After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 774557 - gnome-shell freezes frequently with 3 monitors (wayland)
gnome-shell freezes frequently with 3 monitors (wayland)
Status: RESOLVED FIXED
Product: mutter
Classification: Core
Component: wayland
3.22.x
Other Linux
: Normal normal
: ---
Assigned To: mutter-maint
mutter-maint
https://fedoraproject.org/wiki/Common...
: 772449 (view as bug list)
Depends on:
Blocks: WaylandRelated
 
 
Reported: 2016-11-16 15:23 UTC by Kamil Páral
Modified: 2017-05-29 08:02 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
MetaRendererNative: Flush all pending swap notifies on idle (3.77 KB, patch)
2016-12-05 18:52 UTC, Rui Matos
none Details | Review
MetaRendererNative: Flush all pending swap notifies on idle (3.35 KB, patch)
2016-12-06 16:43 UTC, Rui Matos
committed Details | Review

Description Kamil Páral 2016-11-16 15:23:10 UTC
There seems to be a race condition which makes gnome-shell completely freeze from time to time, but only if you have 3 monitors. If you have 1 or 2 monitors, this problem never occurs. The freeze means that all rendering stops - I see a still picture, not even mouse cursor can be moved. No keyboard shortcuts work, not even Ctrl+Alt+Fx. But the system is not completely frozen, for example I can still hear music from youtube videos.

Once this happens, usually I'm able to regain keyboard control using sysrq, switch to a text TTY, and reboot. (I must not try to restart gdm service, because that end up in an endless display-flickering loop.)

There is absolutely nothing printed in the journal when this happens. The last time it hit me (today), I've had this line in the journal, which timestamp roughly matches:

Nov 16 15:58:40 dryad kernel: [drm:intel_pipe_update_end [i915]] *ERROR* Atomic update failure on pipe C (start=26158 end=26159) time 2 us, min 1073, max 1079, scanline start 1088, end 1080

However, I don't recall seeing anything similar previously, so this might not be related.

I have no idea how to trigger the problem, but it occurs quite frequently, easily once per day. It doesn't seem to map to any particular action - I've seen it during workspaces switching, but also when doing nothing at all, and even during logging in (when I was nearly logged in and apps started to populate the system tray, it froze).

We have seen this problem occur frequently on two different systems, for two different users (my colleague). The laptop hardware is the same - ThinkPad T450, the monitor layout is different (he doesn't have a rotated monitor, one of his monitors has a smaller resolution).

We are very certain that once we disconnect the third monitor, the problem stops occurring (never happened). It only happens with 3 monitors.

My monitor setup:
$ xrandr 
Screen 0: minimum 320 x 200, current 4920 x 1920, maximum 8192 x 8192
XWAYLAND0 connected 1920x1080+1920+0 510mm x 290mm
   1920x1080     59.96*+
XWAYLAND1 connected 1920x1080+0+0 310mm x 170mm
   1920x1080     59.96*+
XWAYLAND2 connected 1080x1920+3840+0 510mm x 290mm
   1080x1920     59.96*+


gnome-session-wayland-session-3.22.1-2.fc25.x86_64
gnome-shell-3.22.1-2.fc25.x86_64
libwayland-client-1.12.0-1.fc25.x86_64
libwayland-cursor-1.12.0-1.fc25.x86_64
libwayland-server-1.12.0-1.fc25.x86_64
mesa-libwayland-egl-12.0.3-3.fc25.x86_64
mutter-3.22.1-8.fc25.x86_64
xorg-x11-server-Xwayland-1.19.0-0.8.rc2.fc25.x86_64
Comment 1 Kamil Páral 2016-11-16 15:38:20 UTC
Correction, we've seen this on 3 different T450s machines for 3 different people.
Comment 2 Rostislav Svoboda 2016-11-28 08:01:03 UTC
Hi, I see this issue too, I have T440s + 2x NEC monitor (23"+24"). I have Fedora 25.

Any estimation, guess what is the problem?
Comment 3 Rostislav Svoboda 2016-11-28 09:34:57 UTC
"freezes occasionally" - I would say my desktop freezes within 5 minutes.
I had to unplug one monitor :(
Comment 4 Jonas Ådahl 2016-11-28 10:17:55 UTC
Could anyone who can reproduce the issue try out 3.23.2? Sadly there is no RPM build available yet for Rawhide because compilation issus on x86.
Comment 5 Kamil Páral 2016-11-28 10:33:19 UTC
Once there is an RPM, I can definitely try it, just please let me know.
Comment 6 Brandon 2016-11-28 17:00:11 UTC
I am seeing something very similar with my Lenovo X1 Yoga, however I am only using two monitors (built in and an external 30").

As soon as a plug in my 30" the system freezes.  If i do it at GDM then GDM sits a loops and flashes like crazy.  Mine also freezes nearly immediately.
Comment 7 Kamil Páral 2016-11-29 08:59:17 UTC
We have another report from a colleague with Thinkpad T540p with 3 monitors.

