Screen appears as blank after Booting up on Fedora 41

Ask Fedora
1

So I installed some Game with Heroic Launcher today, first game I installed on Linux. And some time after that I was unable to open up my terminal app so I rebooted the system.

Since then I have been unable to login into the Desktop Environment I have 3 kernel versions on my System and all of them result in this Blank screen.

I think this is not relevant tbh
I have a backup kernel with the old Fedora 38 installation in the grub list which is the what i installed fedora for the first time. When logging into the Fedora 38 kernel I get some Error more messages about failing to mount boot-efi.mount and more( that’s the 3rd image).

Thanks in advance for any helps.

So I figures out how to get Logs from the Last Boot Session and most of the error messages are in relation to a missing file called libexpat.so.1
Here are the Logs again

3000×3000 1010 KB

3000×3000 834 KB

3000×3000 1.44 MB

2

Since it is saying that it failed to mount /boot/efi, I would suggest running a filesystem check on your ESP. Unfortunately, since your root account is locked, I think you will have to create a rescue boot disk of some sort. The Fedora Linux live installer should work (i.e. the normal installation image that you can download from fedoraproject.org).

3

I have the same problem.
during boot from the f41-live USB I see the line:
…
nouveau 0000:01:00: DRM: failed to create ce channel, -22
…
in light blue
Then once the installed system is coming up, I receive the message in the centre of a blank screen:

NVIDIA kernel module missing. Falling back to nouveau

Then we reach the blank screen with the mouse arrow active but nothing else.
Then, using Ctrl+Alt+F3 i can log in and use the tty terminal command line.

I have tried various ways using rpmfusion to install NVidia drivers but it seems there is nothing compatible for my graphics card (GeForce 210).
Is this correct? Is it thus NOT possible to install F41 on my machine (old i5 with 4 cores and 16GB RAM)?

4

I don’t know exactly which nvidia cards do or do not work with the nouveau driver, but all those video cards should “work” in a fallback VESA mode which ought to be enough for basic use (web browsing, word processing, etc.). Can you not install/use Fedora Linux in “Basic Graphics Mode” (i.e. with nomodeset supplied on the kernel command line at boot)?

5

Thank you for your message. I’ve managed to re-install F40 and now have some catching up to do before the holidays. In I’ll return to the chase in the New Year. Best wishes for the Festive season.

1 Like
6

Now, with the renewed energy from the first signs of spring, I’m trying to get Fedora 41 to work with my nvidia card which worked fine with previous versions of Fedora. When the system boots I get the blank screen but I can login in terminal mode using Ctrl+Alt+F3.
Following nvidia instructions I tried running their recommended installation package NVIDIA-Linux-x86_64-340.108.run
On running this file, it starts up but then stops with an error message which is recorded in the /var/log/nvidia-installation.log file. This latter file also shows other messages as shown here below:
FROM nvidia-installer.log
…
warning: the compiler differs from the one used to build the kernel
The kernel was built by: gcc (GCC) 14.2.1 20250110 (Red Hat 14.2.1-7)
You are using: cc (GCC) 14.2.1 20250110 (Red Hat 14.2.1-7)
…
I know nothing about kernel building and certainly not the difference between the kernel being built by gcc OR cc
Could this be part of my graphics problem? How to preferentially build with gcc or cc?

The last part of the log file shows the terminating installation message:

