Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

shaurz · 2015-07-02 18:57:13

Since the latest package updates for Plasma Desktop 5, it segfaults every time I shut down. It also randomly segfaults from time to time. Anyone else have this problem?

Ever since I updated to version 5, pretty much as soon as the packages were in [current], sleep has never worked for me (I just got used to not having it). When the PC awakes from sleep it shows the login screen. When I log in, all applications have been shut down uncleanly, like as if the system has been rebooted. Does sleep work for anyone else?

x11tete11x · 2015-07-03 18:54:07

I can confirm, same here

pb · 2015-07-03 21:15:06

I can confirm seftaults.

phoenix · 2015-07-04 13:45:30

For me it only segfaults on shutdown, no random crashes.

shaurz · 2015-07-04 23:10:06

It looks like for me, the random crashes are in the userland Intel HD graphics MESA driver:

Thread 1 (Thread 0x7f261ee1e800 (LWP 3232)):
[KCrash Handler]
#5  0x00007f25feb399cc in ?? () from /usr/lib/xorg/modules/dri/i965_dri.so
#6  0x00007f25feac50cc in ?? () from /usr/lib/xorg/modules/dri/i965_dri.so
#7  0x00007f261c8b7d4a in QSGBatchRenderer::Renderer::renderBatches() () from /usr/lib/libQt5Quick.so.5
#8  0x00007f261c8bc50a in QSGBatchRenderer::Renderer::render() () from /usr/lib/libQt5Quick.so.5
#9  0x00007f261c8c715c in QSGRenderer::renderScene(QSGBindable const&) () from /usr/lib/libQt5Quick.so.5
#10 0x00007f261c8c75db in QSGRenderer::renderScene(unsigned int) () from /usr/lib/libQt5Quick.so.5
#11 0x00007f261c8d5cde in QSGRenderContext::renderNextFrame(QSGRenderer*, unsigned int) () from /usr/lib/libQt5Quick.so.5
#12 0x00007f261c91f95c in QQuickWindowPrivate::renderSceneGraph(QSize const&) () from /usr/lib/libQt5Quick.so.5
#13 0x00007f261c8f02bc in ?? () from /usr/lib/libQt5Quick.so.5
#14 0x00007f261c8f0de1 in ?? () from /usr/lib/libQt5Quick.so.5
#15 0x00007f261a47862c in QApplicationPrivate::notify_helper(QObject*, QEvent*) () from /usr/lib/libQt5Widgets.so.5
#16 0x00007f261a47dd10 in QApplication::notify(QObject*, QEvent*) () from /usr/lib/libQt5Widgets.so.5
#17 0x00007f261912157b in QCoreApplication::notifyInternal(QObject*, QEvent*) () from /usr/lib/libQt5Core.so.5
#18 0x00007f2619177b1d in QTimerInfoList::activateTimers() () from /usr/lib/libQt5Core.so.5
#19 0x00007f2619178021 in ?? () from /usr/lib/libQt5Core.so.5
#20 0x00007f2614b139fd in g_main_context_dispatch () from /usr/lib/libglib-2.0.so.0
#21 0x00007f2614b13ce0 in ?? () from /usr/lib/libglib-2.0.so.0
#22 0x00007f2614b13d8c in g_main_context_iteration () from /usr/lib/libglib-2.0.so.0
#23 0x00007f2619178cff in QEventDispatcherGlib::processEvents(QFlags<QEventLoop::ProcessEventsFlag>) () from /usr/lib/libQt5Core.so.5
#24 0x00007f261911effa in QEventLoop::exec(QFlags<QEventLoop::ProcessEventsFlag>) () from /usr/lib/libQt5Core.so.5
#25 0x00007f2619126a4c in QCoreApplication::exec() () from /usr/lib/libQt5Core.so.5
#26 0x000000000042ed66 in main ()

Last edited by shaurz (2015-07-04 23:11:46)

pb · 2015-07-05 07:56:21

It's rather not related to the graphics card. I have Radeon with open source driver.

Fedes · 2015-07-06 16:36:18

Same here... it's getting annoying. This happened after an upgrade of intel drivers, kernel, mesa and QT.. so anything could be the issue. It used to be more stable, though never bullet proof.

pb · 2015-07-07 04:52:35

For me, it starts with upgrade to Plasma 5.3.2.

Guidobelix · 2015-07-10 14:53:35

Same problem with segfault here after upgrade to Plasma 5.3.2
Has anybody got a suggestion?

WorMzy · 2015-07-10 15:33:44

Has anybody got a suggestion?

Less 'me too's and more logs for a start. Has anyone opened a bug report?

martinc2 · 2015-07-10 15:41:52

Hi, there's these two reports on the kde site (see here). The crashes seem to be intel-specific (apart from @pb's report above), in the Mesa code as @shaurz suggests. I've changed my intel conf from sna to uxa, with good results so far.

--martinc

aocab · 2015-07-10 18:38:53

I'm seeing a segfault on both logout and restart and the session gets messed up.

