mirror of
https://github.com/Ponce/slackbuilds
synced 2024-11-28 10:02:43 +01:00
c2c34c5f56
Signed-off-by: Dave Woodfall <dave@slackbuilds.org> Signed-off-by: Willy Sudiarto Raharjo <willysr@slackbuilds.org> |
||
---|---|---|
.. | ||
doinst.sh | ||
nvidia-driver.info | ||
nvidia-driver.SlackBuild | ||
nvidia-switch | ||
rc.nvidia-persistenced | ||
README | ||
README.nvidia-persistenced | ||
skip_conflict-GPU_detect.patch | ||
slack-desc |
This is the proprietary binary video driver from NVidia for the X.Org X-server. This package does not include the kernel-module - it is part of the nvidia-kernel package at SlackBuilds.org. NOTE: This driver is ONLY for Slackware64. For x86 32-bit support, use nvidia-legacy390-{driver,kernel}. Installing the 32-bit compatibility libraries is available: if desired, do: COMPAT32="yes" ./nvidia-driver.SlackBuild If you plan to test this buildscript on current/15.0, do: CURRENT="yes" ./nvidia-driver.SlackBuild This avoids stomping on the libglvnd libraries, plus it forces the script to use the prebuilt nvidia-installer since it no longer builds in current. As a consequence, this script may no longer work in a VM. This script now gives the option of installing the nvidia-persitenced daemon. For details, see README.nvidia-persitenced. Several useful utilities are included: nvidia-xsettings, nvidia-smi, and nvidia-settings. Please refer to the Nvidia documentation and man pages for details and usage. For CUDA/OpenCL to work after reboot, and for utilites such as nvidia-smi, you might need to include the following line in your /etc/rc.d/rc.local (This is REQUIRED for nvidia-persistenced to work): # Create missing nvidia device nodes after reboot if [ -x /usr/bin/nvidia-modprobe ]; then /usr/bin/nvidia-modprobe -c 0 -u fi Notes: 1. You still need to blacklist the open source nouveau driver. This is easily done by installing xf86-video-nouveau-blacklist in extra/xf86-video-nouveau-blacklist on the distibution DVD. 2. If nvidia-driver is already installed, the script will abort with a prompt to remove the existing driver (it is not necessary to remove the kernel module). 3. As of version 450.xx, Nvidia no longer includes libGL.la in their distribution package. On 14.2 ONLY, nvidia-switch saves these files to /var/log/nvidia/{32,64}: libGL{,ESv1_CM,ESv2}.la libGL.so.1.2.0 libGLESv1_cm.so.1.1.0 libGLESv2.so.2.0.0 It then modifies the .la files in /usr/lib{,64} to point to the nvidia versions of these libraries. If you want to revert back to the open source drivers, run 'nvidia-switch --remove' to restore the original mesa libraries, then 'removepkg nvidia-driver' "should" work in most cases. If you forget to run nvidia-switch --remove, just copy the original files saved in /var/log/nvidia/{32,64} back to /usr/lib{,64} (don't forget to run /sbin/ldconfig). (THIS STEP IS NOT NECESSARY ON CURRENT.) 4. (IMPORTANT ON 14.2) When it comes time to remove this package with 'removepkg' (or just about any other package for that matter), you MUST run '/sbin/ldconfig' to restore any symlinks. This is due to removepkg not running ldconfig after removing a package. (This has been added to current/15.0.)