You are not logged in.
jonathon wrote:Did you build the module into your initramfs?
Ok, that got it, thanks jonathon.
now i'm left to understand why I don't remember having to do that before and why it didn't occur to me to try that.
first time i had to run mkinitcpio -P after an update to nvidia-390xx-dkms too, but what matters is that it works!
Offline
Cant start dGPU (geforce 525m) on my Dell xps 15z (intel igpu) with nvidia-xrun or bumblebee with kernel errors:
[ 34.262159] NVRM: GPU 0000:01:00.0: Failed to copy vbios to system memory.
[ 34.262265] NVRM: GPU 0000:01:00.0: RmInitAdapter failed! (XXXXXXXXXX)
[ 34.262303] NVRM: GPU 0000:01:00.0: rm_init_adapter failed, device minor number 0
with kernel 6.0 patch from BlackIkeEagle.
PS: on kernel 5.19 work well as and lts kernel 5.15.73-1-lts
Last edited by badger_aav (2022-10-13 08:07:40)
Offline
Hi all,
Laptop with Intel and Nvidia gpus.
Just upgraded to kernel 6.0.1 ( 6.0.1-zen2-1-zen #1 ZEN SMP PREEMPT_DYNAMIC Thu, 13 Oct 2022 19:01:21 +0000 x86_64 GNU/Linux), had problems with the driver module not compiling, but was able to apply the patch mentioned above, then saw the AUR package has also been updated, so now using that ( nvidia-390xx-dkms 390.154-2). The system boots now, but Xorg doesn't start. I found the following in dmesg output:
[ 108.021040] NVRM: failed to copy vbios to system memory.
[ 108.021321] NVRM: RmInitAdapter failed! (0x30:0xffff:663)
[ 108.021387] NVRM: rm_init_adapter failed for device bearing minor number 0
The nvidia kernel modules do load, but when Xorg starts, its log files report:
[ 752.629] (II) NVIDIA(0): Creating default Display subsection in Screen section
"nvidia" for depth/fbbpp 24/32
[ 752.629] (==) NVIDIA(0): Depth 24, (==) framebuffer bpp 32
[ 752.629] (==) NVIDIA(0): RGB weight 888
[ 752.629] (==) NVIDIA(0): Default visual is TrueColor
[ 752.629] (==) NVIDIA(0): Using gamma correction (1.0, 1.0, 1.0)
[ 752.629] (**) NVIDIA(0): Option "AllowEmptyInitialConfiguration"
[ 752.630] (**) NVIDIA(0): Option "Coolbits" "28"
[ 752.630] (**) NVIDIA(0): Enabling 2D acceleration
[ 752.678] (EE) NVIDIA(GPU-0): Failed to initialize the NVIDIA GPU at PCI:1:0:0. Please
[ 752.678] (EE) NVIDIA(GPU-0): check your system's kernel log for additional error
[ 752.678] (EE) NVIDIA(GPU-0): messages and refer to Chapter 8: Common Problems in the
[ 752.678] (EE) NVIDIA(GPU-0): README for additional information.
[ 752.678] (EE) NVIDIA(GPU-0): Failed to initialize the NVIDIA graphics device!
[ 752.678] (EE) NVIDIA(0): Failing initialization of X screen 0
[ 752.679] (II) UnloadModule: "nvidia"
[ 752.679] (II) UnloadSubModule: "wfb"
"Check your system's kernel log" is what led me to the messages in the dmesg output.
I managed to recover by removing the grub boot parameters
nvidia-drm.modeset=1 video=VGA-2:d
and removing all nvidia modules from my xorg.conf so that i'm now running Xorg with the Intel graphics card.
Any ideas what's going on?
Thanks!
Offline
Any ideas what's going on?
From NVIDIA-Linux-x86_64-390.154/README.txt
Q. Why does the VBIOS fail to load on my Optimus system?
A. On some notebooks with Optimus graphics, the NVIDIA driver may not be able
to retrieve the Video BIOS due to interactions between the System BIOS and
the Linux kernel's ACPI subsystem. On affected notebooks, applications that
require the GPU will fail, and messages like the following may appear in
the system log:
NVRM: failed to copy vbios to system memory.
