command . Either X windows will come up or it will fail horribly. In the latter case, I look in the file /var/log/Xorg.0.log which normally gives me an idea on what happened.

There are some restrictions on how to upgrade those nvidia packages; we can talk about those after you are running the proprietary drivers.

There are various "camps" of users in Slackware. Some feel that Nvidia hardware and proprietary software is no big deal. Some choose to install the driver using the .run from Nvidia, while others choose to use the SlackBuilds. Both are valid options, but the second allows Slackware' package manager to be aware of the files.

Another group of users only want to use Nvidia hardware with open-source software (ie nouveau) and don't want Nvidia's proprietary driver to touch their system.

Then there are other users that want to use a GPU manufacturer that better supports open source development, like Intel or AMD, as both manufacturers directly work with X, mesa, and kernel developers to provide better drivers. They may not always be the best, because you can't always open source all the code, but they do seem to be more friendly towards open development.

The options you prefer depend on what your goals are. There is no single right answer. What you choose to do will be based on what your overall goals are for that GPU.

You would use the nvidia-kernel SlackBuild first, followed by the nvidia-driver SlackBuild, but based on your version string, it looks like you might already be running the correct proprietary version.

noarch means "no architecture", which again means there are no binaries in it for either 32-bit Intel or 64-bit AMD/Intel-64, you can install this package (which either contains data files or scripts) in either architecture.

I have always installed the Nvidia drivers for the .run file. However you have to pay attention between the kernel version and the driver version and it is mandatory that you have the kernel source too If you are going to use Steam, you will need to install multilibs and then reinstall the nvidia drivers for the .run file