Open sn4201 opened 5 days ago
Suspend is currently broken in the Nvidia driver.
Suspend is currently broken in the Nvidia driver.
It did work when i first tested it after rebasing to bluefin-stable, so could there be something that I did which caused it to break again? I enabled secure boot and tweaked a couple other things yesterday
Or is it just universally broken for everyone with nvidia?
Is there any workaround? @tulilirockz mentioned possibly using nvidia-open yesterday, is there a way to install that to try it out or is it affected by the same issue?
Thanks
Bluefin currently doesn't build an nvidia-open image.
It's been broken on 565 series driver. The Nvidia GPU doesn't recover from suspend.
So i used a workaround:
sudo rpm-ostree rebase ostree-unverified-registry:ghcr.io/ublue-os/bluefin:stable
Using the non-nvidia release seems to be perfect for my use case since I dont game. The NVK driver seems to work great in terms of suspend/wake. No performance issues in browsers using HW accel video (at least none i've found yet). Bonus is that power consumption on my laptop is even lower now than before! Hope this helps someone else in the future
ujust toggle-nvk
will rebase you properly, you've inadvertantly rebased without confirming the signature.
sudo bootc switch ghcr.io/ublue-os/bluefin:stable --enforce-container-sigpolicy
should get you back on track!
ujust toggle-nvk
will rebase you properly, you've inadvertantly rebased without confirming the signature.
sudo bootc switch ghcr.io/ublue-os/bluefin:stable --enforce-container-sigpolicy
should get you back on track!
Yeah! Its just that they were switching off of Fedora Silverblue, they had to do the unsigned -> signed step thing before that
sudo bootc switch ghcr.io/ublue-os/bluefin:stable --enforce-container-sigpolicy
I tried this command (without first doing ujust toggle-nvk
, and received:
(process:13526): GLib-CRITICAL **: 17:41:34.373: g_atomic_ref_count_dec: assertion 'old_value > 0' failed
(process:13526): GLib-CRITICAL **: 17:41:34.705: g_atomic_ref_count_dec: assertion 'old_value > 0' failed
(process:13526): GLib-CRITICAL **: 17:41:34.705: g_atomic_ref_count_dec: assertion 'old_value > 0' failed
(process:13526): GLib-CRITICAL **: 17:41:34.705: g_atomic_ref_count_dec: assertion 'old_value > 0' failed
(process:13526): GLib-CRITICAL **: 17:41:34.727: g_atomic_ref_count_dec: assertion 'old_value > 0' failed
(process:13526): GLib-CRITICAL **: 17:41:34.729: g_atomic_ref_count_dec: assertion 'old_value > 0' failed
(process:13526): GLib-CRITICAL **: 17:41:34.729: g_atomic_ref_count_dec: assertion 'old_value > 0' failed
(process:13526): GLib-CRITICAL **: 17:41:34.732: g_atomic_ref_count_dec: assertion 'old_value > 0' failed
(process:13526): GLib-CRITICAL **: 17:41:34.745: g_atomic_ref_count_dec: assertion 'old_value > 0' failed
Queued for next boot: ostree-image-signed:docker://ghcr.io/ublue-os/bluefin:stable
Version: 41.20241117.3
Digest: sha256:0062e05c043c6887d4d9f01c7c9f27abf27440b089d31a172372e1baef8ca9f6
Is this expected?
The last message looks fine, the other stuff is just noise.
Describe the bug
Sadly although suspend/wake was working when I first rebased to bluefin, it has since become broken again. Manual suspend seems to just shut off the laptops screen and does not allow it to turn back on, requiring forced power off.
Not sure if maybe enabling secure boot broke it again, or something else? @tulilirockz
What did you expect to happen?
Laptop sleep/wake should put the system to sleep and wake it up when interacted with
Output of
rpm-ostree status
Output of
groups
Extra information or context