Package Details: nvidia-libgl-full-beta-all 390.12-1

Git Clone URL: https://aur.archlinux.org/nvidia-full-beta-all.git (read-only)
Package Base: nvidia-full-beta-all
Description: NVIDIA driver library symlinks for 'nvidia-utils-full-beta-all'
Upstream URL: http://www.nvidia.com/
Keywords: lib32-opencl-driver lib32-vulkan-driver libglvnd nvidia-settings opencl-driver vulkan-driver
Licenses: custom:NVIDIA
Conflicts: libegl, libgl, libgles, nvidia-libgl
Provides: libegl, libgl, libgles, nvidia-libgl=390.12
Submitter: Det
Maintainer: Det
Last Packager: Det
Votes: 23
Popularity: 1.147393
First Submitted: 2013-06-27 12:48
Last Updated: 2018-01-04 23:30

Required by (592)

Sources (3)

Latest Comments

Det commented on 2018-01-07 21:44

It's the same PKGBUILD for the whole thing, no you can't change it for an individual one, nor do you really want to install the lib32-* ones directly.

Just change the 0 to 1, and it'll update all the packages in the future until you remove them.

cadence commented on 2018-01-07 21:42

Hello,

The PKGBUILD for this package has the _lib32 flag on line 7 set to zero by default. This results in only the non-32bit nvidia packages being installed.

Since this is the lib32 package, this should changed to : "_lib32=1" by default.

Once this flag is set to 1 the lib32 package installs perfectly.

Thank you!

Det commented on 2017-12-08 14:18

Well, my automation takes patches in the form of linux-[kernel version].patch, e.g. linux-4.15.patch

ronoverdrive commented on 2017-12-08 02:36

Thanks Det, this is what I came up with https://pastebin.com/7aAr0U0z but putting my patch file in the same directory as the PKGBUILD gives me this error ==> ERROR: A failure occurred in prepare().

Det commented on 2017-12-08 02:09

$ mv NVIDIA-Linux-x86_64-387.34-no-compat32 a

$ cp -r a b

— change the file(s) in 'b' —

$ diff -ur a b > lol.patch

— apply in the kernel* dir: $ patch -p2 -i lol.patch —

ronoverdrive commented on 2017-12-08 02:00

Ok looks like the problem is located in /kernel/nvidia-drm/nvidia-drm-linux.c and is the same problem the kernel 4.15 guys are having. The work around is to change MODULE_LICENSE("MIT"); to MODULE_LICENSE("GPL"); unfortunately I'm not familiar with making *.patch files to try it.

ronoverdrive commented on 2017-12-08 01:23

With the recent GCC update I'm getting this error on linking:

LD [M] /home/ronoverdrive/nvidia-full-beta-all/src/NVIDIA-Linux-x86_64-387.34/kernel-4.14.4-1-lqx/nvidia-drm.o Building modules, stage 2. MODPOST 4 modules FATAL: modpost: GPL-incompatible module nvidia-drm.ko uses GPL-only symbol 'ex_handler_refcount' make[2]: [scripts/Makefile.modpost:92: __modpost] Error 1 make[1]: [Makefile:1511: modules] Error 2 make[1]: Leaving directory '/usr/lib/modules/4.14.4-1-lqx/build' make: *** [Makefile:84: modules] Error 2 ==> ERROR: A failure occurred in build(). Aborting...

Commander commented on 2017-11-24 19:54

New version:387.34
http://www.nvidia.com/download/driverResults.aspx/127149/en-us

Commander commented on 2017-10-30 17:08

New driver out: 387.22
https://devtalk.nvidia.com/default/topic/1025794

Some nice new updates this time.

Det commented on 2017-10-03 18:12

Real boring these things. They just need improvement.

387.12: https://devtalk.nvidia.com/default/topic/1024719/unix-graphics-announcements-and-news/linux-solaris-and-freebsd-driver-387-12-beta-/

All comments