NVRM: RmInitAdapter failed! (0x30:0xffffffff:858)
NVRM: rm_init_adapter(0) failed
Such problems are typically beyond the control of the NVIDIA driver, which
relies on proper cooperation of ACPI and the System BIOS to retrieve
important information about the GPU, including the Video BIOS.
Linux 6.0 removed some ACPI functionality [1] so the patch for nvidia-390xx which added linux 6.0 support disabled ACPI support when built for linux 6.0 or newer as a result it lost the functionality it requires to access vbios in some hybrid systems.
Edit:
acpi_dev_for_each_child is only GPL-exported however I belive acpi_get_next_object is not. Either way the ACPI rewrite is none trivial.
[1] https://github.com/torvalds/linux/commi … dc433a0c74
Last edited by loqs (2022-10-15 22:08:05)
Offline
From NVIDIA-Linux-x86_64-390.154/README.txt
Ah, thank you. I remember reading that long ago and immediately forgetting about it. And i saw something about ACPI in that patch, just didn't make the connection.
Linux 6.0 removed some ACPI functionality [1] so the patch for nvidia-390xx which added linux 6.0 support disabled ACPI support when built for linux 6.0 or newer as a result it lost the functionality it requires to access vbios in some hybrid systems.
Edit:
acpi_dev_for_each_child is only GPL-exported however I belive acpi_get_next_object is not. Either way the ACPI rewrite is none trivial.
Guess i'll roll my system back to 5.19.13 for now, then. Is the fix likely to come from the Nvidia-supplied driver, or something in the kernel, do you know?
Thanks!
Offline
s the fix likely to come from the Nvidia-supplied driver, or something in the kernel, do you know?
It will have to come from Nvidia. See https://www.kernel.org/doc/Documentatio … nsense.rst
Offline
Guys, does someone have the clear instruction how to install 390xx AUR package like jonathon's guide to install Manjaro's corresponding driver? Do I need to blacklist something like nouveau along with the installation? On Manjaro I installed that driver according to jonathon's guide correctly, but still can't win that task on Arch. I made package on 6.0.2 Arch, installed it with yay, but glxinfo doesn't show any nvidia card (520MX) on my notebook. Please help and sorry for bothering (I'm new to Linux and Arch).
Last edited by JohnTheRevelator (2022-10-20 12:59:36)
Offline
And on manjaro, do you use a 6.x kernel?
…
glxinfo doesn't show any nvidia card
glxinfo generally doesn't show any cards, but the GL vendor.
Offline
And on manjaro, do you use a 6.x kernel?
…glxinfo doesn't show any nvidia card
glxinfo generally doesn't show any cards, but the GL vendor.
Nope, 5.15.
Offline
"quelle surprise"
Read the previous 5 posts or so and install and use linux-lts
Offline
I think @jonathon updated the nvidia-390xx-dkms package a few days ago to work with kernel 6,0.2. Version is 390.154-2. Before that I had to either use linux-lts or downgrade the kernel 5.19.x.
Never argue with an idiot, they will drag you down to their level and then beat you with experience.
It is better to light a candle than curse the darkness.
A journey of a thousand miles begins with a single step.
Online
"quelle surprise"
Read the previous 5 posts or so and install and use linux-lts
I've installed lts kernel 5.15 something, installed the package and utils. Then I've tried to run nvidia-settings, got the message that xorg.conf had been created, and still no luck, after the reboot I got the following error:
https://ibb.co/1KGwCGd
Do I have to edit some configs before reboot?
Last edited by JohnTheRevelator (2022-10-21 18:19:25)
Offline
That's an IO error from your hard drive and unrelated to any nvidia situation.
Edit: techinically the bus - it doesn't mean that your disk is broken or so.
Remove the xorg.conf, it's utterly unnecessary and nvidia-settings creates a bunch of junk anyway.
