Page: 1, 2  Next

Latest 5.17 wont boot Nvidia
panzerlop
Status: Curious
Joined: 05 Apr 2022
Posts: 7
Reply Quote
usci ccg init failed


Operating System: Kubuntu 22.04
KDE Plasma Version: 5.24.4
KDE Frameworks Version: 5.92.0
Qt Version: 5.15.3
Kernel Version: 5.15.0-25-generic (64-bit)
Graphics Platform: X11
Processors: 12 × AMD Ryzen 5 3600X 6-Core Processor
Memory: 23.4 GiB of RAM
Graphics Processor: NVIDIA GeForce GTX 1660 SUPER/PCIe/SSE2
Back to top
damentz
Status: Assistant
Joined: 09 Sep 2008
Posts: 982
Reply Quote
Nvidia version?

9to5linux.com/nvidia-510-47-03-graphics-driver-released-with-linux-kernel-5-17-and-vulkan-1-3-support

:: Quote ::
Two major changes are included in the NVIDIA 510.47.03 graphics driver that you should be aware of if you own an NVIDIA GPU. The first one is support for the upcoming Linux 5.17 kernel series, and the second one is support for the recently released Vulkan 1.3 graphics API.

Back to top
techAdmin
Status: Site Admin
Joined: 26 Sep 2003
Posts: 4075
Location: East Coast, West Coast? I know it's one of them.
Reply Quote
Also, it's useful to use the right language, the kernel doesn't 'boot' anything but itself. Or better put, the kernel is booted.

Nvidia drivers failing to build on new kernels is a constant, it's always been that way, this is why in general using nvidia non free with latest bleeding edge kernels is not a very good idea. Nvidia tends to support latest kernel major version pretty well, but if you are on the bleeding edge, you are inviting failure. As damentz notes, in this case, they have 5.17 support out in latest nvidia, but that doesn't do you any good if you are using the 470 or 390 legacy drivers.

Showing: inxi -bxxz

is useful since it shows the actually relevant and meaningful data.

Note that if you are using a frozen pool distro like ubuntu, you can't really use latest kernels if you are also using non free nvidia drivers in a practical sense, the distro packaged drivers are made to work with the distro packaged kernels. Why this is so difficult for people to understand has always somewhat mystified me.
Back to top
panzerlop
Status: Curious
Joined: 05 Apr 2022
Posts: 7
Reply Quote
It boots fine with the stock 5.15.3 Kernal and it's driver 510.60.02
Back to top
panzerlop
Status: Curious
Joined: 05 Apr 2022
Posts: 7
Reply Quote
:: techAdmin wrote ::
Also, it's useful to use the right language, the kernel doesn't 'boot' anything but itself. Or better put, the kernel is booted.

Nvidia drivers failing to build on new kernels is a constant, it's always been that way, this is why in general using nvidia non free with latest bleeding edge kernels is not a very good idea. Nvidia tends to support latest kernel major version pretty well, but if you are on the bleeding edge, you are inviting failure. As damentz notes, in this case, they have 5.17 support out in latest nvidia, but that doesn't do you any good if you are using the 470 or 390 legacy drivers.

Showing: inxi -bxxz

is useful since it shows the actually relevant and meaningful data.

Note that if you are using a frozen pool distro like ubuntu, you can't really use latest kernels if you are also using non free nvidia drivers in a practical sense, the distro packaged drivers are made to work with the distro packaged kernels. Why this is so difficult for people to understand has always somewhat mystified me.


People don't understand it because most people don't even know how to compile a kernel dude...

It was working fine for me on the old Ubuntu LTS, Mint ( based off Ubuntu LTS ) it's only since the latest version on 22.04 beta the error occurred. The only feedback I have is what I wrote nothing else happens when I try to boot with it.