In file included from /tmp/selfgz1777/NVIDIA-Linux-x86_64-340.108/kernel/nv.c:12:
/tmp/selfgz1777/NVIDIA-Linux-x86_64-340.108/kernel/os-interface.h:27:10: fatal error: stdarg.h: No such file or directory
27 | #include <stdarg.h>
| ^~~~~~~~~~
compilation terminated.
make[3]: *** [scripts/Makefile.build:229: /tmp/selfgz1777/NVIDIA-Linux-x86_64-340.108/kernel/nv.o] Error 1
make[2]: *** [/usr/src/kernels/6.12.15-200.fc41.x86_64/Makefile:1977: /tmp/selfgz1777/NVIDIA-Linux-x86_64-340.108/kernel] Error 2
make[1]: *** [Makefile:236: __sub-make] Error 2
make[1]: Leaving directory ‘/usr/src/kernels/6.12.15-200.fc41.x86_64’
NVIDIA: left KBUILD.
nvidia.ko failed to build!
make: *** [Makefile:197: nvidia.ko] Error 1
→ Error.
ERROR: Unable to build the NVIDIA kernel module.
ERROR: Installation has failed. Please see the file ‘/var/log/nvidia-installer.log’ for details.
(THIS FILE)
You may find suggestions on fixing installation problems in the README available on the Linux driver download page at www.nvidia.com.

ANY SUGGESTIONS ON HOW TO PROCEED?
Thank you.
Derek Jones

7

I’m not sure what that warning is saying, but sudo dnf install gcc ought to be enough. I usually install the full c-development group (sudo dnf install @c-development). There is also a @development-tools group that provides commands like patch, but that probably isn’t necessary.

It looks like the above was the fatal error. That file is provided by the gcc package:

gcc-14.2.1-7.fc41.x86_64 : Various compilers (C, C++, Objective-C, ...)
Repo         : @System
Matched From : 
Filename     : /usr/lib/gcc/x86_64-redhat-linux/14/include/stdarg.h

Hmm, its also in the kernel-devel package:

kernel-devel-6.12.15-200.fc41.x86_64 : Development package for building kernel modules to match the kernel
Repo         : @System
Matched From : 
Filename     : /usr/src/kernels/6.12.15-200.fc41.x86_64/include/linux/stdarg.h

You might want to install/update the kernel-devel package.

8

You might want to install/update the kernel-devel package.

I think 6.12.15-200 kernel-devel is no longer in the repos and you have to get it from Koji. (I ran into this last week when rebuilding nVidia drivers after downgrading from 6.13.4).

9 
 sudo dnf install fedora-repos-archive
 sudo dnf config-manager setopt updates-archive.enabled=no
 sudo dnf install kernel-devel-6.12.15-200.fc41.x86_64  --enablerepo=updates-archive
2 Likes
10

Dear Gregory Lee Bartholomew.

Thank you for your kind replies and suggestions and I do hope all this is not taking up too much of your time.

Following your comment:

I’m not sure what that warning is saying, but sudo dnf install gcc ought to be enough. I usually install the full c-development group ( sudo dnf install @c-development ). There is also a @development-tools group that provides commands like patch , but that probably isn’t necessary.

I have installed

gcc and @c-development and @development-tools as well as kernel-devel (no problem finding in the repos). And rebooted, of course!

The “gcc vs. cc” problem and the “fatal error: stdarg.h: No such file or directory” persist.

Is there a “kernel-source” anywhere which could be installed?

Maybe this could solve the “gcc vs. cc” problem?

Thanks once again for all your help.

Best regards,

Derek Jones

11

I’d focus on the missing stdarg.h problem first. It’s possible your installation of kernel-devel is corrupted somehow. Try running rpm -V kernel-devel. No news is good news, but if that command outputs that some files are missing or changed/corrupted, then you might need to run dnf reinstall kernel-devel to get the files replaced with valid copies. You might do the same with the gcc package.

12

PS. Before running the NVIDIA “run” program, the inxi -G command gives me an Xwayland (very low resolution!!!) server

image
image1600×1200 126 KB

1 Like
13

Hi Derek,

It looks like you are trying to install the drivers downloaded from Nvidia, correct?
If you just install the kernel-devel packages, you should have everything you need.

