Rmnscnce/kernel-xanmod

Hi im running the latest xanmod-edge kernel (5.16.2 at the time of writing) and im having issues building the nvidia kernels, using akmods --force i get a fail (log attached below). kernel headers as well as the edge-devel package are installed

https://pastebin.ubuntu.com/p/DdCfN6cDNY/

495 driver is not compatible with 5.16 kernel. you should wait for update for nvidia drivers. Or optionally you can download rpms from rpmfusion rawhide repo and update from command line. Thats a bit work as you have to download a lot of packages

akmod-nvidia-510.39.01-3.fc36.x86_64.rpm
nvidia-settings-510.39.01-1.fc36.x86_64.rpm
xorg-x11-drv-nvidia-cuda-libs-510.39.01-2.fc36.i686.rpm
xorg-x11-drv-nvidia-kmodsrc-510.39.01-2.fc36.x86_64.rpm
kmod-nvidia-510.39.01-3.fc36.x86_64.rpm 
nvidia-xconfig-510.39.01-1.fc36.x86_64.rpm 
xorg-x11-drv-nvidia-cuda-libs-510.39.01-2.fc36.x86_64.rpm 
xorg-x11-drv-nvidia-libs-510.39.01-2.fc36.i686.rpm
nvidia-modprobe-510.39.01-1.fc36.x86_64.rpm  
xorg-x11-drv-nvidia-510.39.01-2.fc36.x86_64.rpm 
xorg-x11-drv-nvidia-devel-510.39.01-2.fc36.i686.rpm  
xorg-x11-drv-nvidia-libs-510.39.01-2.fc36.x86_64.rpm
nvidia-persistenced-510.39.01-1.fc36.x86_64.rpm 
xorg-x11-drv-nvidia-cuda-510.39.01-2.fc36.x86_64.rpm 
xorg-x11-drv-nvidia-devel-510.39.01-2.fc36.x86_64.rpm 
xorg-x11-drv-nvidia-power-510.39.01-2.fc36.x86_64.rpm

and copy them in an new directory and run

 sudo dnf install ./*
1 Like

anyone got amd p-state working on zen2?

Nope, and nobodys answering question why it isn’t working. I asked on xanmod forum and github repo and even here but nobody cares. Althou on the website they advertise that it works

Hi,

The main stable kernel from xanmod debian/ubuntu is not available on copr. Is that intended?
Unfortunately it’s the only one I would like to use.

Thank you.

I use an AMD Zen 2 family processor and it works for me.

> lscpu | grep "Model name" && grep "" /sys/devices/system/cpu/cpufreq/policy0/scaling_driver
Model name:                      AMD Ryzen 5 5500U with Radeon Graphics
amd-pstate

what governor do you use? I’m on schedutil

➜ lscpu | grep "Model name" && grep "" /sys/devices/system/cpu/cpufreq/policy0/scaling_driver
Model name:                      AMD Ryzen 7 4800H with Radeon Graphics
acpi-cpufreq

I’m on ondemand. Maybe that’s why.

> cat /sys/devices/system/cpu/cpufreq/policy0/scaling_governor
ondemand
> cat /sys/devices/system/cpu/cpufreq/policy0/scaling_driver
amd-pstate

I’m not sure where the compatibility problem lies, but there are some struggles getting these kernels to run on Silverblue/ostree: Installation of xanmod kernel fails · Issue #3546 · coreos/rpm-ostree · GitHub. TL;DR logs:

Generating initramfs... done
error: Finalizing kernel: Unexpected / in .vmlinuz.hmac: 1c6da614504a667d4e6d6e0aada248342354aeb6e5e1e5cd34193537c3f2873c8d5251f3c4d3f878d653679b12d70a0bbf3ab324703521b8fd9071c04d3467b5  /lib/modules/5.17.0-xm1.0.fc36.x86_64/vmlinuz

Nevermind, I found the culprit: rpm-ostree/rpmostree-kernel.cxx at 86070c18996fd43c4a2d04012edfe2c9ad15dafb · coreos/rpm-ostree · GitHub

Edit: hmm, it looks like ostree is expecting the shasum to be something along the lines of /boot/vmlinuz-<something>, which is according to the kernel spec. It might be necessary to patch the shasum in the .spec.

