You are not logged in.
tl;dr certain events seem to trigger what appears to be a display reconnection event
Filed a bug against gnome-shell here: https://gitlab.gnome.org/GNOME/gnome-shell/issues/2075 -- but I'm beginning to suspect I'll need to file a kernel bug report as well
From my description:
> I have an interesting issue, and I'm not necessarily sure if it's related to Shell yet, but I figure I'll start here. I have a new Ice Lake 2-in-1 (XPS 13 7390) and am experiencing what I can best describe as unwanted display re-connection issues.
> That is, without any prompting, the display will go suddenly go black, maybe flicker once, and then come back online as if nothing happened. I would expect this to happen if I plugged in an external monitor, but it happens on the internal display without that.
> However, it's not completely at random. I can generally get the event to reliably happen in a few circumstances, though I can't figure out the commonalities between them.
Has anybody experienced anything similar? Posted three videos in the comments of that report.
Offline
Check your dmesg/journal log, in particular check whether you aren't running into the issue discussed in https://bbs.archlinux.org/viewtopic.php?id=250765
Offline
Check your dmesg/journal log, in particular check whether you aren't running into the issue discussed in https://bbs.archlinux.org/viewtopic.php?id=250765
Yea, stumbled across that earlier. Unfortunately there is nothing in the logs at all when the issue occurs.
I suppose the next step is to test out linux-git.
Offline
Well, linux 5.5-rc4 made no difference. Interestingly, sway does not show the same behavior, so I believe this is somehow related to Shell. I have a sneaking suspicion that this has something to do with input handling, will need to do some more triaging.
Offline
I have the exact same issue on the same laptop but using bspwm and lightdm, so no gnome involved.
At first I thought that was somehow related with pulseaudio as I am able to reproduce this by triggering something audio related (inserting/removing earphones, starting a video or an audio trace). But then it also happens randomly when I don't do anything.
I can confirm that there is nothing related in the logs. The only thing that I can add is that 5.3.6arch1 is not affected, but 5.3.6arch3 is (I did not try 5.3.6arch2, but looking at the commits I suspect it would be affected too).
My temporary solution is downgrading the kernel to 5.3.6arch1 (and also linux-firmware otherwise the wifi is broken).
Offline
I run sway on my 7390 2-in-1 and, as per other comments, I've not experienced these issues myself. However, I've only just recently re-enabled the LPSS (Low Power SubSystem) module as it was broken through 5.3.x. I noticed that some of my input names and devices changed once the i2c (and similar) busses were working. It might, therefore, be an option to try and disable or temporarily blacklist the intel_lpss_pci module and see if it's related to that?
Offline
i know other models that have adaptive brightness control enabled have had similar issues, not sure if yours is effected or not but there's a simple test linked on this wiki
https://wiki.archlinux.org/index.php/De … ss_Control
Offline
I've just tried the link and it doesn't seem to be enabled for me (I didn't see any fade in), nor can I see any options for it being enabled in the "BIOS".
Offline
I have the exact same issue on the same laptop but using bspwm and lightdm, so no gnome involved.
At first I thought that was somehow related with pulseaudio as I am able to reproduce this by triggering something audio related (inserting/removing earphones, starting a video or an audio trace). But then it also happens randomly when I don't do anything.
I can confirm that there is nothing related in the logs. The only thing that I can add is that 5.3.6arch1 is not affected, but 5.3.6arch3 is (I did not try 5.3.6arch2, but looking at the commits I suspect it would be affected too).
My temporary solution is downgrading the kernel to 5.3.6arch1 (and also linux-firmware otherwise the wifi is broken).
that's an interesting finding, could you test arch2? if that does have it, it significantly reduces the problem space. not sure how to get the set of patches that were applied though.
Offline
edoars wrote:I have the exact same issue on the same laptop but using bspwm and lightdm, so no gnome involved.
At first I thought that was somehow related with pulseaudio as I am able to reproduce this by triggering something audio related (inserting/removing earphones, starting a video or an audio trace). But then it also happens randomly when I don't do anything.
I can confirm that there is nothing related in the logs. The only thing that I can add is that 5.3.6arch1 is not affected, but 5.3.6arch3 is (I did not try 5.3.6arch2, but looking at the commits I suspect it would be affected too).
My temporary solution is downgrading the kernel to 5.3.6arch1 (and also linux-firmware otherwise the wifi is broken).
that's an interesting finding, could you test arch2? if that does have it, it significantly reduces the problem space. not sure how to get the set of patches that were applied though.
I'm not seeing a 5.3.6-arch2 or 5.3.6-arch3, did you mean 5.4.6-arch{2,3}?
Offline
[craigcabrey@enterprise ~]$ pacman -Q linux
linux 5.4.6.arch1-1
I can confirm that with this kernel, the issue is no longer present. I used the fix from the other thread (symlinking the wifi firmware) to fix the Wifi. Now that we have a known good state, can start bisecting to find the problem child.
tracking this issue in https://bugs.archlinux.org/task/65025
Last edited by craigcabrey (2020-01-03 05:22:04)
Offline
I'm not seeing a 5.3.6-arch2 or 5.3.6-arch3, did you mean 5.4.6-arch{2,3}?
Yes, sorry for the late response, I meant 5.4.6.
tracking this issue in https://bugs.archlinux.org/task/65025
Thank you
Offline
I have the exact same device and problem on kernel 5.4.8-arch1.
The event is reproducible on certain triggers: Starting gdm, selecting the user in gdm, and starting playback of media in Chromium, as well as intermittent during audio playback.
Offline
I have the exact same device and problem on kernel 5.4.8-arch1.
The event is reproducible on certain triggers: Starting gdm, selecting the user in gdm, and starting playback of media in Chromium, as well as intermittent during audio playback.
There is no fix available yet, check the discussion on the Gitlab issue linked above. For now, you can patch a revert of f6ec9483091f8e67adab0311a4e2f90aab523310 locally (or I can put up my build somewhere, but I imagine most would just prefer to build themselves).
Offline
grinko wrote:I have the exact same device and problem on kernel 5.4.8-arch1.
The event is reproducible on certain triggers: Starting gdm, selecting the user in gdm, and starting playback of media in Chromium, as well as intermittent during audio playback.
There is no fix available yet, check the discussion on the Gitlab issue linked above. For now, you can patch a revert of f6ec9483091f8e67adab0311a4e2f90aab523310 locally (or I can put up my build somewhere, but I imagine most would just prefer to build themselves).
Thank you! I compiled the kernel with the commit reverted and the issue is gone after a power cycle.
Offline
The bug should be resolved in the upcoming kernel version 5.4.12-arch1
Offline
The bug should be resolved in the upcoming kernel version 5.4.12-arch1
It is, but another bug was introduced that causes the system to struggle on video playback and eventually hard freeze, at least on my machine.
I had to revert back to 5.4.8 with the patch applied and the system is stable again.
Offline