Brandon, your case is probably a different bug. This does not freeze immediately, but from time to time randomly (can be once, can be several times per day). I think you should file it separately.
Comment 8 Vincent Maurin 2016-11-29 09:33:10 UTC
Same for me with a Lenovo X1 Carbon using the laptop screen + an docking station with two external screens. Nothing relevant in the journal. Using a Xorg session solved the issue, I haven't tried to unplug one screen
Comment 9 vtsorakidis 2016-11-29 21:42:48 UTC
Same here. 3 monitors plugged and system crashes. I have noticed though that this issue came up with the latest system update I received couple hours ago. I was working fine before that. Is there a way to restore to my previous state?
Comment 10 Jonas Ådahl 2016-11-30 08:52:51 UTC
The current workaround is to login to the X11 session instead of the Wayland session.
Comment 11 Rui Matos 2016-12-01 23:55:42 UTC
Can those on F25 try the rpm in http://koji.fedoraproject.org/koji/taskinfo?taskID=16700675 and report back if they still reproduce this or, if not, if there's any noticeable performance impact?
Comment 12 Mauri Sahlberg 2016-12-02 06:50:42 UTC
(In reply to Rui Matos from comment #11)
> Can those on F25 try the rpm in
> http://koji.fedoraproject.org/koji/taskinfo?taskID=16700675 and report back
> if they still reproduce this or, if not, if there's any noticeable
> performance impact?

I have HP EliteBook 8740w and after upgrading from FC23 to FC25 the 3rd display almost instantly froze the machine when the third display was accessed. For example moving chromium to the 3rd display and opening a menu froze the machine every time. Another example, moving x2go to third display instantly froze the machine.

Installed the mutter packaged from your link. Now I'm writing this on the 3rd display. I don't know about the performance though. Sorry. This is work laptop and I'm a "developer" and nothing I do requires top notch performance related to the display. 

So far, I'm happy with your rpm.
Comment 13 vtsorakidis 2016-12-02 07:12:10 UTC
I am running fedora 25 under gnome xorg and everything works fine. I won't give it a try for now because this is my working environment and my ass is on fire. So far there are plenty of bugs and especially on graphics. Gnome tweak tool extension "List windows at the bottom" doesn't work well also. Filezilla gui also the same. In my opinion this version should be still as beta.
Comment 14 Vincent Maurin 2016-12-02 11:02:22 UTC
(In reply to Rui Matos from comment #11)
> Can those on F25 try the rpm in
> http://koji.fedoraproject.org/koji/taskinfo?taskID=16700675 and report back
> if they still reproduce this or, if not, if there's any noticeable
> performance impact?

I am testing the RPM since 4-5 hours and so far no freeze. Not noticeable performance impact
Comment 15 Leo Ufimtsev 2016-12-05 15:43:43 UTC
(In reply to Rui Matos from comment #11)
> Can those on F25 try the rpm in
> http://koji.fedoraproject.org/koji/taskinfo?taskID=16700675 and report back
> if they still reproduce this or, if not, if there's any noticeable
> performance impact?

I would like to test this. How do I install/rollback the koji package above?

(I have this issue. Freeze up occurs sometimes within 5 minutes, sometimes once a day. MacBook Pro 2016: Fedora 25. Main Screen + 2 side screens). @Gnome 3.22.2
Comment 16 Vincent Maurin 2016-12-05 17:06:42 UTC
(In reply to Leo Ufimtsev from comment #15)
> 
> I would like to test this. How do I install/rollback the koji package above?
> 

On my side, I have downloaded the main rpm ( https://kojipkgs.fedoraproject.org//work/tasks/675/16700675/mutter-3.22.2-2.fc25.x86_64.rpm ) and run this command with root permission :
dnf install mutter-3.22.2-2.fc25.x86_64.rpm

After few days of use, still no freezes, performances are ok (no so smooth, but hard to compare)
Comment 17 Rui Matos 2016-12-05 18:52:18 UTC
Created attachment 341423 [details] [review]
MetaRendererNative: Flush all pending swap notifies on idle

We need to do swap notifications asynchronously from flip events since
these might be processed during swap buffers if we are waiting for the
previous frame's flip to continue with the current.

This means that we might have more than one swap notification queued
to be delivered when the idle handler runs. In that case we must
deliver all notifications for which we've already seen a flip event.

Failing to do so means that if a new frame, that only swaps buffers on
such a swap notification backlogged Onscreen, is started, when later
we get its flip event, we'd notify only an old frame which would hit
this MetaStageNative's frame_cb() early exit:

  if (global_frame_counter <= presented_frame_counter)
    return;

and we'd never finish the new frame and thus clutter's master clock
would be waiting forever stuck.
--

Think I finally understood the problem and this patch should fix it
while not delaying the start of new frames.

