-
-
Notifications
You must be signed in to change notification settings - Fork 15.4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
mesa_drivers_glvnd: init #34282
mesa_drivers_glvnd: init #34282
Conversation
I feel, better to have Real backend can still be dispatched via |
We discussed this on IRC already:
So not now, sorry. The next step is also drop that |
On Mon, Jan 29, 2018 at 06:45:40AM +0000, gnidorah wrote:
@gnidorah
We discussed this on IRC already:
> vcunat: ... but it's difficult to find time for larger changes (or more risky ones).
> gchristensen: it still requires holistic thinking about how it interacts with the rest of the systems
So not now, sorry. The next step is also drop that `LD_LIBRARY_PATH=/run/opengl-driver/lib` hack, once all drivers are ready to by picked by libglvnd.
I seen discussion, but decide answer here, because I seen it few hours
after it happens.
I would like to drop `LD_LIBRARY_PATH=/run/opengl-driver/lib` hack as
well, but libglvnd still need something like this, to found their
`libGL_$VENDOR.so.0`. But it would be nice, to move this hack to
dedicated variable like LIBGL_VENDOR_PATH.
Anyway, I like to support move into this direction, so I would like to
join efffort, and working on small steps.
One of following steps should be make default aliases `libGL` and
`libGLU` (for mesa-nonglu and mesa respectively), and replace direct
references on mesa everywhere with `libGL`/`libGLU` (this shouldn't
cause any hash changes).
Then we can one time replace `libGL` with `libglvnd` as default libGL
provider. (One reason, why I not tried it yet -- it take 2-3 days to full
rebuild).
… --
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub:
#34282 (comment)
|
I'm definitely interested in replacing direct mesa references everywhere. With that in mind I started testing this. I've applied this patch and done the full rebuild of a Plasma 5 desktop:
xserver starts, and reports the expected GLX, but Plasma 5 does not:
I think I've seen that error before but can't recall what the fix was. That said, the Plasma 5 desktop was working (with HW opengl) prior to this patch and enable of GLVND. Note this config does not use primus. Checking the |
I think the issue with NVidia with glvnd but no optimus is related to this logic here: I was able to restore OpenGL by applying this hack of a patch: |
@avnik Unfortunately I'm not in position to implement that, but I would like to be a tester 😃 And we need someone from NixOS team with interest to support and approve that change. @coreyoconnor Fixed, thank you! |
@gnidorah Thanks! No issues with current commits. |
Motivation for this change
See also #34461
This allows nothing more that co-existing of mesa and proprietary drivers in same /run/opengl-driver/ prefix e.g. for Nvidia Optimus setups. Then correct GL drivers are picked dynamically depending on what X/DRI drivers are used.
cc maintainers @eduarrrd @vcunat
Things done
build-use-sandbox
innix.conf
on non-NixOS)nix-shell -p nox --run "nox-review wip"
./result/bin/
)