0

It seems that this bug is an old one, and a really persistent one, as has been reported by several users. It was also reported that the bug has been resolved in patch updates, but, somehow, I am still facing the same problem!

Machine: MSI GV62 7RD Processor: Intel® Core™ i7-7700HQ CPU @ 2.80GHz × 8 Graphics: GeForce GTX 1050/PCIe/SSE2 Operating System: Ubuntu 18.10 Kernel: 4.18.0-15-generic

Sorry about the limited details, but I am not Ubuntu expert (not a novice either).

I am filing this bug report as I was asked to file a new bug report.

Also, please note that I have already tried all the workarounds to this problem available on the internet (including disabling the Secure Boot from the BIOS as well completely purging and reinstalling the nvidia-driver-390), but nothing worked till now!

Expecting the Ubuntu team to resolve this soon.

With best regards, Abbas.

Edit: Found the latest driver, but couldn't install it!

After looking at Nvidia's official website, I found out that the latest driver they have release is 410.93. After downloading the .run file, I proceeded as follows:

  1. Purged all Nvidia packages, including nouveau, as well as xorg.
  2. Made the .run file executable.
  3. Stopped the gdm service.
  4. Tried to install it. Got stuck here: enter image description here

  5. Reinstalled all the dependencies of the driver package, as described here.

  6. Repeated steps 3 & 4. Got the same response.

The installer gets stuck at 5% infinitely until the terminal window is forcibly closed.

Any idea what's going on?

Details!

Every time I run the installer (.run file), I get this prompt:

The distribution-provided pre-install script failed! Are you sure you want to continue?

Does this hold any significance?

Found the culprit!

Although I had all the dependencies installed already, as mentioned earlier, something definitely wasn't right! With this thought, I rechecked each of the dependencies through the commands mentioned on the same page. What I found was glibc could not be found, even when I had the latest version installed. I researched the internet and found that I am not the first one facing this issue, and the NVIDIA driver is not the first package making this mistake.

Actually, the NVIDIA driver was looking for glibc in

/lib/

While it was installed in

/lib/x86_64-linux-gnu

So, I solved the problem by creating a simlink, using the following command.

# ln -s /lib/x86_64-linux-gnu/libc-2.13.so /lib/libc.so.6

This allowed the installation to proceed beyond the 5% mark (where it was stuck, as reported in my last update).

A bigger problem!

Well, the driver installation proceeds beyond the 5% mark, but this doesn't lead to a happy ending! The installer concludes with the following message:

ERROR! Unable to load the 'nvidia-drm' kernel module. Installation Failed!

And..... I haven't been able to find a solution to this, yet!

Reply to heynnema's reply #2

I even tried the latest version (418.43) and that also fails with the exact same message.

A good news!

I think it must be reported that I do not see any pkcs #7 messages at the startup anymore. Though I am not sure which of my experiments cured it!

Another attempt

I even tried the latest NVIDIA-Linux-x86_64-418.56.run.

I would also like to mention that I am trying to install the drivers in rescue.target in order to avoid any conflicts with the running applications.

This time, the ERROR: Unable to load the 'nvidia-drm' kernel module message was accompanied with a bit more detail.

[227.245253] CPU4: Core temperature above threshold, cpu clock throttled (total events = 1)
[227.245255] CPU4: Package temperature above threshold, cpu clock throttled (total events = 1)
[227.245258] CPU6: Package temperature above threshold, cpu clock throttled (total events = 1)
[227.245260] CPU5: Package temperature above threshold, cpu clock throttled (total events = 1)
[227.245260] CPU1: Package temperature above threshold, cpu clock throttled (total events = 1)
[227.245261] CPU2: Package temperature above threshold, cpu clock throttled (total events = 1)
[227.245262] CPU7: Package temperature above threshold, cpu clock throttled (total events = 1)
[227.245263] CPU3: Package temperature above threshold, cpu clock throttled (total events = 1)
[227.245264] CPU0: Package temperature above threshold, cpu clock throttled (total events = 1)

It would be really great if someone could help me with this NVIDIA driver installation!