I'd like to ask those on F25 to try this scratch build:

http://koji.fedoraproject.org/koji/taskinfo?taskID=16760532
Comment 18 Rui Matos 2016-12-05 18:57:23 UTC
*** Bug 772449 has been marked as a duplicate of this bug. ***
Comment 19 Rui Matos 2016-12-05 18:57:30 UTC
*** Bug 775284 has been marked as a duplicate of this bug. ***
Comment 20 Michael Mejias 2016-12-06 03:56:11 UTC
(In reply to Rui Matos from comment #17)
> Created attachment 341423 [details] [review] [review]
> MetaRendererNative: Flush all pending swap notifies on idle
> 
> We need to do swap notifications asynchronously from flip events since
> these might be processed during swap buffers if we are waiting for the
> previous frame's flip to continue with the current.
> 
> This means that we might have more than one swap notification queued
> to be delivered when the idle handler runs. In that case we must
> deliver all notifications for which we've already seen a flip event.
> 
> Failing to do so means that if a new frame, that only swaps buffers on
> such a swap notification backlogged Onscreen, is started, when later
> we get its flip event, we'd notify only an old frame which would hit
> this MetaStageNative's frame_cb() early exit:
> 
>   if (global_frame_counter <= presented_frame_counter)
>     return;
> 
> and we'd never finish the new frame and thus clutter's master clock
> would be waiting forever stuck.
> --
> 
> Think I finally understood the problem and this patch should fix it
> while not delaying the start of new frames.
> 
> I'd like to ask those on F25 to try this scratch build:
> 
> http://koji.fedoraproject.org/koji/taskinfo?taskID=16760532


Installed as request and no freezing so far. I have a Dell E6440 with docking station and two external HP L2105tm monitors hooked up. One via VGA and the other via DVI. 

Prior to installing this the screen would freeze and the underlying system was still operational as mentioned by others.
Comment 21 Jonas Ådahl 2016-12-06 04:37:59 UTC
Review of attachment 341423 [details] [review]:

::: src/backends/native/meta-renderer-native.c
@@ +136,3 @@
+              cogl_object_unref (info);
+            }
+          while (info != onscreen_native->last_flipped);

Trying to spell out what the code does, since I had a hard time to follow it easily.

This part will invoke the frame callback with frames up to (and including) the one that has received its flip callback.

meta_onscreen_native_queue_swap_notify() is changed to detect when its called recursively from meta_onscreen_native_swap_buffers_with_damage() to set the 'most recently flipped' frame info while avoiding the one being processed a few frames up.

meta_onscreen_native_swap_buffers_with_damage() is changed to allow *_queue_swap_notify() to know its called recursively, more or less.

It believe we could fix this without detecting recursive calls. What do you think about this: ?

1. Add a "pending_queue_swap_notify_frame_count" to MetaOnscreenNative (next to pending_queue_swap_notify on the master branch)

2. Set pending_queue_swap_notify_frame_count to renderer_native->frame_counter in the end of meta_onscreen_native_flip_crtcs()  (on the master branch next to setting pending_queue_swap_notify) to mark the frame count that has now been swapped.

3. Add a "pending_swap_notify_frame_count" to MetaOnscreenNative next to pending_swap_notify.

4. In meta_onscreen_native_queue_swap_notify() set pending_swap_notify_frame_count to queued_pending_swap_notify_frame_count.

5. In flush_pending_swap_notify() invoke the frame callback with frames with global_frame_counter <= onscreen_native->pending_swap_notify_frame_count.

This means we carry the pending state around in the same way as we carry the related state, without relying on check-if-im-being-called-recursively kind of thing that took quite an effort to understand.
Comment 22 Rui Matos 2016-12-06 16:43:54 UTC
Created attachment 341495 [details] [review]
MetaRendererNative: Flush all pending swap notifies on idle

--

(In reply to Jonas Ådahl from comment #21)
> This means we carry the pending state around in the same way as we carry the
> related state, without relying on check-if-im-being-called-recursively kind
> of thing that took quite an effort to understand.

Yeah, this is much cleaner, sorry for the previous spaghetti solution.
Comment 23 Jonas Ådahl 2016-12-07 01:47:26 UTC
Review of attachment 341495 [details] [review]:

Looks good to me.

coding style nits: the long lines (far beyond 80 chars) stand out in this file.

::: src/backends/native/meta-renderer-native.c
@@ +132,2 @@
+          while ((info = (CoglFrameInfo *) g_queue_peek_head (&onscreen->pending_frame_infos)) &&
+                 info->global_frame_counter <= onscreen_native->pending_swap_notify_frame_count)

No need for the "(CoglFrameInfo *)" casting thing there, g_queue_peek_head() returns a gpointer.
Comment 24 Kamil Páral 2016-12-07 08:30:09 UTC
(In reply to Rui Matos from comment #17)
> I'd like to ask those on F25 to try this scratch build:
> 
> http://koji.fedoraproject.org/koji/taskinfo?taskID=16760532

So far no problem in a 1-2 days of usage, 3 different people in our office.
Comment 25 Rui Matos 2016-12-07 15:52:13 UTC
(In reply to Jonas Ådahl from comment #23)
> coding style nits: the long lines (far beyond 80 chars) stand out in this
> file.

