Gaming Linux® From Scratch (System V Edition) - Version #8869

Chapter 3. Shared Dependencies

NVIDIA-575.51.02

Introduction to NVIDIA

The NVIDIA proprietary driver contains firmware, kernel drivers, userland drivers pertaining to OpenGL, Vulkan, and hardware acceleration, and provides useful utilites for NVIDIA cards.

Additional Downloads

NVIDIA Dependencies

Required

GBM from Mesa-25.0.4, libglvnd-1.7.0, and Xorg Libraries

Recommended

Installation of NVIDIA

[Note]

Note

Make sure to execute the following commands when an Xorg instance isn't running, or else the driver will complain and quit. Alternatively, you can pass --allow-installation-with-running-driver --no-x-check to nvidia-installer to override NVIDIA's warnings. This is not recommended as it skips sanity checks, but is generally safe. If you do this, you should reboot afterwards.

Along with that, be sure to have the kernel source that you have compiled the kernel you are currently using with. Make sure the source tree has not been moved, or else you will have to specify the kernel source tree directory using --kernel-source-path=<path>. If you don't have the source tree on your system, the driver will fail to install.

For a complete list of flags, pass -A.

First enable XFree86 DRI and Nouveau [8] support in the kernel and recompile if necessary. 

Device Drivers --->
  Graphics support --->
    <*/M> Direct Rendering Manager (XFree86 4.1.0 and higher DRI support) --->
                                                                      ...  [DRM]
      <M> Nouveau (NVIDIA) cards                                   [DRM_NOUVEAU]
    Frame buffer Devices --->
      <M>   nVidia Framebuffer Support                               [FB_NVIDIA]
      < /M> Simple framebuffer support                               [FB_SIMPLE]
[Important]

Important

Build these drivers as kernel modules so that they can properly be blacklisted.

Install NVIDIA by running the following command as the root user: 

sh NVIDIA-Linux-x86_64-575.51.02.run -x &&
pushd NVIDIA-Linux-x86_64-575.51.02 &&
  patch -Np1 -i ../nvidia-575.51.02-gcc15.patch \
    -d kernel &&
  patch -Np1 -i ../nvidia-575.51.02-gcc15.patch \
    -d kernel-open &&
  ./nvidia-installer --no-abi-note \
                     --ui=none     \
                     --no-systemd  \
                     --install-compat32-libs &&
  popd &&
rm -rf NVIDIA-Linux-x86_64-575.51.02
[Tip]

Tip

If you aren't booted in the LFS system you are installing GLFS for, then it is recommended to copy the above command to a shell script, then once you are booted into the LFS system, you can run the script. Running the command in a chroot is likely to cause kernel modules to be unable to be built due to a different kernel being used on the host.

And just follow the prompts. It will first ask if you wish to install the NVIDIA Proprietary or MIT/GPL kernel modules. If you have the NVIDIA Grace Hopper or NVIDIA Blackwell cards, you must choose the MIT/GPL option. If you have a Turing, Ampere, Ada Lovelace, or Hopper card, it is recommended to now choose the MIT/GPL option. If you have a Maxwell, Pascal, or Volta card, choose NVIDIA Proprietary. If you are using both an older and newer card, but not the newest cards, go with NVIDIA Proprietary.

[Note]

Note

If you are having trouble installing the driver, it might be because it has issues with the Ncurses library. Pass --silent to the above command for your architecture as it may solve the issue. Also pass -M <kernel_type> where <kernel_type> can be open or proprietary. For the notice above, NVIDIA Proprietary is proprietary and MIT/GPL is open.

You may also encounter an error saying the the driver failed to build a kernel module. This is usually because the kernel version you are using is “too high.” If this happens, it may be necessary to downgrade the kernel to a lower version. When a new driver release happens, the driver will then most likely support the kernel version at the time. At the top of this section, the recommended kernel version to use with this driver is mentioned to ensure that a build failure does not occur. The current kernel version used for kernel parameters and such in this book is linux-6.14.2.

[Important]

Important

When you update the kernel, you will also need to reinstall the NVIDIA driver.

This is because there is glue when kernel modules are built. This glue doesn't matter too much with first party modules (modules in the kernel source tree) as the modules are updated at the same time the kernel is. At that point, there is new glue holding them together. However, when it comes to third party modules, they aren't updated when the kernel updates, so this glue the third party modules had on the kernel breaks. This leads to broken kernel modules. This also applies if you stay on the same kernel version but change some options, this can also break the glue.