I really like the performance from this project and came to just try and help, I'm not sure if I misread you but , I'm just giving feedback bro, if you could give some steps on inxi -bxz ( run when? From normal boot with stock or ? ) then i';d be very happy to update with it.
Back to top
techAdmin
Status: Site Admin
Joined: 26 Sep 2003
Posts: 4075
Location: East Coast, West Coast? I know it's one of them.
Reply Quote
Non free nvidia driver has to support the kernel. Every kernel's api and abi changes in unpredictable ways, native free modules are updated along with the kernel so they work together more or less by default. Non free modules do not, the company, nvidia, in this case, has to release updated drivers that support the latest kernels, sometimes their drivers work with new kernels, sometimes they don't, there is no science to it, if they work, they work, if they don't, they don't. On more rolling release distros like Arch, if a driver can be patched to run on a new kernel, it will work, but if can't be patched, it won't, this stuff is not magic, it's code interacting with code.

inxi -bxxz is a command, you run it, and paste it into a code block here. One you paste it, it's possible to roughly see what's going on with your system without having to ask you for every detail one bit at a time, which gets old and tedious after the 1000th time or so doing it.

If you don't understand why a non free kernel module like nvida will break routinely on cutting edge kernels then you should probably stick to your distro frozen pool kernels. If you don't know what a non free vs a free kernel module are then you should definitely not be using one on a cutting edge kernel, you should stick to your frozen pool kernel and driver package, and be happy.

It's up to you to learn this stuff, not damentz or liquorix, or, do yourself a big favor, and switch to a graphics device that uses a free driver, then you don't have to deal with this stuff, and can more or less use any kernel you want, within reason.

Basically you are trying to have your cake and eat it too, that is, you want the stability of a frozen pool distro AND you want a rolliing release cutting edge kernel AND you want to use a non free graphics driver. That doesn't work in the real world, you can do 2 of those things, but not all 3. In other words, you can usually mix frozen pool distro with cutting edge kernel if you use a free driver, OR you can use a rolling release distro with cutting edge kernel with a non free driver [but this will fail if nvidia or other non free modules fail to add support for new kernels, which happens sometimes, I've seen 3-6 month delays in the past], OR you can use a frozen pool distro with a frozen kernel and a frozen non free driver.

Only the 1st and 3rd options are totally reliable as a general rule, option 2 will often but not always work, though it usually does, but it's not guaranteed.

Note that if you reported a failure with the free drivers and current Liquorix, that would be a significant issue, but a failure of a frozen pool distro with nvidia and current liquorix is an expected event, so very little is to be gained or learned by reporting such a failure, other than a specific nvidia driver on a specific kernel on your specific hardware failed, which isn't really very interesting since that happens all the time.
Back to top
panzerlop
Status: Curious
Joined: 05 Apr 2022
Posts: 7
Reply Quote
:: Code ::
inxi -bxz
System:
  Kernel: 5.15.0-25-generic x86_64 bits: 64 compiler: gcc v: 11.2.0
    Desktop: KDE Plasma 5.24.4 Distro: Ubuntu 22.04 (Jammy Jellyfish)
Machine:
  Type: Desktop Mobo: ASUSTeK model: ROG STRIX B450-F GAMING v: Rev 1.xx
    serial: <superuser required> UEFI: American Megatrends v: 4602
    date: 08/17/2021
CPU:
  Info: 6-core AMD Ryzen 5 3600X [MT MCP] arch: Zen 2 speed (MHz): avg: 2649
    min/max: 2200/4505
Graphics:
  Device-1: NVIDIA TU116 [GeForce GTX 1660 SUPER] vendor: ASUSTeK
    driver: nvidia v: 510.60.02 bus-ID: 08:00.0
  Display: x11 server: X.Org v: 1.21.1.3 driver: X: loaded: nvidia
    gpu: nvidia resolution: 1440x1080
  OpenGL: renderer: NVIDIA GeForce GTX 1660 SUPER/PCIe/SSE2
    v: 4.6.0 NVIDIA 510.60.02 direct render: Yes
Network:
  Device-1: Intel I211 Gigabit Network vendor: ASUSTeK driver: igb v: kernel
    port: e000 bus-ID: 03:00.0
Drives:
  Local Storage: total: 2.73 TiB used: 1.25 TiB (45.9%)
Info:
  Processes: 468 Uptime: 13h 36m Memory: 23.4 GiB used: 3.08 GiB (13.2%)
  Init: systemd runlevel: 5 Compilers: gcc: 11.2.0 Packages: 2315 Shell: Bash
  v: 5.1.16 inxi: 3.3.13


