Bug 1398964

Summary: Wayland hang after a few minutes
Product: [Fedora] Fedora Reporter: Jean-Christophe Baptiste <jc>
Component: waylandAssignee: Adam Jackson <ajax>
Status: CLOSED CURRENTRELEASE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 25CC: ajax, nphilipp, whenry
Target Milestone: ---   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2016-12-11 19:10:24 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Attachments:
Description Flags
kernel logs after attaching all screens none

Description Jean-Christophe Baptiste 2016-11-27 16:50:43 UTC
Description of problem:

The Wayland session totally freeze after a few minutes of use.
A hard reboot is then required.

Version-Release number of selected component (if applicable):

Stock Wayland, Fedora 25.

How reproducible:

So far I can reproduce after any session.
I could never get a Wayland session to survive more than 5 minutes.


Steps to Reproduce:
1. Open a Gnome session (Wayland)
2. Do some normal work or just wait doing nothing
3. The mouse cursor will freeze after a while.

Actual results:

Total session freeze. Even VTY cannot be opened.

No error found in dmesg or journalctl.

Comment 1 William Henry 2016-11-28 17:07:34 UTC
I've been seeing the same thing. Looking through journalctl after reboot I don't see anything obvious.  Rebooting into using X11 and everything is fine. I'm in Wayland now and waiting on it to freeze. I have a journalctl -f running to see what is the last thing that happens when it does. 

I will update this if I see that it continues to stay working today. So far so good - I'm surprised 'cause non of my updates seem related.

Comment 2 Jean-Christophe Baptiste 2016-11-28 18:16:07 UTC
I have done the "journalctl -f", nothing occurs in my case.

Comment 3 William Henry 2016-11-28 19:11:26 UTC
FYI I have not had a crash today. I'm not sure what package may have changed to fix this problem. 

I had performed a dnf update late last week but hadn't seen any change. I also performed one today. (6 packages were upgraded) 

I will update if I have further problems.

Comment 4 William Henry 2016-11-28 20:55:46 UTC
Just happened when I started XChat. Nothing suspicious looking in journalctl.

Comment 5 William Henry 2016-11-28 20:58:34 UTC
Related? BZ 1397471?

Comment 6 Jean-Christophe Baptiste 2016-11-30 19:28:14 UTC
No, I deactivated easyscreencast, applied all updates and it still hangs.
Not sure but it seems that displaying a contextual menu (right click) helps triggering the bug.
I could reproduce it several times with Firefox.

Comment 7 Jean-Christophe Baptiste 2016-12-03 09:49:48 UTC
Ok, I could more or less isolate the problem for my case.

In a single or dual monitor setting, the session is perfectly stable.

But it freezes immediately or within minutes whenever I plug in a third screen.

Both external screen of my T460 laptop are connected in HDMI:

- monitor 1 is plugged directly on the HDMI port
- monitor 2 is connected to the Display port through an adapter.

I am still seeing nothing in the logs.

Should I close this bug report and re-open a new one with these more specific information?

Comment 8 Jean-Christophe Baptiste 2016-12-03 09:53:47 UTC
Created attachment 1227578 [details]
kernel logs after attaching all screens

Comment 9 Jean-Christophe Baptiste 2016-12-09 17:46:46 UTC
For some reason, the last kernel update is more verbose.

I finally get an error when it freezes :

déc. 09 18:34:48 localhost.localdomain kernel: [drm:intel_pipe_update_end [i915]] *ERROR* Atomic update failure on pipe A (start=29311 end=29312) time 162 us, min 1043, max 1049, scanline start 1040, end 1051

Comment 10 Jean-Christophe Baptiste 2016-12-11 19:10:24 UTC
I am going to close my monologue even if nobody gives it a shit.

Yesterday's kernel update (4.8.12-300) apparently fixed the issue. I have been using Wayland for almost two day with no crash.