inttf NVIDIA 340.108 / 390.154 Patcher for Kernel 6.2/6.1/6.0/5.19/5.18/5.17/5.16/5.15/5.14/5.13 [BASH Script] - Comment Page: 4
This is a BASH script, which download NVIDIA installer, extract it, patch it and make new patched installer package. This is very quickly tested alpha version so if you have any problems please let me know or if this works as it should you can also let me know. I use here currently ARCH Linux Patches + my own patches. Currently this works with NVIDIA 340.108 / 390.154 / 418.113 / 435.21 drivers and latest 6.2, 6.1, 6.0, 5.19, 5.18, 5.17, 5.16, 5.15, 5.14, 5.13, 5.12, 5.11, 5.10 and 5.9 kernels.
This should work with any distribution. You just need install wget and patch packages.
Check video version of guide:
Watch More Videos and Subscribe to
if-not-true-then-false.com Youtube...
not working on kernel 5.10.7
Hello asenler,
What Linux distro+Kernel you are using and what version of NVIDIA drivers you are using? What errors you get?
I used the inttf-nvidia-patcher.sh with Ubuntu 20.04.3 LTS & kernel 5.11.0.38
and got the this patch:
NVIDIA-Linux-x86_64-340.108-patched-kernel-5.14.run
is the the 5.14 patch is the correct one for me? I thought I should get something like
NVIDIA-Linux-x86_64-340.108-patched-kernel-5.11.run not the …5.14.run
Am I right???
Update. Ubuntu 20.04 lts and kernel 5.11.0.38
After using the batch I got …. patched-5.14.run instead of …patched-5.11
Anyway I used the patched-5.14.run
After rebooting I got a black screen
I had to uninstall the drivers using
….patched-5.14.run — uninstall
Could you please fix the patch file. Thanks
update2:
I used this:
https://launchpad.net/~kelebek333/+archive/ubuntu/nvidia-legacy
and was able to install the nvidia drivers 340.108
also upgraded my kernel to 5.15.1 and it is working.
Thanks anyway
Hello usama,
Thanks for this, I will check soon kernel 5.15 patches.
I have kernel 5.15.3 so the script doesn’t patch the driver. İf you added the kernel-5.15.patch it will be very good job.
Hello Ali,
Kernel 5.15 patches are available.
I have 5.11 kernel (and only one), when run:
./inttf-nvidia-patcher.sh -v 418.113
I got NVIDIA-Linux-x86_64-418.113-patched-kernel-5.14.run
file.
but this is version 5.14
so what to do?
try to install from this file anyway?
#uname -mrs
Linux 5.11.0-40-generic x86_64
It’s backward compatible at some point, but if it fails, then it’s maybe easiest to drop some patches.
Many thanks for putting this together. I’m still having issues after running this and following the instructions to install NVIDIA drivers. After installing and rebooting, I get the error, “Failed to start Dispatch Pmts to Console Directory Watch” during start up.
One possible issue is that when I run the Patcher, it builds a file NVIDIA-Linux-x86_64-340.108-patched-kernel-5.14.run even though my current kernel is 5.11. Is this potentially causing the issue?
I’m trying to install this on an old Macbook (mid-2010) running Ubuntu with a GeForce 320M.
Any ideas what may be causing the error? I re-installed and reverted back to the Nouveau driver for now.
Hello BR Stoll,
Looks like this is harmless error and depends on dracut. What version of Ubuntu you are using?
I was using Ubuntu 20.04 LTS. However, despite multiple attempts I wasn’t ever able to get the NVIDIA driver working. (The error above was the last message I received while attempting to boot before then the system seemed to hang.
Given the age of my laptop, I’ve decided to give Mint 20.2 (Cinnamon) a try, as this might result in better performance. Not unexpectedly, I’ve seeing the same Nouveau driver errors in Mint that I saw in Ubuntu, so I’m going to try installing the NVIDIA driver via the method above. (Already tried using the standard Driver Manager option, but that left me with a black screen after installing the NVIDIA 340 driver.
Found a few other files that need patching. In the 5.14 kernel on Fedora 34, I still get these errors:
CC [M] /var/lib/dkms/nvidia/390.144/build/nvidia-uvm/uvm_common.o
/var/lib/dkms/nvidia/390.144/build/nvidia/nv-mmap.c: In function ‘nv_encode_caching’:
/var/lib/dkms/nvidia/390.144/build/nvidia/nv-mmap.c:255:16: warning: this statement may fall through [-Wimplicit-fallthrough=]
255 | if (NV_ALLOW_CACHING(memory_type))
| ^
/var/lib/dkms/nvidia/390.144/build/nvidia/nv-mmap.c:257:9: note: here
257 | default:
| ^~~~~~~
CC [M] /var/lib/dkms/nvidia/390.144/build/nvidia-uvm/uvm_linux.o
CC [M] /var/lib/dkms/nvidia/390.144/build/nvidia-uvm/nvstatus.o
CC [M] /var/lib/dkms/nvidia/390.144/build/nvidia-uvm/nvCpuUuid.o
CC [M] /var/lib/dkms/nvidia/390.144/build/nvidia-uvm/uvm8.o
CC [M] /var/lib/dkms/nvidia/390.144/build/nvidia-uvm/uvm8_tools.o
/var/lib/dkms/nvidia/390.144/build/nvidia/nvlink_linux.c: In function ‘nvlink_sleep’:
/var/lib/dkms/nvidia/390.144/build/nvidia/nvlink_linux.c:600:18: error: ‘struct task_struct’ has no member named ‘state’; did you mean ‘__state’?