Also may I add to your last statements, I don't WANT to have to use non-free, I simply can't afford to buy an AMD GPU and this computer was a gift. No need to be so aggressive man I signed up because I simply hoped the feedback would help and I like the project.
Back to top
damentz
Status: Assistant
Joined: 09 Sep 2008
Posts: 982
Reply Quote
Ok, now that we have system information, it appears you are running a driver that should be compatible with kernel 5.17.
:: Code ::
OpenGL: renderer: NVIDIA GeForce GTX 1660 SUPER/PCIe/SSE2
v: 4.6.0 NVIDIA 510.60.02 direct render: Yes


So now that we've ruled that out, I googled usci ccg init failed, and found that the i2c_nvidia_gpu module is at fault: askubuntu.com/questions/1278399/dual-system-ubuntu-20-04nvidia-gpu-i2c-timeout-error-ucsi-ccg-i2c-transfer
bugzilla.kernel.org/show_bug.cgi?id=206653

It appears to be unresolved with no patch and you need to blacklist the module, then reboot:
echo 'blacklist i2c_nvidia_gpu' | sudo tee /etc/modprobe.d/blacklist_i2c-nvidia-gpu.conf

Now, in regards to the conversation so far, most open source communities work on the concept of help me, help you. Or in other words, if I need help, I try to exhaust as many things as I reasonably can before asking for help and showing my work so far.

Your post opens up as if you saw some text on your screen and the first thing in your mind was to post here instead of doing a basic search of what you saw. I don't know if that's what you did but it gives the sense that you're wasting our time, regardless of intention.

I appreciate that Liquorix is working out for you (or was up to this point), but the next time you submit an issue, just be mindful of whether the post you write feels like it come off as low effort.
Back to top
panzerlop
Status: Curious
Joined: 05 Apr 2022
Posts: 7
Reply Quote
:: damentz wrote ::
Ok, now that we have system information, it appears you are running a driver that should be compatible with kernel 5.17.
:: Code ::
OpenGL: renderer: NVIDIA GeForce GTX 1660 SUPER/PCIe/SSE2
v: 4.6.0 NVIDIA 510.60.02 direct render: Yes


So now that we've ruled that out, I googled usci ccg init failed, and found that the i2c_nvidia_gpu module is at fault: https://askubuntu.com/questions/1278399/dual-system-ubuntu-20-04nvidia-gpu-i2c-timeout-error-ucsi-ccg-i2c-transfer
https://bugzilla.kernel.org/show_bug.cgi?id=206653

It appears to be unresolved with no patch and you need to blacklist the module, then reboot:
echo 'blacklist i2c_nvidia_gpu' | sudo tee /etc/modprobe.d/blacklist_i2c-nvidia-gpu.conf

Now, in regards to the conversation so far, most open source communities work on the concept of help me, help you. Or in other words, if I need help, I try to exhaust as many things as I reasonably can before asking for help and showing my work so far.

Your post opens up as if you saw some text on your screen and the first thing in your mind was to post here instead of doing a basic search of what you saw. I don't know if that's what you did but it gives the sense that you're wasting our time, regardless of intention.

I appreciate that Liquorix is working out for you (or was up to this point), but the next time you submit an issue, just be mindful of whether the post you write feels like it come off as low effort.


Noted; I will try to add as much info as possible in future, I tried the command to blacklist the i2c but sadly the same message persists, I took a picture with my phone so I could write the full output that I get on the screen, it is black with the following lines in the top left:

nvidia-gpu 0000:08:00.3: ic2 timeout error e0000000
ucsi_ccg 0-0008: i2c_transfer failed -110
ucsi_ccg 0-0008: ucsi_ccg_init failed -110

After that it just hangs until I restart,

I also tried the safemode option from grub of the 5.17 and that let me boot in ( where i did the command again) but on restart same again. If you have any other ideas please let me know , if not thanks for trying anyway

< Edited by panzerlop :: Apr 6, 22, 5:51 >

Back to top
damentz
Status: Assistant
Joined: 09 Sep 2008
Posts: 982
Reply Quote
Hmm, maybe the module is still being loaded some how?

Looking online, it appears that System76 adds these lines depending on the type of GPU you have: github.com/pop-os/system76-power/blob/master/src/graphics.rs#L21

Can you add alias i2c_nvidia_gpu off to the blacklist file, this may be the missing piece.

Also, I saw some mentions that you may need to run sudo update-initramfs -u afterwards for it to work properly, before rebooting.
Back to top
Display posts from previous:   
Page: 1, 2  Next
All times are GMT - 8 Hours