Then boot the multi-user.target (2nd link below) and post a complete system journal (1st link below on how to do that from the console)
Last edited by seth (2022-10-21 19:06:48)
Offline
Anyone knows what happened to the nvidia-390xx package? It disappeared from AUR
Offline
https://aur.archlinux.org/packages?O=0&K=nvidia-390xx
You wanted to use nvidia-390xx-dkms anyway.
Offline
I've downgraded my kernel to linux-zen-5.19.13.zen1-1 (and also the linux-zen-headers-5.19.13.zen1-1 package). I then tried to downgrade the nvidia-390xx-dkms package by removing the kernel-6.0 patch from PKGBUILD in nvidia-390xx-utils and re-running makepkg.
It seems to build the modules fine, runs mkinitcpio without errors, but at that point i can't actually install the modules:
sudo modprobe -vv nvidia
modprobe: INFO: custom logging function 0x55a1d19e2af0 registered
insmod /lib/modules/5.19.13-zen1-1-zen/updates/dkms/nvidia.ko.zst
modprobe: INFO: Failed to insert module '/lib/modules/5.19.13-zen1-1-zen/updates/dkms/nvidia.ko.zst': Invalid argument
modprobe: ERROR: could not insert 'nvidia': Invalid argument
modprobe: INFO: context 0x55a1d1f9a450 released
What am i missing?
Offline
Did you also downgrade pahole? 1.23 is needed for the 5.19 kernels.
Offline
That was it, thank you!
Offline
404 - Page Not Found in https://aur.archlinux.org/packages/nvidia-390xx
what happened ?
Offline
404 - Page Not Found in https://aur.archlinux.org/packages/nvidia-390xx
what happened ?
Here's to you: https://aur.archlinux.org/packages?O=0&K=nvidia-390
<49,17,III,I> Fama di loro il mondo esser non lassa;
<50,17,III,I> misericordia e giustizia li sdegna:
<51,17,III,I> non ragioniam di lor, ma guarda e passa.
Offline
Does nvidia-open support these old cards such as the 820M ?
An open source driver seems to be the only viable way forward, otherwise compatibility will always break when a kernel interface updates and the old binary driver that's supposed to be stable and work simply gets obsolete and broken. This is exactly the current situation with this stupid driver.
ATI/AMD has already switched to the open source model for linux, long ago. and nothing bad happened, on the contrary, support, compatibility and even sales have increased for them, not only for desktop usage, also for servers and crypto mining.
I really don't know what's wrong with these NVIDIA people. I think they simply are ingorant, and this ignorance leads to fear, preventing them to advance to a situation that is mumch better not only for us, but for them as well.
What's wrong with their stupid brains ? Don't they understand that open sourcing is the best for everybody ?
Come on people, somebody need to tell them this is ridiculous and that they need to weak up now. Too much damage has been made. Damage for the end users, but also for the people that is supposed to get their hardware working and simply cannot because of this ridiculous blockade.
Offline
Does nvidia-open support these old cards such as the 820M ?
Offline
Anyway, back to the original issue, I found these patches in the debian package:
https://sources.debian.org/src/nvidia-g … n/patches/
Will try them out.
EDIT: the patches apply fine but I get the same acpi error (and error when trying to start the Xsever)
Last edited by canolucas (2022-11-05 18:00:54)
Offline
After pestering Nvidia's tech support by email every few weeks, I finally got a reply this morning:
"We have fixed this issue and fix will be available in next driver release.
We found the kernel patch you were used to workaround the issue is not good. It will crash X. Please don't use "kernel-6.0.patch" .
With NVIDIA's fixed driver the driver compile with kernel 6.0+ successfully and X also starts without any error."
There's also a message to this effect on the Nvidia support forum post i started, the first time anyone from Nvidia has bothered to post in the thread! So fingers crossed, hoping this will be fixed soon. Support for new kernels for this line runs out at the end of the year, though...
Offline
390.157 was released by Nvidia today, and it's working for me with kernel 6.0.9, GeForce GT 630M, running Xorg.
Offline