600 | current->state = TASK_INTERRUPTIBLE;
| ^~~~~
| __state
CC [M] /var/lib/dkms/nvidia/390.144/build/nvidia-uvm/uvm8_global.o
CC [M] /var/lib/dkms/nvidia/390.144/build/nvidia-uvm/uvm8_gpu.o
CC [M] /var/lib/dkms/nvidia/390.144/build/nvidia-uvm/uvm8_gpu_isr.o
CC [M] /var/lib/dkms/nvidia/390.144/build/nvidia-uvm/uvm8_procfs.o
make[2]: *** [scripts/Makefile.build:271: /var/lib/dkms/nvidia/390.144/build/nvidia/nvlink_linux.o] Error 1
make[2]: *** Waiting for unfinished jobs….
CC [M] /var/lib/dkms/nvidia/390.144/build/nvidia-uvm/uvm8_va_space.o
/var/lib/dkms/nvidia/390.144/build/nvidia/os-interface.c: In function ‘os_delay’:
/var/lib/dkms/nvidia/390.144/build/nvidia/os-interface.c:583:18: error: ‘struct task_struct’ has no member named ‘state’; did you mean ‘__state’?
583 | current->state = TASK_INTERRUPTIBLE;
| ^~~~~
| __state
make[2]: *** [scripts/Makefile.build:272: /var/lib/dkms/nvidia/390.144/build/nvidia/os-interface.o] Error 1
make[1]: *** [Makefile:1858: /var/lib/dkms/nvidia/390.144/build] Error 2
make[1]: Leaving directory ‘/usr/src/kernels/5.14.16-201.fc34.x86_64’
make: *** [Makefile:81: modules] Error 2
Hello John,
Did you used patched version of NVIDIA 390.144? It works with 5.14 and 5.15 kernels.
I upgraded kernel to 5.15 then nvidia 340.1098 driver fail. i hope to find a solution
Kernel 5.15 should work now!
I have download sh file today but the file does’t update for kernel-5.15. for this reason installation of driver failed.
Hi Ali,
It have to be cached copy, some reason, what you get. If you try again:
Then run:
You should see:
Yes pathces and script allright I install tke nvidia 340.108 driver with kernel 5.15.10.
But my monitor have 1366*768 but driver set only 1024*768 there is no more option.xorg conf file doesn’t changed this situation.
perhaps nvidia leaves gt300m series card.
Literal hours spent trying to get 340.108 to install on 5.11. 10mins after finding this site its up and running. L E G E N D
Bro can you guide me i have kernel 5.11 and did the exact same steps but it is stuck in black screen showing FAILED .
Check this.
Sounds awesome, thanks!
Debian Sid on kernel 5.15
FANTASTIC.
Excellent Petr!
Good times – bad times – with new installation and new patch for GeForce 340.108 and kernel 5.16 there is some issue – ABI driver error ?
Hello,
Add following to /etc/X11/xorg.conf:
This was needed !
Thank You very much.
It is showing error:
[FAILED] Failed to start Dispatch Pots to Console Directory Watch
Hello Mohit,
Could you tell me a more about your system? Could you post output of following command:
A moment ago I posted a video trying to help with this problem:
Unfortunately didn’t work for me…
But at least I was able to see the screen again, after suspend, then got CPU stuck messages in systemlog… the system almost unstable, just maybe able to open some app like notes… So I was forced to use Intel again… Don’t know what to do, maybe the next Linux kernel will help… Don’t know.
Asus K56CB, Nvidia GeForce GT 740M
Ubuntu 20.04
Tried about 40 different versions include server versions…
Thanks to whoever wrote this!! Fantastic
The patches looks like are only for x86_64 (64 bits) platforms, but I’m running openSuse Tumbleweed 32 bits (old Pentium 4 with a FX Quadro).
Is possible apply the patch to NVIDIA-Linux-x86-340.100.run ?
The original nvidia setup fails with “kernel/nv-linux.h:122:10: fatal error: asm/kmap_types.h: No such file or directory”
Hello MrSparc,
I dropped 32-bit support, when Fedora dropped their 32-bit support. It’s possible to patch 32-bit NVIDIA installer too, but do you find working patches already somewhere?
I downloaded both (x86 and x86_64) 340.108 setups from nvidia website. The 340.108 is the least legacy version for Linux 32-bit that still is available for downloads that supports Quadro FX-1700. Extracted both and did source compare. The source is the same for both architectures,just the compilation target host will be different. So your kernel patches should work. Indeed I modified your bash script for 340.108, changing the url download (just remove _x64 prefix from url and file names), calculated the bsum for the new setups and changed the target used to build the patched setup as x86.
The patched bash works well, downloads, applies the patches and genererates the new setup. I run this in my box and the setup installed without errors.
But after boot looks like lhe black screen of your last video, restarting the X and not being able to switch to terminal because it keeps restarting the X manager.
I have not time right now to continue (Christmas time) but I need to investigate the boot log
JR:” do you find working patches already somewhere?”
If you’re talking about Linux 32bits distro, openSuse Tumbleweed still supports Intel or AMD 32-bit desktops, laptops, and servers (i686) in the Alternative Downloads page.
I’m running this distro with 5.15 kernel.
I found the issue looking at jornal boot log:
“`
This server has a video driver ABI version of 25.2 that this
driver does not officially support. Please check
http://www.nvidia.com/ for driver updates or downgrade to an X
server with a supported driver ABI.
=================================================================
(EE) NVIDIA: Use the -ignoreABI option to override this check.
“`
I need to add a the section in /etc/X11/xorg.conf
“`
Section “ServerFlags”
Option “IgnoreABI” “true”
EndSection
“`
After that it runs flawlessly
NVIDIA Driver Version 340.108
Quadro FX 1700
Linux 32bits kernel 5.15.8-1-pae i686
Distro openSuse Tumbleweed
Thanks JR!