GNOME Bugzilla – Bug 786569
Night Light stays orange after overnight suspend
Last modified: 2018-06-06 09:43:32 UTC
Night Light stays orange after waking from sleep. Turing it on or off etc. doesn't fix it. Ubuntu GNOME 17.04 Kernel: 4.10.0-32-generic x86_64 (64 bit gcc: 6.3.0) Gnome 3.24.2 (Gtk 3.22.11-0ubuntu3) Graphics: Card: Advanced Micro Devices [AMD/ATI] Ellesmere [Radeon RX 470/480] Display Server: X.Org 1.19.3 driver GLX Renderer: Gallium 0.4 on AMD POLARIS10 (DRM 3.9.0 / 4.10.0-32-generic, LLVM 4.0.0) GLX Version: 3.0 Mesa 17.0.7 Direct Rendering: Yes This was also reported on launchpad where others have the issue too. https://bugs.launchpad.net/ubuntu/+source/gnome-settings-daemon/+bug/1686895
A workaround for this is to change the display resolution and then revert original resolution.
The Display panel just has a toggle, the code that implements it is in gnome-settings-daemon.
I've noticed the same issue. Night light when set to automatic (not tested manual) will not change back in the morning once the sun is up. I have automatic suspend turned _off_ Night light controls stop working as Veron describes. The workaround does work. -- Fedora 26 Kernel 4.12.5 GNOME 3.24.2
On Fedora 26 with GNOME 3.24.2, I get the same behavior, but the only workaround I've found seems to be restarting. Changing the resolution as suggested above did not work.
I'm experiencing the same issue on Fedora 27 with gnome-settings-daemon-3.26.2-2.fc27.x86_64. The workaround described in https://bugs.launchpad.net/ubuntu/+source/gnome-settings-daemon/+bug/1686895/comments/8 works for me: Open Gnome "Settings", go to "Devices" -> "Color" and toggle the color management for the affected screen(s) off and on again.
Created attachment 369795 [details] [review] color: Cancel old smooth transition on color temperature change A smooth transition between two color temperatures might still be active when the next color temperature is set. Cancel any pending transitions in this case. This possibly fixes issues where the colour temperature is stuck in night mode even though it is daytime.
Created attachment 369796 [details] [review] color: Add test to check that smooth transitions are cancelled
Could someone try whether the patch actually fixes the bug. It definitely fixes a possible race condition that might be causing this, but there could be further issues.
Someone tried attachment 36795 [details] [review] here https://bugzilla.redhat.com/show_bug.cgi?id=1564030 , apparently with success
Yanko, the link goes to some odd attachment. Did you mean attachment 369796 [details] [review] from this bug which cancels the smooth transition? At this point, I believe we do have a second bug that is related to attaching a new monitor while the laptop is suspended. However, I have no clue where it is, a dbus-monitor log from a resume that exhibits the bug might be useful (I have personally never reproduced the issue).
Hmm, interesting, so the user reports that this specific bug was fixed by the patch. I really did not expect that to happen, but that would be really good news.
Sorry for the typo. I meant attachment 369795 [details] [review]
Review of attachment 369795 [details] [review]: ::: plugins/color/gsd-night-light.c @@ +217,3 @@ /* small jump */ if (ABS (temperature - self->cached_temperature) < GSD_TEMPERATURE_MAX_DELTA) { + poll_smooth_destroy (self); Could be more readable if we do this call right above the if() so it catches the big and small jump cases, and replace the one from poll_smooth_create() with a warning/assert checking that the idle is unset. But LGTM.
Review of attachment 369796 [details] [review]: Nice :), LGTM.
Thanks, committed to both master and stable with the changes!
We had a number of issues that were fixed recently. Most notable: * Cancelling of smooth color temperature transitions after an abrupt change happens. This has been reported to fix issues where one monitor has the wrong state after resuming. https://bugzilla.gnome.org/show_bug.cgi?id=786569 * There was an issue in certain mutter versions which caused night light to only work on one monitor https://gitlab.gnome.org/GNOME/mutter/merge_requests/71 I am closing bugs now that look like duplicates of these issues. If you believe you are hitting a different bug, then please open a new issue on gitlab. Please note that some of these fixes are not yet part of a stable release. They will become available in gnome-settings-daemon 3.28.2.
OK, there are reports that the issue is less likely but persists. I opened https://gitlab.gnome.org/GNOME/gnome-settings-daemon/issues/60 to track this.