-
-
Notifications
You must be signed in to change notification settings - Fork 16.7k
nixos/kmscon: Fix session & gpu clashing #391574
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
base: master
Are you sure you want to change the base?
Conversation
As per fbcon fashion, there's rendering issues. I guess the PR will get much more complicated if this doesn't turn out to be a fluke. |
Ok, this might just be a greetd issue I'm having. Either way, it would be great if someone would test this under "normal" circumstances |
Enabled with GNOME and it seems to work like it used to. Didn’t test extensively, but at least GDM is able to start now. |
Yeah, I suspect the current rev we're in actually works as intended and we actually have to disable DRM specifically. There also is the option to patch kmscon to allow it to free the gpu on demand. However, we would also have to modify every single derivation that starts a graphical session or needs /dev/dri/cardX so it also tells kmscon to free the GPU, just so we use DRM for kmscon (and I don't know any benefits of using DRM over fbcon for kmscon) |
how do i undo this git has played me |
This pull request has been mentioned on NixOS Discourse. There might be relevant details there: https://discourse.nixos.org/t/prs-ready-for-review/3032/5352 |
do we expect multiple seats? |
Don't you get a new "seat" if you switch ttys aka Ctrl+Alt+FX? |
I'm gonna get to do some testing and I'll report back. I guess we do need to set the seat to master, although this should also support multiple seats? I should also make the hwAccel option actually work |
Does getty do multiseat? I don't know if I'm missing anything |
Okay, so switching VT's is way less sluggish now and I can now let kmscon rest again at the master seat used by systemd. Promptly testing for DRM & friends |
what the fuck? what did git just do? |
This is ridicolous why is git co-authoring commits? |
@SuperSandro2000 The issues are fixed on my end. Try to see if they're fixed on yours too. The highlights of the current changes are that it doesn't screw with my system while resting on seat0 anymore and that it's way less sluggish. |
I try to find a bit of time in the coming days to test this. |
6578e32
to
433c615
Compare
what the hell is this CI error |
f959fa4
to
fa0d87a
Compare
Well, a stable version just dropped. |
pkgs/libtsm: 4.0.2-unstable-2023-12-24 -> 4.1.0
Start testing again. Having more than 1 session doesn't seem to give it a stroke anymore. There's not much we can do for DRM yet (well, we could still use the patch that enables kmscon to release the GPU but you'd have to run a wrapper beforehand launching something like Sway when kmscon is running using DRM) |
Apparently VT's are all for seat0 |
I have some module changes i'll upload soon |
Things done
Adjusted the systemd unit for kmscon to fix #385497
This however will completely disable DRM for kmscon and now seats will be unspecified.
If someone absolutely requires DRM for kmscon, let me know.
nix.conf
? (See Nix manual)sandbox = relaxed
sandbox = true
nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD"
. Note: all changes have to be committed, also see nixpkgs-review usage./result/bin/
)Add a 👍 reaction to pull requests you find important.