Fixed the one case going above 100 chars. I don't think keeping lines
under 80 chars is reasonable with our average variable and type name
lengths...

> ::: src/backends/native/meta-renderer-native.c
> @@ +132,2 @@
> +          while ((info = (CoglFrameInfo *) g_queue_peek_head
> (&onscreen->pending_frame_infos)) &&
> +                 info->global_frame_counter <=
> onscreen_native->pending_swap_notify_frame_count)
>
> No need for the "(CoglFrameInfo *)" casting thing there, g_queue_peek_head()
> returns a gpointer.

Indeed, dropped it.

   a943c0f..497a94f  gnome-3-22 -> gnome-3-22
   71077d5..cfafb0b  master -> master

Attachment 341495 [details] pushed as cfafb0b - MetaRendererNative: Flush all pending swap notifies on idle
Comment 26 jackieb 2016-12-17 04:23:02 UTC
So this is fixed in mutter 3.23.3 of 2016-12-15

I can confirm freeze for 3-displays as soon as windows are placed on external monitors for Ubuntu 16.10 and GNOME3 Staging repo:
c89 4.8.0-30-generic #32-Ubuntu SMP Fri Dec 2 03:43:27 UTC 2016
2016-12-17
Ubuntu 16.10
yakkety
Package: gnome-session
Version: 3.22.2-1ubuntu1~ubuntu16.10.1
Package: mutter
Version: 3.22.2-1ubuntu1~ubuntu16.10.1
Package: libwayland-server0
Version: 1.11.0-2
Package: lightdm
Version: 1.19.5-0ubuntu1
Comment 27 jackieb 2016-12-19 22:58:54 UTC
Could we have a deb of mutter 3.23.3 for 16.10, please?

either at
https://packages.debian.org/sid/mutter
or at
https://launchpad.net/~gnome3-team/+archive/ubuntu/gnome3-staging?field.series_filter=yakkety

I tried to compile the 17.04 3.22.2 package with upgraded source on 16.10, but it failed. It could take 7 weeks for the regulars to compile it.
Comment 28 Florian Müllner 2016-12-19 23:43:45 UTC
We don't distribute .deb (or .rpm, .pkg.tar.xz, <insert-your-favorite-package-format>) packages upstream. You can ask your distributor to include the patches from the gnome-3-22 branch in the package until we release a 3.22.3 tarball upstream.
Comment 29 jackieb 2016-12-20 08:00:50 UTC
Working for the GNOME now ;)

b/c Ubuntu uses -Werror=format-security, please check-in this fix:

diff ../mutter-3.23.3-exported/src/backends/meta-egl.c src/backends/meta-egl.c 
145c145
<                error_str);
---
>                "%s", error_str);

or it won’t compile on Ubuntu. I’ma test my franken-package
Comment 30 Jeremy Bicha 2016-12-20 14:01:11 UTC
Any chance you'll do a mutter 3.22.3 release soon? The 3.22.2 release was broken since distros had to cherry-pick 5c46094d if they wanted GNOME Shell to work on X.

Thanks!
Comment 31 jackieb 2016-12-20 18:05:28 UTC
It segfaults:
Dec 20 09:38:26 c89 kernel: [314682.865029] gnome-shell[24173]: segfault at 18 ip 00007ff09748c33c sp 00007ffd6f2ad600 error 4 in libmutter.so.0.0.0[7ff097427000+125000]
Dec 20 09:38:27 c89 org.gnome.Shell.desktop[24173]: (EE)
Dec 20 09:38:27 c89 org.gnome.Shell.desktop[24173]: Fatal server error:
Dec 20 09:38:27 c89 org.gnome.Shell.desktop[24173]: (EE) failed to dispatch Wayland events: Connection reset by peer
Dec 20 09:38:27 c89 org.gnome.Shell.desktop[24173]: (EE)
Dec 20 09:38:27 c89 gnome-session-binary[23853]: Unrecoverable failure in required component org.gnome.Shell.desktop
Dec 20 09:38:27 c89 gnome-session-binary[23853]: WARNING: Application 'org.gnome.Shell.desktop' killed by signal 11

game over
Comment 32 jackieb 2016-12-20 18:43:05 UTC
it works with mutter version 3.22.2-2ubuntu1~ubuntu16.10.1
that was just built at
http://ppa.launchpad.net/gnome3-team/gnome3-staging/ubuntu
Comment 33 jackieb 2016-12-20 18:54:32 UTC
This bug is still present:
https://bugs.launchpad.net/ubuntu/+source/mutter/+bug/1647084

if you drag LibreOffice or a Virtual Machine window from a low DPI onto a HiDPI display, gnome shell freezes
Comment 34 jackieb 2016-12-20 19:19:09 UTC
Closing pop-under dialogs now freezes, too
https://bugs.launchpad.net/ubuntu/+source/libreoffice/+bug/1649307
Comment 35 jackieb 2016-12-24 17:28:10 UTC
I have uptime of 4 days so GNOME Staging for Ubuntu can be run crash-free on multiple-display (I cannot paste the versions b/c clipboard is broken.)