Hi, i want to remove the older version of this kernel, i tried to sudo dnf remove (package name) and disable the repository but the kernel is still there. Is there other on how to do it?

sudo dnf remove --oldinstallonly

I am having trouble using Docker CE while running kernel-xanmod-edge. Docker CE runs fine when using the stock Fedora 35 x86_64 kernel in my machine. However, when booting kernel-xanmod-edge, the dockerd service fails to start. When checking the logs, I find messages about failure to setup networking:


abr 05 11:11:17 karlalex-asus kernel: Bridge firewalling registered
abr 05 11:11:17 karlalex-asus audit[16015]: AVC avc:  denied  { write } for  pid=16015 comm="iptables" path="pipe:[55284]" dev="pipefs" ino=55284 scontext=system_u:system_r:iptables_t:s0 tcontext=system_u:system_r:kernel_t:s0 tclass=fif>
abr 05 11:11:17 karlalex-asus audit[16015]: AVC avc:  denied  { write } for  pid=16015 comm="iptables" path="pipe:[55286]" dev="pipefs" ino=55286 scontext=system_u:system_r:iptables_t:s0 tcontext=system_u:system_r:kernel_t:s0 tclass=fif>
abr 05 11:11:17 karlalex-asus dockerd[15975]: time="2022-04-05T11:11:17.321640455-05:00" level=warning msg="Running iptables --wait -t nat -L -n failed with message: `iptables v1.8.7 (legacy): can't initialize iptables table `nat': Bad >
abr 05 11:11:17 karlalex-asus kernel: bpfilter: Loaded bpfilter_umh pid 16016
abr 05 11:11:17 karlalex-asus kernel: bpfilter: write fail -13
abr 05 11:11:17 karlalex-asus kernel: bpfilter: Loaded bpfilter_umh pid 16017
abr 05 11:11:17 karlalex-asus kernel: bpfilter: write fail -13
abr 05 11:11:17 karlalex-asus dockerd[15975]: time="2022-04-05T11:11:17.326370153-05:00" level=info msg="Firewalld: docker zone already exists, returning"
abr 05 11:11:17 karlalex-asus audit[16018]: AVC avc:  denied  { write } for  pid=16018 comm="iptables" path="pipe:[55996]" dev="pipefs" ino=55996 scontext=system_u:system_r:iptables_t:s0 tcontext=system_u:system_r:kernel_t:s0 tclass=fif>
abr 05 11:11:17 karlalex-asus audit[16018]: AVC avc:  denied  { write } for  pid=16018 comm="iptables" path="pipe:[57053]" dev="pipefs" ino=57053 scontext=system_u:system_r:iptables_t:s0 tcontext=system_u:system_r:kernel_t:s0 tclass=fif>
abr 05 11:11:17 karlalex-asus kernel: bpfilter: Loaded bpfilter_umh pid 16019
abr 05 11:11:17 karlalex-asus kernel: bpfilter: write fail -13
abr 05 11:11:17 karlalex-asus kernel: bpfilter: Loaded bpfilter_umh pid 16020
abr 05 11:11:17 karlalex-asus kernel: bpfilter: write fail -13
abr 05 11:11:17 karlalex-asus firewalld[883]: WARNING: COMMAND_FAILED: INVALID_IPV: 'ipv4' is not a valid backend or is unavailable
abr 05 11:11:17 karlalex-asus firewalld[883]: WARNING: COMMAND_FAILED: INVALID_IPV: 'ipv4' is not a valid backend or is unavailable
abr 05 11:11:17 karlalex-asus firewalld[883]: WARNING: COMMAND_FAILED: INVALID_IPV: 'ipv4' is not a valid backend or is unavailable

Running setenforce 0 does nothing to solve the issue.

Hiya, is there a way to make xanmod edge override the fedora kernel in the grub list? cos every boot I need to select the xm1 kernel or it defaults to the fedora kernel, thanks

I get errors on systemd-oomd and nohang on this kernel.
WARNING: PSI metrics are not provided by the kernel: [Errno 2] No such file or directory: '/proc/pressure/memory'
edit: looks like I have to enable psi

Is it possible to setup package symlinks or aliases so that i can remove the package kernel-*. “akmods” requires the default kernel as a dependency rather than kernel-xanmod-edge.

Ideally i would like to just run Xanmod kernels.

use dnf swap command

I’m having trouble running any kernel newer than kernel-xanmod-edge-5.17.5-xm1.0.fc37.x86_64. When I do I the system won’t start the GUI to login. I tried setting the system to multi-user and used systemctl start graphical.target with the exact same results. I’m new to fedora so I’m not sure how to login to kde plasma without a login gui. The commands I’m used to from debian (like startx startplasma-x11) don’t work.

Just unsure why everything is like clockwork on 5.17.5, but 5.17.6, 8 and 9 all gave me issues.

What kinds of hardware are you using (specifically GPU)? Does the issue persist while using the stock Fedora kernel? This may be a kernel regression that has to be reported to the mainline.

Ryzen 7 3700X and GeForce 1660 TI are the CPU and GPU. I need to check if the newer stock kernels work for me and then i’ll report back. Thank you.