Be sure that when you reinstall the driver, make sure you are using the new kernel and have the kernel source you built from. Otherwise the driver will fail to install.

Command Explanations

-x: This switch extracts the runfile.

patch ...: These commands apply patches that force the installation to use -std=gnu17 when compiling kernel modules. This is necessary to allow building with GCC-15.x.x.

--no-abi-note: This parameter removes a mostly useless note about a minimum kernel version for NVIDIA OpenGL libraries.

--ui=none: This disables a TUI menu which can get graphical issues during installation.

--no-systemd: This parameter disables Systemd support.

--install-compat32-libs: This option enables the installation of 32-bit libraries by default instead of being prompted.

Configuring NVIDIA

Config Files

/etc/modprobe.d/nvfb.conf

Configuration Information

Nouveau support was compiled into the kernel to make the NVIDIA driver build. However, Nouveau conflicts with NVIDIA, so it should be blacklisted.

As the root user, create the file /etc/modprobe.d/nvfb.conf to blacklist Nouveau: 

mkdir -pv /etc/modprobe.d &&
cat >> /etc/modprobe.d/nvfb.conf << "EOF"
# Begin /etc/modprobe.d/nvfb.conf

blacklist nvidiafb
blacklist nouveau

# End /etc/modprobe.d/nvfb.conf
EOF

Furthermore, this driver does not install DRI [9] drivers or driver stubs. Likewise, it does not install a pkg-config file saying where to find the associated files. The files aren't needed with this driver. However, some applications expect the pkg-config file to exist, namely Xorg-Server-21.1.16. Create one now as the root user: 

cat > /usr/lib/pkgconfig/dri.pc << "EOF"
prefix=/usr
includedir=${prefix}/include

dridriverdir=/usr/lib/dri

Name: dri
Description: Direct Rendering Infrastructure
Version: 575.51.02
Requires.private: libdrm >=  2.4.109
Cflags: -I${includedir}
EOF

Applications that do require this file typically will not link against any library in /usr/lib/dri. If you are worried that an application will try to link against non-existent libraries, you can compile Mesa-25.0.4 to get these libraries and a real pkg-config file. Note that those libraries will not be used when this driver is in use.

[Note]

Note

Now you should skip to XCB Utils.

Contents

Installed Programs: nvidia-bug-report.sh, nvidia-cuda-mps-control, nvidia-cuda-mps-server, nvidia-debugdump, nvidia-installer, nvidia-modprobe, nvidia-ngx-updater, nvidia-settings, nvidia-smi, nvidia-uninstall, and nvidia-xconfig
Installed Firmware: gsp_ga10x.bin and gsp_tu10x.bin
Installed Libraries: libEGL_nvidia, libGLESv1_CM_nvidia, libGLESv2_nvidia, libGLX_nvidia, libcuda, libcudadebugger, libglxserver_nvidia, libnvcuvid, libnvidia-allocator, libnvidia-api, libnvidia-cfg, libnvidia-egl-gbm, libnvidia-egl-wayland, libnvidia-eglcore, libnvidia-encode, libnvidia-fbc, libnvidia-glcore, libnvidia-glsi, libnvidia-glvkspirv, libnvidia-gpucomp, libnvidia-gtk2, libnvidia-gtk3, libnvidia-ml, libnvidia-ngx, libnvidia-nvvm, libnvidia-opencl, libnvidia-opticalflow, libnvidia-pkcs11-openssl3, libnvidia-pkcs11, libnvidia-ptxjitcompiler, libnvidia-rtcore, libnvidia-tls, libnvidia-wayland-client, libnvoptix, for Wine-10.6 (_nvngx and nvngx; both are DLLs), libvdpau_nvidia, nvidia-drm_gbm, and nvidia_drv
Installed Directories: /usr/lib/firmware/nvidia/575.51.02, /usr/lib/gbm, /usr/lib/modprobe.d, /usr/lib/modules-load.d, /usr/lib/nvidia/wine, /usr/lib/nvidia/xorg, /usr/lib/sysusers.d, /usr/lib/vdpau, /usr/lib/xorg/modules/drivers, /usr/share/X11/xorg.conf.d, /usr/share/egl/egl_external_platform.d, /usr/share/nvidia, /etc/share/vulkan/icd.d, /etc/share/vulkan/implicit_layer.d, and /usr/src/debug/nvidia-utils

