No need to apply my new patch, as the result indicates all 3 modules are of 560.28.03. My new (correct) patch fixes the issue for pre-560 series of drivers only. To be honest, I've missingly uploaded from wrong directory where I've backed up the previous patch before fixing the issue.everything is ok ? Do I need to apply your new patch ? It seems that nvidia-drm.ko does not support 14.1-RELEASE,but it works on the 14.0.
kld_list
variable in /etc/rc.conf (nvidia-modeset.ko, nvidia.ko and other dependencies are automatically pulled in) and having hw.nvidiadrm.modeset=1
in /boot/loader.conf.kldxref_enable="YES"
and /etc/rc.d/kldxref does the job.Metaports graphics/drm-kmod and graphics/nvidia-drm-kmod chooses basically the newest supported version for running base. So for 14.1, 61 is selected. This is because newer version of (non-nvidia) DRM driver has usually wider range of supported CPUs (iGPUs). So if your CPU is properly supported by 515, it would work, theoretically.I've compiled the nvidia-drm-515-kmod instead of the nvidia-drm-61-kmod and it has been accepted.
View attachment 20008
Metaports graphics/drm-kmod and graphics/nvidia-drm-kmod chooses basically the newest supported version for running base. So for 14.1, 61 is selected. This is because newer version of (non-nvidia) DRM driver has usually wider range of supported CPUs (iGPUs). So if your CPU is properly supported by 515, it would work, theoretically.
Freebsd 14.1 p3 is out
might be worth checking if the update has any fixes
It is just working here, not Wayland, though.There is no definitive evidence that drm-61 will work on 14.1 unless someone tries your procedure on that system.
% pkg version -v | rg 'nvidia'
linux-nvidia-libs-560.31.02 = up-to-date with index
nvidia-driver-560.31.02 = up-to-date with index
nvidia-drm-61-kmod-560.31.02_1 = up-to-date with index
nvidia-drm-kmod-550.107.02 > succeeds index (index has 550.54.14)
nvidia-settings-535.146.02_1 = up-to-date with index
nvidia-xconfig-525.116.04 = up-to-date with index
% pkg version -v | rg 'drm'
drm-61-kmod-6.1.92 = up-to-date with index
libdrm-2.4.122,1 = up-to-date with index
linux-c7-libdrm-2.4.97 = up-to-date with index
nvidia-drm-61-kmod-560.31.02_1 = up-to-date with index
nvidia-drm-kmod-550.107.02 > succeeds index (index has 550.54.14)
Wayland is rather fragmentedSo what is remaining is someone to try with Wayland.
pkg version -v | rg 'nvidia'
libva-nvidia-driver-0.0.12 = up-to-date with port
linux-nvidia-libs-550.54.14 = up-to-date with port
nvidia-driver-550.54.14 = up-to-date with port
nvidia-drm-515-kmod-550.54.14_1 = up-to-date with port
nvidia-settings-535.146.02_1 = up-to-date with port
wlroots-0.17.4 Modular Wayland compositor library
hw.nvidiadrm.modeset="1"
exec dbus-launch --exit-with-session ck-launch-session startplasma-wayland
qt6-wayland-6.7.2 Qt6 wrapper for Wayland
egl-wayland-1.1.13 EGLStream-based Wayland external platform
/boot/loader.conf
hw.nvidiadrm.modeset="1"
export __NV_PRIME_RENDER_OFFLOAD=1
export __GLX_VENDOR_LIBRARY_NAME=nvidia
export QT_QPA_PLATFORMTHEME=qt6ct
export QT_QPA_PLATFORM=wayland
export LIBSEAT_BACKEND=consolekit2
export QT_QPA_PLATFORM=wayland
then there are the issue getting kde and nvidia to work
1 -what is the command you use to start kde
you have to start kde from a tty because the gui login screen doesnt work with wayland
so what is the correct command to start
dbus
consolekit
startplasma-wayland
Code:exec dbus-launch --exit-with-session ck-launch-session startplasma-wayland
2 - qt6 and wayland packages
what are the qt6 packages needed
Code:qt6-wayland-6.7.2 Qt6 wrapper for Wayland
Code:egl-wayland-1.1.13 EGLStream-based Wayland external platform
3 - hw.nvidiadrm.modeset
Code:/boot/loader.conf
Code:hw.nvidiadrm.modeset="1"
4 - environmental variables
Wayland - ArchWiki
wiki.archlinux.org
Code:export __NV_PRIME_RENDER_OFFLOAD=1 export __GLX_VENDOR_LIBRARY_NAME=nvidia
Code:export QT_QPA_PLATFORMTHEME=qt6ct export QT_QPA_PLATFORM=wayland export LIBSEAT_BACKEND=consolekit2
lots of different permutations to test
i have run qt based applications in an ubuntu jail and displayed the window on the Freebsd host
by setting the following in the ~/.zshenv file in the Ubuntu jail
Code:export QT_QPA_PLATFORM=wayland
my hunch is the running KDE using Wayland is either to do with the way kde is started using the dbus, consolekit and the startplasma-wayland script
or that an environmental variable that needs to be set for Nvidia
might be an idea to post the contents of the startplasma-script to see if that where the issue is
Renaming /usr/lib64/xorg/modules/libglamoregl.so makes by twm-~/.xinitrc-startx working again....
