NVIDIA-Linux-x86_64-418.113 wouldn't build

There have been changes to the driver layout and libs meanwhile, most of all glvnd and especially its egl counterpart. So if centos 8 doesn’t support this, installing the runfile driver without thinking about the options to use, you’ll break it. That’s why I recommended using the repo driver.

Again, I tried repo driver. Same result.
And repodriver is seems much less aware about Centos kernel.
For e.g Centos has it’s current kernel modules installed in /lib/modules/4.18.0-240.22.1.el8_3.x86_64
While the repo driver installs nvidia driver in /lib/modules/4.18.0-240.22
It creates a new directory 4.18.0-240.22 and installs drivers there. Of course the kernel doesn’t see it until you have to manually copy them to the right directory 4.18.0-240.22.1.el8_3.x86_64
But anyway… The result is stays the same - a black screen.

People, you are official nvidia guys, aren’t you? Do you provide a tech support here? To be honest I don’t understand if I can get a some sort of support here.
I described my problem, sent you all logs I have. Could you please stop telling fortunes, find the reason and give me some exact advice? Are there driver developers here? Can they say what does the hecking undefined symbol: glamor_egl_get_driver_name mean at all?
Do they have CentOS? Would they just test the driver against a regular CentOs distro? Do they test the driver at all?
I have:

  1. Nvidia GPU
  2. Nvidia official driver.

They don’t work together. You have all the logs you asked for and I have followed all the steps you recommended.
What’s next?

This is a user2user forum, I’m not in any form affiliated or what-so-ever with nvidia.
In your case, not the nvidia driver is failing but the modesetting driver which is part of the xorg server which is distributed alongside centos.
Moreover, you’re hitting a bug which is perfectly new to me. I’d recommend to install a distribution which supports hybrid graphics notebooks, check if that works and revert to centos bug tracker if so.

Thanks for the help, but sorry, I don’t agree with you.
As I said all is working fine with an older kernel and an older nvidia driver.
The differences are the kernel and the driver. NOT the xorg server. It stays the same.
The older 418 driver is not built against a newer kernel with a stupid error I mentioned before . fatal error: drm/drmP.h: No such file or directory Whose the issue is it? xorg? I don’t think so.
While the current driver doesn’t work with the same xorg server because … I don’t know… it requires some calls which xorg can’t find and call… But only with the certain nvidia driver. There is no problem with the native i915 driver or 418 nvidia driver which the same xorg works well with.
All what I see this - the current 460 nvidia driver was not well tested with a regular CentOS distro. I’m sure a hybrid graphics configuration doesn’t mean anything significant here.

So who is responsible to make the driver working with a certain Linux kernel and a certain distro? I’m pretty sure that this a driver developers problem, isn’t it? Not some 3-rd party repos and so on…

$ apt-file search modesetting_drv.so
xserver-xorg-core: /usr/lib/xorg/modules/drivers/modesetting_drv.so
xserver-xorg-core-hwe-18.04: /usr/lib/xorg/modules/drivers/modesetting_drv.so

I’m on a ubuntu based system but that isn’t relevant. The modesetting driver is not part of the nvidida driver.
Glamor also isnt:
https://www.freedesktop.org/wiki/Software/Glamor/

To enable Glamor, the following steps are needed:

  1. Rebuild the mesa using the parameter: --with-egl-platforms=x11,drm --with-gallium-drivers= --enable-gbm --enable-shared-glapi --enable-glx-tls.
  2. Rebuild the xf86-video-intel driver, add parameter --enable-glamor to enable glamor module which is embedded in intel driver.
  3. Build and install glamor source. The Glamor source can be get at git://anongit.freedesktop.org/git/xorg/driver/glamor

So this looks very much like a problem the distro has to take care of.


Nvidia only provides one generic .run file installer. With the huge variety in linux distributions and their more or less subtle differences, it would be a huge task to package the driver for every distribution and every release of them.
That is the downside of the diversity in the linux world. Distros pick up the nvidia driver and adapt it to their structure. That’s just the way it is. So if the repo driver does not integrate with the distros system, please report this to them. Nvidia will not help you out there.