Installed kernel as of this writing is:
6.13.5-200.fc41.x86_64 #1 SMP PREEMPT_DYNAMIC Thu Feb 27 15:07:31 UTC 2025 x86_64 GNU/Linux

Here is how I do mine:
You must do these as root or via sudo

  1. sudo dnf clean all
  2. sudo dnf upgrade -y (let’s get completely up to date)
  3. download the latest drivers from Nvidia (570.124.04 as of this writing)
  4. sudo chmod 755 {Nvidia_driver_package-version}.run
  5. sudo systemctl set-default multi-user (this completely disables the GUI desktop interfaces at boot)
  6. reboot
  7. login as root or as user with sudo access
  8. exec the nvidia driver install {Nvidia_driver_package-version}.run
  9. just select the default answers as the package installer does its work
  10. reboot and do a quick test by loging in and executing startx at the command line
    IF the desktop GUI environment starts without issue, sudo systemctl set-default graphical and reboot
    ELSE find and fix the reason which is most likely going to be the driver compile did not work/install correctly
    TO start over/uninstall the driver sudo {Nvidia_driver_package-version}.run -uninstall
1 Like
14

Dear Einer,

The message I received during running your suggested NVIDIA driver (NVIDIA-Linux-x86_64-570.124.04.run) which did not work giving me the following message:
WARNING: The following NVIDIA GPUs are supported through NVIDIA legacy Linux graphics drivers and will be ignored by the NVIDIA 570.124.04 Linux graphics driver:
GeForce 210 requires 340.xx
Please visit "Unix Drivers | NVIDIA "
for more information.

Please go to Unix Drivers | NVIDIA, follow the Archive link under the Linux x86_64 heading, follow the link for the 340.108 driver, and then go to the Supported Products List:
Appendix A. Supported NVIDIA GPU Products

"You may find suggestions on fixing installation problemsin the README available for the 340.xx driver download page
I then followed the 340.108 link to:
https://us.download.nvidia.com/XFree86/Linux-x86_64/340.108/README/index.html

And then the successive link to:

and there the Supported Products tab:

Here I found the whole range of the GeForce 200 Series:
including my card: GeForce 210
Therefore the 340.108 driver is the one to use.

Thanks (danke schoen?) for your kind help.
Best regards,
Derek Jones

15

Thank you once again.
I followed your advice and run:
rpm -V kernel-devel. and dnf reinstall kernel-devel and the news was no news (good or not!).
I also did the same with the gcc package.

I also did a thorough check on nvidia websites (see my reply to One Einer below) which confirmed the 340.108 installation file for my nvidia graphics card.

16

I’m no expert at building the Nvidia driver, but at this point I’d probably try downgrading to the 6.11 kernel and seeing if you can get the Nvidia driver to work with that.

To downgrade the kernel, you can use dnf --repo=fedora downgrade kernel\*. You’ll have to manually pick the older kernel when booting your system because it will default to the newest one. You’ll probably have to hold the shift key while booting to get the boot menu to show.

You can use uname -r to verify that you are running the 6.11 kernel before you attempt to run the Nvidia installer.

17

Hi Derek,
So, if you are still getting that compiler mismatch error:
export IGNORE_CC_MISMATCH=“1”
Then try the install / compile again … :slight_smile:

1 Like
18

Thanks once again Gregory for your help. I have tried with your 6.11 kernel suggestion but the nvidia installation stops at 43% with the same error message in /var/log/nvidia-installation.log (warning gcc compilation… but you are using cc…etc.) .

19

Well, I wish my suggestions had been helpful enough to get your driver to build. If there isn’t a problem with the installed packages/dependencies, I think the problem must be with the configuration of your system. You might want to examine your environment variables and make sure there isn’t anything “unusual” set like LD_LIBRARY_PATH or maybe something in your PATH. If you do see something set that might be interfering with the build, you can try fixing the value with export <VAR>=<new-value> and the rerun the installer. Alternatively, you might try creating a whole new user account just for a “clean” environment.

20

Thank you once again.

I’ll be back in my institute on Wednesday and will try out your suggestions.
It may sound strange but my system doesn’t seem to have an “export” command.