Nvidia GeForce 930M not working properly

DragonEggBedrockBreaking

ReillyBrogan I was unable to revert the packages this time and had to reinstall the OS.

WetGeek

DragonEggBedrockBreaking had to reinstall the OS

Been there, and done that before. At least installing a Solus system is fast and easy. Sometimes a fresh start turns out to be a good thing.

DragonEggBedrockBreaking

WetGeek Yes. I’ve reinstalled solus at least a dozen times since making this post. I’ve also had to reinstall other distributions countless times, ranging from easier options like mint and fedora, to more time consuming options like arch and void.

ReillyBrogan

Before you updated to the broken kernel/driver version do you know which versions of the kernel and driver you were on?

DragonEggBedrockBreaking

ReillyBrogan The kernel was something.50 i think, and idk about the driver.

Edit: wait nevermind, I misunderstood, no, I have no idea,

ReillyBrogan

Could you try installing the 470 driver by following the directions in the other post? At the very least that'll help narrow down the issue.

DragonEggBedrockBreaking

ReillyBrogan I have tried the 470 and 390 drivers, neither work.

ReillyBrogan

It's unfortunate that we don't know what version of the kernel and drivers you were on before. That makes it difficult to identify what exactly broke your system, was it a bug in a newer kernel or something to do with the driver update? Or something else unrelated. Do you happen to know roughly when the last time you updated your system was before it broke?

DragonEggBedrockBreaking

ReillyBrogan 7th June

ReillyBrogan

I mean when was the last time you updated your system and it worked afterwards. If you know that then perhaps we can estimate which kernel/driver version was the last working version for you.

ReillyBrogan

Anyway, here are a few more test packages to give a try:

to_install="sudo eopkg it "

# systemd
installed=($(eopkg li | grep systemd | awk '{print  $1 "|"}'))
for i in "systemd" "systemd-32bit" "systemd-32bit-dbginfo" "systemd-32bit-devel" "systemd-dbginfo" "systemd-devel" ; do
  if [[ " ${installed[*]} " =~ " ${i}| " ]]; then
    to_install="$to_install https://solus.reillybrogan.com:8443/systemd-250/$i-250.7-105-1-x86_64.eopkg"
  fi
done

# linux-current
installed=($(eopkg li | grep linux-current | awk '{print  $1 "|"}'))
for i in "linux-current" "linux-current-headers" "linux-current-dbginfo"  ; do
  if [[ " ${installed[*]} " =~ " ${i}| " ]]; then
    to_install="$to_install https://solus.reillybrogan.com:8443/kernel-5.15.50-patched/$i-5.15.50-217-1-x86_64.eopkg"
  fi
done

# nvidia-glx-driver
installed=($(eopkg li | grep nvidia-glx-driver | awk '{print  $1 "|"}'))
for i in "nvidia-glx-driver-32bit" "nvidia-glx-driver-common" "nvidia-glx-driver-current" "nvidia-glx-driver-modaliases" ; do
  if [[ " ${installed[*]} " =~ " ${i}| " ]]; then
    to_install="$to_install https://solus.reillybrogan.com:8443/kernel-5.15.50-patched/$i-515.57-431-1-x86_64.eopkg"
  fi
done

# kmod
installed=($(eopkg li | grep kmod | awk '{print  $1 "|"}'))
for i in "kmod" "kmod-devel" "kmod-dbginfo" ; do
  if [[ " ${installed[*]} " =~ " ${i}| " ]]; then
    to_install="$to_install https://solus.reillybrogan.com:8443/kernel-5.15.50-patched/$i-30-12-1-x86_64.eopkg"
  fi
done

eval $to_install

This adds a few more changes to the linux-current package that may help. I've included rebuilt nvidia-glx-driver packages as well as kmod and systemd packages since I have those on my system and I accidentally built the linux-current package against them. They won't hurt though.

ReillyBrogan

Also, rather than re-installing the system when you aren't able to rollback the packages it would be better to instead follow the instructions on the boot rescue page in order to mount your system and chroot into it. That would at least allow us to extract the journal logs from the failed boot and hopefully shed some more insight into the issue.

And once we know which versions of the kernel and drivers DO work on your system it should be a simple matter to re-build those versions and then you can copy them to a folder on your system. Then you can re-install them directly from that folder while chrooted into your system. That should speed up troubleshooting significantly as well as allow you to just reinstall those packages after a system update. Sure the updater will try to overwrite the packages anytime you do an update after that point but you can always just reinstall the packages after the update and should still be good to go.

ReillyBrogan

Also, if the previous packages don't help what I want you to do is rapidly tap down when the system is booting in order to bring up the boot menu. Hit e after the 5.15.50-217 kernel and then add this to the boot entry: nvidia-drm.modeset=1 (there should be a space before it). This isn't a fix per se but it will force the system to try to use the other nvidia kernel module (nvidia-drm) which might not have whatever the issue is.

« Previous Page