The page you gave a link to says PLEASE NOTE: Glamor has been merged into Xorg server 1.16 (released July 2014)
And that page was last edited Mon May 7 16:24:22 2018…
So I don’t think one should make some special steps to enable this stuff in 2021… …And again, even a few weeks ago everything worked well with 418 driver. I don’t buy this shit. Sorry.

As for Nvidia… Well I thought Nvidia is highly interested in order their driver to work with popular distros. I don’t have some exotic distro. It’s Centos, guys! I’m pretty sure Centos is in a top 10 most used distros in the world.
But you just said me that Nvidia almost gives a flying fuck if their driver works with it or not.
Seriously?
Ok then. Sorry for bothering you, guys. I’m bouncing away…

This was just to show, that it’s an xorg and intel thing.
On a hybrid system those two components and the nvidia driver have to work together. The distro must take care of the correct packaging. Many of them then even leave it to the user to manually configure their interaction.
Please mind your language.


Nvidia does not provide a packaged driver for any distro. Nothing any of us users can do about it.

glamor_egl_get_driver_name is a function of the Xserver’s glamoregl module which was added in Xserver 1.20.7:
https://cgit.freedesktop.org/xorg/xserver/commit/glamor?id=195c2ef8f9f07b9bdabc0f554a9033b7857b99c7
Which the accompanied modesetting driver is relying on.
The glamoregl module provided by centos on your system is

(II) Module glamoregl: vendor="X.Org Foundation"
         compiled for 1.20.3, module version = 1.0.1

so this might be the problem.

Further investigation done, extracting libglamoregl.so from the currently installed package
https://rpmfind.net/linux/RPM/centos/8.3.2011/appstream/x86_64/Packages/xorg-x11-server-Xorg-1.20.8-6.1.el8_3.x86_64.html
yields that the included library is indeed outdated thus broken.
Plain Centos packaging failure.

Thanks. So nothing of any previous suggestions might help… 3rd party repos, glamor enabling stuff…

Well could you please send me a “working” libglamoregl.so library from other disto? I could replace mine with the working one and check.

I don’t know if that works that way but you could try to use it from the previous centos package:
https://rpmfind.net/linux/centos/8.3.2011/AppStream/x86_64/os/Packages/xorg-x11-server-Xorg-1.20.8-6.el8.x86_64.rpm
The bug was obviously introduced by the latest centos update.
The path is /usr/lib64/xorg/modules/

1 Like

NB: don’T install the package, just unzip the needed file.

I replaced the library libglamoregl.so with one, taken from the previous centos xorg package as you suggested.
And guess what?

Everything started working!!!  
With current 460 driver installed from *.run  and the config files  optimus.desktop and 10-nvidia.conf 

I’m happy!!! At least until the next CentOS update happens))
Thank you very much for your help!
Great job!

Hi, MikeZag!
Would you mind to post the procedures you did to install the nvidia driver? I am new with this procedures and I`m having the same issues you had. The only diference between my video card and yours is that mine is Geforce GTX 960M.

Thank you!

Could you please clarify, what versions of Centos and xorg-x11-server-Xorg package do you have now?

Hi, MikeZag!
I have the version xorg-x11-server-Xorg-1.20.10-1 and I`m trying to install the nvidia driver on CentOS 8.
Before my kernel was upgraded to any version closer to 4.18.0-240 I was able to do the automated driver installation following the guidelines provided by bumblebee – Linux at NC State, but somehow from what I see in the logs the NVIDIA-Linux-x86_64-418.56.run and current drivers don’t compile anymore.

The installation mode suggested by bumblebee – Linux at NC State that I chose was the one that uses an automated script that uses VirtualGL and DKMS to generate the necessary compiled drivers each time that occurs a kernel update. By analyzing the logs I see that the problem is occurring in the compilation of drivers after having updated the CentOS kernel to the latest versions. I’ll attach here the compilation script and the installation failure log, if anyone can help, I appreciate it. The script extracts the build files from the NVIDIA-Linux-x86_64-418-xx executable and in some cases can patch some file, which was not my case, even though I tried.

bumblebee-nvidia (24.4 KB)
nvidia-installer.log (215.8 KB)

Dicksonrafael
We’ve got absolutely different problems. My problem was in black screen cased by improper version of Xserver’s libglamoregl.so library . Compilation passes without a hitch though .
And you are getting compilations errors.
I’m sure someone here can help you.