I don’t understand how you can make the wayland session default, b/c this is not ready for users.
c. The clipboard is broken
b. The two freezes listed above
c. A large number of bugs related to LibreOffice

The claim that the clipboard was fixed in 3.22 is a false one. The clipboard now longer crashes GNOME Shell in 3.22, but it is seriously broken.

- It frequently does nothing in cross-app paste or even in-app paste
- Paste frequently takes a minute or more, in which the destination app is frozen
- It seems sometimes the paste is succeeds in app-1 if a subsequent paste is carried out in another app
- Paste from Android Studio or Visual Code Studio has a Unicode translation problem: every character becomes two characters. It seems paste via gedit is the get-around

It will be necessary to back-port clipboard fixes to 3.22. It is not currently useful.
Comment 36 jackieb 2016-12-24 19:04:38 UTC
And the repeated character bug is still present, so

if you do cut in visual studio code, it cuts many times eating all sorts of source text

the getaround is to instead do copy followed by delete
Comment 37 Moises Bites 2017-03-22 12:50:15 UTC
Hello Guys,

I have this problem many times in these days. Fedora Core 25. 

I have a computer with 3 monitors, and a laptop with a monitor pluged. So, the system freeze totaly many times, with wayland or with xorg.

My gnome, 
$ cat /usr/share/gnome/gnome-version.xml
<?xml version="1.0" encoding="UTF-8"?>
<gnome-version>
 <platform>3</platform>
 <minor>22</minor>
 <micro>2</micro>
 <distributor>GNOME.Org</distributor>
 <date>2017-01-19</date>
</gnome-version>

$  uname -r -v 
4.9.14-200.fc25.x86_64 #1 SMP Mon Mar 13 19:26:40 UTC 2017

$ lshw -short
H/W path         Device      Class          Description
=======================================================
                             system         POS-EIQ87CY (1502969)
/0                           bus            POS-EIQ87CY
/0/0                         memory         64KiB BIOS
/0/4                         memory         256KiB L1 cache
/0/5                         memory         1MiB L2 cache
/0/6                         memory         6MiB L3 cache
/0/7                         memory         8GiB System Memory
/0/7/0                       memory         DIMM [empty]
/0/7/1                       memory         DIMM [empty]
/0/7/2                       memory         4GiB DIMM DDR3 Synchronous 1600 MHz (0,6 ns)
/0/7/3                       memory         4GiB DIMM DDR3 Synchronous 1600 MHz (0,6 ns)
/0/3d                        processor      Intel(R) Core(TM) i5-4570 CPU @ 3.20GHz
/0/100                       bridge         4th Gen Core Processor DRAM Controller
/0/100/2                     display        Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics Controller
/0/100/3                     multimedia     Xeon E3-1200 v3/4th Gen Core Processor HD Audio Controller
/0/100/14                    bus            8 Series/C220 Series Chipset Family USB xHCI
/0/100/14/0      usb3        bus            xHCI Host Controller
/0/100/14/0/1                input          USB Optical Mouse
/0/100/14/0/2                input          USB Keyboard
/0/100/14/0/d                storage        FreeAgent Go
/0/100/14/1      usb4        bus            xHCI Host Controller
/0/100/16                    communication  8 Series/C220 Series Chipset Family MEI Controller #1
/0/100/16.3                  communication  8 Series/C220 Series Chipset Family KT Controller
/0/100/19        eno1        network        Ethernet Connection I217-LM
/0/100/1a                    bus            8 Series/C220 Series Chipset Family USB EHCI #2
/0/100/1a/1      usb1        bus            EHCI Host Controller
/0/100/1a/1/1                bus            USB hub
/0/100/1b                    multimedia     8 Series/C220 Series Chipset High Definition Audio Controller
/0/100/1c                    bridge         8 Series/C220 Series Chipset Family PCI Express Root Port #1
/0/100/1c.2                  bridge         8 Series/C220 Series Chipset Family PCI Express Root Port #3
/0/100/1c.2/0                bridge         IT8893E PCIe to PCI Bridge
/0/100/1d                    bus            8 Series/C220 Series Chipset Family USB EHCI #1
/0/100/1d/1      usb2        bus            EHCI Host Controller
/0/100/1d/1/1                bus            USB hub
/0/100/1d/1/1/4              generic        uTrust 2700 R Smart Card Reader
/0/100/1f                    bridge         Q87 Express LPC Controller
/0/100/1f.2                  storage        8 Series/C220 Series Chipset Family 6-port SATA Controller 1 [AHCI mode]
/0/100/1f.3                  bus            8 Series/C220 Series Chipset Family SMBus Controller
/0/1             scsi0       storage        
/0/1/0.0.0       /dev/sda    disk           128GB MZMPC128HBFU-000
/0/1/0.0.0/1     /dev/sda1   volume         1GiB EXT4 volume
/0/1/0.0.0/2     /dev/sda2   volume         118GiB Linux LVM Physical Volume partition
/0/2             scsi2       storage        
/0/2/0.0.0       /dev/cdrom  disk           DVDRAM GH24NSC0
/0/3             scsi5       storage        
/0/3/0.0.0       /dev/sdb    disk           250GB FreeAgent Go
/0/3/0.0.0/1     /dev/sdb1   volume         232GiB Linux filesystem partition