Short Descriptions

nvidia-bug-report.sh

is the NVIDIA bug reporting shell script

nvidia-cuda-mps-control

allows multiple CUDA processes to share a single GPU context

nvidia-cuda-mps-server

creates the shared GPU context, manages the clients, and issues work to the GPU on behalf of the clients

nvidia-debugdump

collects the internal GPU state

nvidia-installer

installs, upgrades, and uninstalls the NVIDIA driver (running the runfile directly is recommended)

nvidia-modprobe

creates Linux device files and loads the NVIDIA kernel module

nvidia-ngx-updater

updates NGX features without requiring a full application update

nvidia-settings

a GUI application relying on GTK-3 that allows tweaking settings like resolution and refresh rate

nvidia-smi

provides montioring information for Tesla and select Quadro devices

nvidia-uninstall

installs, upgrades, and uninstalls the NVIDIA driver (running the runfile directly is recommended)

nvidia-xconfig

manipulates X11 configuration files to allow the NVIDIA driver to be used when starting X11

libEGL_nvidia

provides NVIDIA implementations of OpenGL functionality

libGLESv1_CM_nvidia

provides the API entry points for all OpenGL ES and EGL function calls

libGLESv2_nvidia

provides the API entry points for all OpenGL ES and EGL function calls

libGLX_nvidia

provides NVIDIA implementations of OpenGL functionality

libcuda

provides runtime support for CUDA

libcudadebugger

provides support for debugging CUDA applications

libglxserver_nvidia

is the NVIDIA GLX extension module for X11

libnvcuvid

provides an interface to hardware video decoding capabilities

libnvidia-allocator

is used internally by other driver components

libnvidia-api

provides an interface for managing properties of GPUs

libnvidia-cfg

is used internally by other driver components

libnvidia-egl-gbm

provides GBM EGL application support

libnvidia-egl-wayland

provides client-side Wayland EGL application support

libnvidia-eglcore

is used internally by other driver components

libnvidia-encode

provides an interface to video encoder hardware

libnvidia-fbc

provides an interface to capture and optionally encode the framebuffer of an X11 server screen

libnvidia-glcore

is used internally by other driver components

libnvidia-glsi

is used internally by other driver components

libnvidia-glvkspirv

is used internally by other driver components

libnvidia-gpucomp

is used internally by other driver components

libnvidia-gtk2

provides the nvidia-settings user interface

libnvidia-gtk3

provides the nvidia-settings user interface

libnvidia-ml

provides a monitoring and management API

libnvidia-ngx

provides functions for AI features

libnvidia-nvvm

provides JIT link-time-optimization for the CUDA driver

libnvidia-opencl

provides NVIDIA's implementation of the OpenCL API standard

libnvidia-opticalflow

can be used for hardware-accelerated computation of optical flow vectors and stereo disparity values

libnvidia-pkcs11-openssl3

is a cryptography library wrapper aiming to provide cryptography operations when the GPU and driver are operating in Confidential Compute mode

libnvidia-pkcs11

is a cryptography library wrapper aiming to provide cryptography operations when the GPU and driver are operating in Confidential Compute mode

libnvidia-ptxjitcompiler

is a JIT compiler which compiles PTX into GPU machine code and is used by the CUDA driver

libnvidia-rtcore

is used internally by other driver components

libnvidia-tls

provides thread local storage support for the NVIDIA OpenGL libraries

libnvidia-wayland-client

is required for nvidia-settings on Wayland

libnvoptix

implements the OptiX ray tracing engine

libvdpau_nvidia

provides the NVIDIA implementation for the VDPAU library

_nvngx

provides DLSS support for use with Wine

nvngx

provides DLSS support for use with Proton

nvidia-drm_gbm

is the NVIDIA implementation of GBM

nvidia_drv

is the NVIDIA X11 driver


[8] Newer versions of the NVIDIA drivers will fail to compile with TTY support unless a graphics driver is included in the kernel. Nouveau is used here, though alternate graphics drivers may also work.

[9] Direct Rendering Infrastructure.