Arch Linux User Repository

Yep thanks for solution worked for me too.. I think an update is on the way according to this :
https://github.com/NVIDIA/open-gpu-kernel-modules/commit/811073c51e1f098834e90e67a074f1746b6e2831#diff-32da369dcd8cd223fa47c8b86028f1a7572ed699c33f2057f4702e4b36f13dd5

And
https://forums.developer.nvidia.com/t/nvidia-modules-build-failure-with-upcoming-gcc-14-and-recent-kernels-due-to-misfiring-conftest-sh-test-heads-up/279072

Patch here
https://bbs.archlinux.org/viewtopic.php?pid=2170219#p2170219

For anyone having issues installing, downgrade GCC back to 13.2. GCC 14 breaks the dkms installation for this driver.

Downgrading gcc and gcc-libs to 13.2.1, then re-installing the current 6.8.9-arch1-2 of linux and linux-headers worked for me.

P.S. Thanks grzechoo and everyone for the solve. Not the way I wanted to spend my morning!

Downgrade gcc and gcc-libs to 13.2.1 and go back with your linux and linux-headers packages to 6.8.9-arch1-1. Latest gcc 14.1.1 causes NVIDIA dkms module installation to fail. Just figure that out after an hour of digging on internet and lately upgraded packages...

@pervokur may you provide more details on how to get that kernel version and how to set clang-17 with lto=none?

I'm guessing you installed that kernel manually ... so I supose if it works for 6.9.0 ... it should work for 6.8.x, should we flag this package as out of date?

Module works on 6.9.0-rc6 + clang-17 lto=none

I can't install the moduel for 6.8.9-arch1-2, like @erickwill

but the module either is installed for 6.6.30-2-lts !

Why is that? I've even tried to downgrade to 6.8.9-arch1-1 where I had the module working previously but still doesn't install!

Got an issue here while updating to the latest kernel linux-6.8.9.arch1-2 linux-headers-6.8.9.arch1-2.

==> dkms install --no-depmod nvidia/470.239.06 -k 6.8.9-arch1-2 Error! Bad return status for module build on kernel: 6.8.9-arch1-2 (x86_64) Consult /var/lib/dkms/nvidia/470.239.06/build/make.log for more information. =

here is the make.log:

https://drive.google.com/file/d/1ySTYgyOFMt-MdaNNwbTpKdHRnTgRkZlS/view?usp=drive_link

Please, let me know if I can provide anything else. In this meantime, is there any workaround? Thanks!

I was able to compile kernel modules with clang-16 (my kernel is x86-64 6.7.8 clang-16 full lto)

1)edit conftest.sh change CC=$1 to CC=gcc, because this script isn't working correctly with clang-16

2) IGNORE_CC_MISMATCH=1 for me reported kernel compiler is something like "clang 16.0.6 ld.lld 16.0.6", so this environment variable helps me to pass compiler check

3) -MMD to -MD I wrapped clang-16, so -MMD option will be replaced with -MD. (maybe it's enough to remove '-MD' from Kbuild file, so there will no conflict between -MMD and -MD)

@wviana and others. Warning the procedures also uninstall steam. I have to typically install steam doing

pacman -Sdd steam

however reinstalling steam the same way was also no problem.

because for some time steam complains about lib32 stuff for nvidia, its clearly broken and has been for some months.

I had the same error message. I took these steps. First I removed all.

yay -Rc nvidia-470xx-dkms lib32-nvidia-470xx-utils

Then I simple reinstalled them, since they are not present on the system, no dependency errors, and since they reinstalled simultaneously, there is no problem.

yay -S nvidia-470xx-dkms lib32-nvidia-470xx-utils

If not using a helper like yay, I guess you would do something like.

makepkg -sri

From each directory of the extracted snapshot files. It should work, although I don't do this anymore since it has become increasingly complicated to install these drivers correctly. However if doing the manual way issue 'sudo mkinitcpio -p linux' after installing, that usually solved most problems for me in the past.

There will come a time when our hardware is no longer updated, breakage is to be expected at some point.