Always freeze after some minutes, when I had started VirtualBox and the lock screen system start. Many times, the system freeze when I using Google Chrome.
Comment 38 Jonas Ådahl 2017-03-23 04:05:14 UTC
Are there anything interesting showing up in the logs? Both the session journal and kernel logs.
Comment 39 Moises Bites 2017-03-23 13:02:45 UTC
(In reply to Jonas Ådahl from comment #38)
> Are there anything interesting showing up in the logs? Both the session
> journal and kernel logs.

I think that these ones are the records at the moment of freezing today:

52 chrome: [2572:2605:0323/074652.562490:ERROR:socket_stream.cc(210)] Closing stream with result -100
07:46:50 systemd-logind: Power key pressed.
07:46:50 kernel: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=fprintd comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
07:46:40 gnome-settings-: Error setting property 'PowerSaveMode' on interface org.gnome.Mutter.DisplayConfig: O tempo limite foi alcançado (g-io-error-quark, 24)
07:46:24 gdm-x-session: (II) input device 'Power Button', /dev/input/event1 is a keyboard
07:46:24 gdm-x-session: (II) input device 'Power Button', /dev/input/event1 is tagged by udev as: Keyboard
07:46:24 gdm-x-session: (II) systemd-logind: got resume for 13:65
07:46:24 gdm-x-session: (II) input device '        USB Keyboard', /dev/input/event5 is a keyboard
07:46:24 gdm-x-session: (II) input device '        USB Keyboard', /dev/input/event5 is tagged by udev as: Keyboard
07:46:24 gdm-x-session: (II) systemd-logind: got resume for 13:69
07:46:24 gdm-x-session: (II) input device '        USB Keyboard', /dev/input/event3 is a keyboard
07:46:24 gdm-x-session: (II) input device '        USB Keyboard', /dev/input/event3 is tagged by udev as: Keyboard
07:46:24 gdm-x-session: (II) systemd-logind: got resume for 13:67
07:46:24 gdm-x-session: (II) input device 'Power Button', /dev/input/event0 is a keyboard
07:46:24 gdm-x-session: (II) input device 'Power Button', /dev/input/event0 is tagged by udev as: Keyboard
07:46:24 gdm-x-session: (II) systemd-logind: got resume for 13:64
07:46:24 gdm-x-session: (II) input device 'PixArt USB Optical Mouse', /dev/input/event2 is a pointer caps
07:46:24 gdm-x-session: (II) Device 'PixArt USB Optical Mouse' set to 1000 DPI
07:46:24 gdm-x-session: (II) input device 'PixArt USB Optical Mouse', /dev/input/event2 is tagged by udev as: Mouse
07:46:24 gdm-x-session: (II) systemd-logind: got resume for 13:66
07:46:24 gdm-x-session: (II) input device 'Video Bus', /dev/input/event4 is a keyboard
07:46:24 gdm-x-session: (II) input device 'Video Bus', /dev/input/event4 is tagged by udev as: Keyboard
07:46:24 gdm-x-session: (II) systemd-logind: got resume for 13:68
07:46:24 gdm-x-session: (II) intel(0): switch to mode 1600x900@60.0 on VGA1 using pipe 2, position (0, 0), rotation normal, reflection none
07:46:24 gdm-x-session: (II) intel(0): switch to mode 1600x900@60.0 on HDMI3 using pipe 1, position (3200, 0), rotation normal, reflection none
07:46:24 gdm-x-session: (II) intel(0): switch to mode 1600x900@60.0 on HDMI2 using pipe 0, position (1600, 0), rotation normal, reflection none
07:46:24 gdm-x-session: (II) AIGLX: Resuming AIGLX clients after VT switch
07:46:24 gdm-x-session: (II) systemd-logind: got resume for 226:0
07:46:24 kernel: CRED_REFR pid=27545 uid=0 auid=1000 ses=2 subj=system_u:system_r:xdm_t:s0-s0:c0.c1023 msg='op=PAM:setcred grantors=pam_unix,pam_gnome_keyring acct="user" exe="/usr/libexec/gdm-session-worker" hostname=usermachine addr=? terminal=/dev/tty1 res=success'
07:46:24 kernel: USER_ACCT pid=27545 uid=0 auid=1000 ses=2 subj=system_u:system_r:xdm_t:s0-s0:c0.c1023 msg='op=PAM:accounting grantors=pam_unix,pam_localuser acct="user" exe="/usr/libexec/gdm-session-worker" hostname=usermachine addr=? terminal=/dev/tty1 res=success'
07:46:24 kernel: USER_AUTH pid=27545 uid=0 auid=1000 ses=2 subj=system_u:system_r:xdm_t:s0-s0:c0.c1023 msg='op=PAM:authentication grantors=pam_unix,pam_gnome_keyring acct="user" exe="/usr/libexec/gdm-session-worker" hostname=usermachine addr=? terminal=/dev/tty1 res=success'
07:46:24 gdm-session-wor: gkr-pam: unlocked login keyring
07:46:20 systemd: Started Fingerprint Authentication Daemon.
07:46:20 kernel: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=fprintd comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
07:46:20 dbus-daemon: [system] Successfully activated service 'net.reactivated.Fprint'
07:46:20 systemd: Starting Fingerprint Authentication Daemon...
07:46:20 dbus-daemon: [system] Activating via systemd: service name='net.reactivated.Fprint' unit='fprintd.service' requested by ':1.19' (uid=42 pid=1231 comm="/usr/bin/gnome-shell " label="system_u:system_r:xdm_t:s0-s0:c0.c1023")