i get you have it working on your other machine which doesnt use nvidiaMan. I've already been able to make work KDE6 and Wayland. I know what are the working parameters. I'm sure that it will work even with an nvidia GPU. Maybe some minor adjustements will be needed. Im pretty sure that it does not work because the bug that I have shown more times. I have even started a discussion on the nvidia forum,exposing that bug,but no one replied. Even on the FreeBSD / KDE ML,no one told anything :
No provider of glEGLImageTargetTexture2DOES found. Requires one of GL_OES_EGL_image
someone wants to help ?forums.developer.nvidia.com
export __EGL_VENDOR_LIBRARY_FILENAMES=/usr/local/share/glvnd/egl_vendor.d/10_nvidia.json eglinfo -B
egl-wayland-1.1.13
i get you have it working on your other machine which doesnt use nvidia
have you seen this thread
from that page ( i changed the path to the freebsd path for the file )
Code:export __EGL_VENDOR_LIBRARY_FILENAMES=/usr/local/share/glvnd/egl_vendor.d/10_nvidia.json eglinfo -B
do you have this installed
Code:egl-wayland-1.1.13
just mention it since your error had egl in it so thought it might be realted
Plasma/Wayland/Nvidia - KDE Community Wiki
community.kde.org
No provider of glEGLImageTargetTexture2DOES found. Requires one of: GL extension "GL_OES_EGL_image"
export __NV_PRIME_RENDER_OFFLOAD=1
export __GLX_VENDOR_LIBRARY_NAME=nvidia
nvidia OptimusThe error is :
Code:No provider of glEGLImageTargetTexture2DOES found. Requires one of: GL extension "GL_OES_EGL_image"
I see that it is related to the libglamoregl.so library (glamor).
hw.nvidiadrm.modeset="1"
NVIDIA GLX implementation
export __NV_PRIME_RENDER_OFFLOAD=1
export __GLX_VENDOR_LIBRARY_NAME=nvidia
nvidia Optimus
Talk:NVIDIA Optimus - ArchWiki
wiki.archlinux.org
The NVIDIA's official page gives a caveat.[1]
Some versions of the “modesetting” driver try to load a sub-module called “glamor”, which conflicts with the NVIDIA GLX implementation.
Please ensure that the libglamoregl.so X module is not installed.
do you have nvidia modeset enable in /boot/loader.conf
Code:hw.nvidiadrm.modeset="1"
makes me think of setting variables for glx
Code:export __NV_PRIME_RENDER_OFFLOAD=1 export __GLX_VENDOR_LIBRARY_NAME=nvidia
hw.nvidiadrm.modeset="1"
export __NV_PRIME_RENDER_OFFLOAD=1
export __GLX_VENDOR_LIBRARY_NAME=nvidia
kld_list="nvidia-modeset nvidia-drm linux linux64 i915kms ext2fs mac_priority fusefs"
thats what im getting atYou told me 1000 times and I replied 1000 times that I started using :
from the beginning. I have also tried these :
Code:export __NV_PRIME_RENDER_OFFLOAD=1 export __GLX_VENDOR_LIBRARY_NAME=nvidia
thanks. What I didn't try has been to remove glamor.
Some versions of the “modesetting” driver try to load a sub-module called “glamor”, which conflicts with the NVIDIA GLX implementation.
hw.nvidiadrm.modeset="0"
kld_list="nvidia-modeset nvidia-drm"