Best Regards, Abbas.

1
ERROR: Unable to load the 'nvidia-drm' kernel module message was accompanied with a bit more detail.

The problem was solved by running the driver-installer (.run) file in rescue.terget, with Secure Boot disabled.

This successfully completed the installation. However, the NVIDIA driver works only with secure boot disabled.

Keeping the Secure Boot permanently disabled definitely leaves the system prone to a lot of other threats and problems!

Link to the NVIDIA forum post.

I appreciate all the help and guidance I received in solving this problem, especially from heynnema. I am really grateful.

This post describes the procedure for signing the driver so that it would work even with Secure Boot enabled. For an in-depth understanding, please refer to this.

  • Unfortunately, the driver doesn't work with Secure Boot enabled. – Abbas Mehdi Mar 29 at 6:38
0

Note: assuming that turning off Secure Boot in your BIOS didn't fix this problem...

In the terminal...

cd /var/log # change to the syslog directory

grep -i pkcs syslog* # search syslog and syslog.1

Notice if it finds pkcs in which log file... syslog... or syslog.1

sudo -H gedit syslog # lets look at syslog

Search for pkcs in the log

Look a few lines before/after the pkcs line, and note which driver is called out

That driver needs to be updated or removed

Update #1:

Feb 20 00:22:57  gdm3[2133]: modprobe: ERROR: could not insert 'nvidia': Operation not permitted 
Feb 20 00:22:57  kernel: [ 81.556911] PKCS#7 signature not signed with a trusted key 
Feb 20 00:22:57  systemd[1]: Stopped /run/user/120 mount wrapper. 
Feb 20 00:22:57  gdm3[2133]: modprobe: ERROR: could not insert 'nvidia': Operation not permitted 
Feb 20 00:22:57  kernel: [ 81.687080] PKCS#7 signature not signed with a trusted key 
Feb 20 00:22:57  gdm3: Child process -2397 was already dead. 

Update #2:

To view the extensive Nvidia installation instructions, paying special attention to the kernel module signing information...

Download the latest Nvidia driver 418.43 which was just released on 2/22/19.

sh ./NVIDIA-Linux-x86_64-418.43.run -x

cd ./NVIDIA-Linux-x86_64-418.43

more README.txt

cd ..

sudo sh ./NVIDIA-Linux-x86_64-418.43.run

Note: be patient when the installer is building the dkms kernel modules

  • It seems I found the culprit. – Abbas Mehdi Feb 19 at 19:01
  • Feb 20 00:22:57 *** gdm3[2133]: modprobe: ERROR: could not insert 'nvidia': Operation not permitted Feb 20 00:22:57 *** kernel: [ 81.556911] PKCS#7 signature not signed with a trusted key Feb 20 00:22:57 *** systemd[1]: Stopped /run/user/120 mount wrapper. Feb 20 00:22:57 *** gdm3[2133]: modprobe: ERROR: could not insert 'nvidia': Operation not permitted Feb 20 00:22:57 *** kernel: [ 81.687080] PKCS#7 signature not signed with a trusted key Feb 20 00:22:57 *** gdm3: Child process -2397 was already dead. – Abbas Mehdi Feb 19 at 19:03
  • Have similar messages in both the files. Which clearly means that the kernel could not identify the nvidia driver to have a valid signature. Now, how to proceed now? – Abbas Mehdi Feb 19 at 19:03
  • @AbbasMehdi I'd first purge all nvidia related software, reboot, and make sure you have a 100% error message free boot. Then research online, I think there's a newer nvidia driver than 390. If there is, download and install it, reboot, and check for PKCS/whatever errors. – heynnema Feb 19 at 19:33
  • @ heynnema Sure I will! – Abbas Mehdi Feb 21 at 20:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.