07:46:17 gnome-settings-: failed to set screen _ICC_PROFILE: Falha ao abrir arquivo "/home/user/.local/share/icc/edid-124a3fc627aa9c1dd15dc66cd0d16108.icc": Permissão negada
07:46:16 gdm-x-session: (II) systemd-logind: got pause for 13:65
07:46:16 gdm-x-session: (II) systemd-logind: got pause for 13:69
07:46:16 gdm-x-session: (II) systemd-logind: got pause for 13:67
07:46:16 gdm-x-session: (II) systemd-logind: got pause for 13:64
07:46:16 gdm-x-session: (II) systemd-logind: got pause for 13:66
07:46:16 gdm-x-session: (II) systemd-logind: got pause for 13:68
07:46:16 gdm-x-session: (II) systemd-logind: got pause for 226:0
07:46:15 gdm-x-session: (II) AIGLX: Suspending AIGLX clients for VT switch
07:46:15 gdm-x-session: (**) Option "fd" "27"
07:46:15 gdm-x-session: (**) Option "fd" "27"
07:46:15 gdm-x-session: (**) Option "fd" "26"
07:46:15 gdm-x-session: (**) Option "fd" "21"
07:46:15 gdm-x-session: (**) Option "fd" "25"
07:46:15 gdm-x-session: (**) Option "fd" "45"
07:46:15 gdm-x-session: (**) Option "fd" "28"
07:46:15 gdm-x-session: (II) input device 'Power Button', /dev/input/event1 is a keyboard
07:46:15 gdm-x-session: (II) input device 'Power Button', /dev/input/event1 is tagged by udev as: Keyboard
07:46:15 gdm-x-session: (II) systemd-logind: got resume for 13:65
07:46:15 gdm-x-session: (II) input device '        USB Keyboard', /dev/input/event5 is a keyboard
07:46:15 gdm-x-session: (II) input device '        USB Keyboard', /dev/input/event5 is tagged by udev as: Keyboard
07:46:15 gdm-x-session: (II) systemd-logind: got resume for 13:69
07:46:15 gdm-x-session: (II) input device '        USB Keyboard', /dev/input/event3 is a keyboard
07:46:15 gdm-x-session: (II) input device '        USB Keyboard', /dev/input/event3 is tagged by udev as: Keyboard
07:46:15 gdm-x-session: (II) systemd-logind: got resume for 13:67
07:46:15 gdm-x-session: (II) input device 'Power Button', /dev/input/event0 is a keyboard
07:46:15 gdm-x-session: (II) input device 'Power Button', /dev/input/event0 is tagged by udev as: Keyboard
07:46:15 gdm-x-session: (II) systemd-logind: got resume for 13:64
07:46:15 gdm-x-session: (II) input device 'PixArt USB Optical Mouse', /dev/input/event2 is a pointer caps
07:46:15 gdm-x-session: (II) Device 'PixArt USB Optical Mouse' set to 1000 DPI
07:46:15 gdm-x-session: (II) input device 'PixArt USB Optical Mouse', /dev/input/event2 is tagged by udev as: Mouse
07:46:15 gdm-x-session: (II) systemd-logind: got resume for 13:66
07:46:15 gdm-x-session: (II) input device 'Video Bus', /dev/input/event4 is a keyboard
07:46:15 gdm-x-session: (II) input device 'Video Bus', /dev/input/event4 is tagged by udev as: Keyboard
07:46:15 gdm-x-session: (II) systemd-logind: got resume for 13:68
07:46:15 gdm-x-session: (II) intel(0): switch to mode 1600x900@60.0 on VGA1 using pipe 2, position (0, 0), rotation normal, reflection none
07:46:15 gdm-x-session: (II) intel(0): switch to mode 1600x900@60.0 on HDMI3 using pipe 1, position (3200, 0), rotation normal, reflection none
07:46:14 gdm-x-session: (II) intel(0): switch to mode 1600x900@60.0 on HDMI2 using pipe 0, position (1600, 0), rotation normal, reflection none
07:46:14 gdm-x-session: (II) AIGLX: Resuming AIGLX clients after VT switch
07:46:14 gdm-x-session: (II) systemd-logind: got resume for 226:0
07:46:12 gnome-shell: g_object_unref: assertion 'G_IS_OBJECT (object)' failed
07:46:12 gnome-shell: g_dbus_connection_signal_unsubscribe: assertion 'G_IS_DBUS_CONNECTION (connection)' failed
07:46:12 gnome-shell: g_dbus_connection_signal_unsubscribe: assertion 'G_IS_DBUS_CONNECTION (connection)' failed
07:46:11 gnome-settings-: failed to set screen _ICC_PROFILE: Falha ao abrir arquivo "/home/user/.local/share/icc/edid-124a3fc627aa9c1dd15dc66cd0d16108.icc": Permissão negada
07:46:05 gdm-x-session: (II) systemd-logind: got pause for 13:65
07:46:05 gdm-x-session: (II) systemd-logind: got pause for 13:69
07:46:05 gdm-x-session: (II) systemd-logind: got pause for 13:67
07:46:05 gdm-x-session: (II) systemd-logind: got pause for 13:64
07:46:05 gdm-x-session: (II) systemd-logind: got pause for 13:66
07:46:05 gdm-x-session: (II) systemd-logind: got pause for 13:68
07:46:05 gdm-x-session: (II) systemd-logind: got pause for 226:0
07:46:05 kernel: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=getty@tty3 comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
07:46:05 systemd: Started Getty on tty3.
07:46:05 gdm-x-session: (II) AIGLX: Suspending AIGLX clients for VT switch
07:46:05 gdm-x-session: (**) Option "fd" "28"
07:46:05 gdm-x-session: (**) Option "fd" "28"
07:46:05 gdm-x-session: (**) Option "fd" "27"
07:46:05 gdm-x-session: (**) Option "fd" "26"
07:46:05 gdm-x-session: (**) Option "fd" "25"
07:46:05 gdm-x-session: (**) Option "fd" "24"
07:46:05 gdm-x-session: (**) Option "fd" "21"

There is a record in kernel log today befora those. I don't think that is the problem but :-) :
00:21:23 trt-42428.trt23.gov.br kernel: perf: interrupt took too long (3147 > 3127), lowering kernel.perf_event_max_sample_rate to 63000


I hope those records help you to find a solution. Thank you.


:-)
Comment 40 Jonas Ådahl 2017-03-24 01:47:11 UTC
This seems to include logs that looks like VT switching, from and to an X11 session. There is also a power button press logged. Does that ring a bell? Does the freeze get triggered by something like that? Does it get triggered in the same way when running Wayland and X11?
Comment 41 Moises Bites 2017-03-24 14:19:17 UTC
When I moved the mouse to login in locked screen, the mouse was frozen. So, I tried to switch the VT by keyboard, but didn't work. Because I didn't get the control of machine I needed to power off and restart the OS. :-(

The system freezes less in X11, but still freezes sometimes, almost always when the lock screen is activated.

I will try more tests, and I will report here. :-)
Comment 42 Kamil Páral 2017-03-24 14:30:50 UTC
Moises, I think this is most definitely a different issue and should be reported separately.
Comment 43 Moises Bites 2017-03-24 14:32:13 UTC
(In reply to Kamil Páral from comment #42)
> Moises, I think this is most definitely a different issue and should be
> reported separately.

Ok. Thank you.
Comment 44 Eduardo 2017-05-28 04:02:32 UTC
This bug is still present under gnome-shell 3.24.2

Besides, it happens as well with only two monitors (laptop + external), with no apparent reason: Qt apps (QOwnNotes, Clementine)? Firefox? VPN?
Comment 45 Kamil Páral 2017-05-29 08:02:12 UTC
Eduardo, it will be most likely a different bug. Please file a new report and attach as much debug information as you can (package versions, journal output, whether you can ssh in during display freeze, etc). Thanks.