 $ loginctl session-status c1
c1 - alpha (1000)
   Since: Thu 2015-07-02 18:26:31 CDT; 4min 15s ago
  Leader: 304
Seat: seat0; vc1
TTY: tty1
Service: login; type tty; class user
   State: closing
Unit: session-c1.scope
  &#9500;&#9472;393 /usr/lib/gamin/gam_server
  &#9500;&#9472;484 /usr/bin/plasmashell --shut-up
  &#9500;&#9472;511 /usr/bin/pulseaudio --start --log-target=syslog
  &#9492;&#9472;553 /usr/bin/kuiserver5

Jul 02 18:27:54 asusm32 kdeinit5[384]: QDBusAbstractAdaptor: Cannot relay signal KDEDModule::moduleDeleted(KDEDModule*): Pointer
Jul 02 18:27:54 asusm32 kdeinit5[384]: QDBusAbstractAdaptor: Cannot relay signal KDEDModule::moduleDeleted(KDEDModule*): Pointer
Jul 02 18:27:54 asusm32 kdeinit5[384]: QThread: Destroyed while thread is still running
Jul 02 18:27:54 asusm32 kdeinit5[384]: QDBusAbstractAdaptor: Cannot relay signal KDEDModule::moduleDeleted(KDEDModule*): Pointer
Jul 02 18:27:54 asusm32 kdeinit5[384]: QDBusAbstractAdaptor: Cannot relay signal KDEDModule::moduleDeleted(KDEDModule*): Pointer
Jul 02 18:27:54 asusm32 kdeinit5[384]: QDBusAbstractAdaptor: Cannot relay signal KDEDModule::moduleDeleted(KDEDModule*): Pointer
Jul 02 18:27:54 asusm32 kdeinit5[384]: QDBusAbstractAdaptor: Cannot relay signal KDEDModule::moduleDeleted(KDEDModule*): Pointer
Jul 02 18:27:54 asusm32 kglobalaccel5[440]: The X11 connection broke (error 1). Did the X11 server die?
Jul 02 18:27:54 asusm32 kactivitymanagerd[408]: The X11 connection broke (error 1). Did the X11 server die?

Program received signal SIGSEGV, Segmentation fault.
PlasmaQuick::View::containment (this=0x3597de0) at /var/abs/local/plasma-framework/src/plasma-framework-5.11.0/src/plasmaquick/view.cpp:248
248	    return d->containment;
#0  PlasmaQuick::View::containment (this=0x3597de0) at /var/abs/local/plasma-framework/src/plasma-framework-5.11.0/src/plasmaquick/view.cpp:248
#1  0x000000000044f845 in ShellCorona::screenForContainment (this=0x2008cf0, containment=0x32a9f80) at /var/abs/local/plasma-workspace/src/plasma-workspace-5.3.2/shell/shellcorona.cpp:1455
#2  0x00007f35160beb71 in ContainmentInterface::availableScreenRect (this=this@entry=0x32affb0) at /var/abs/local/plasma-framework/src/plasma-framework-5.11.0/src/scriptengines/qml/plasmoid/containmentinterface.cpp:220
#3  0x00007f35160c8b88 in ContainmentInterface::qt_metacall (this=0x32affb0, _c=<optimized out>, _id=5, _a=0x7ffd8f14fb30) at /var/abs/local/plasma-framework/src/build/src/scriptengines/qml/moc_containmentinterface.cpp:392
#4  0x00007f353dd86c89 in readProperty (p=0x397f2c4, idx=<optimized out>, obj=<optimized out>, this=0x397f2b0) at qml/qqmlvaluetype_p.h:86
#5  QQmlValueTypeBase<QRect>::read (this=0x397f2b0, obj=<optimized out>, idx=<optimized out>) at qml/qqmlvaluetype_p.h:114
#6  0x00007f353ddcd30a in readReferenceValue (reference=reference@entry=0x7f35041069e0) at qml/qqmlvaluetypewrapper.cpp:125
#7  0x00007f353ddce595 in QV4::QmlValueTypeWrapper::get (m=0x7f35041069e0, name=0x7f35040a6c40, hasProperty=0x0) at qml/qqmlvaluetypewrapper.cpp:275
#8  0x00007f353dd1cd17 in get (hasProperty=0x0, name=0x7f35040a6c40, this=<optimized out>) at jsruntime/qv4object_p.h:252
#9  QV4::Runtime::getProperty (ctx=0x7ffd8f14fdd0, object=..., name=0x7f35040a6c40) at jsruntime/qv4runtime.cpp:654
#10 0x00007f345c41df6e in ?? ()
#11 0x0000000000000000 in ?? ()

ReiserFS · 2015-07-10 19:07:17

martinc2 wrote:

Hi, there's these two reports on the kde site (see here). The crashes seem to be intel-specific (apart from @pb's report above), in the Mesa code as @shaurz suggests. I've changed my intel conf from sna to uxa, with good results so far.
--martinc

It works for me.

baranator · 2015-07-23 00:12:47

martinc2 wrote:

Hi, there's these two reports on the kde site (see here). The crashes seem to be intel-specific (apart from @pb's report above), in the Mesa code as @shaurz suggests. I've changed my intel conf from sna to uxa, with good results so far.
--martinc

had the exact same crash-report as shaurz, switching to uxa fixed it for me, too.

shaurz · 2015-07-23 10:47:10

I switched to uxa as per the instruction at: https://wiki.archlinux.org/index.php/In … SNA_issues

But the problem still persists! How can I check that UXA is being used?

baranator · 2015-07-23 19:34:45

shaurz wrote:

I switched to uxa as per the instruction at: https://wiki.archlinux.org/index.php/In … SNA_issues
But the problem still persists! How can I check that UXA is being used?

when I enable uxa i get the following output on /var/log/Xorg.0.log:

[     7.840] (II) UXA(0): Driver registered support for the following operations:
[     7.840] (II)         solid
[     7.840] (II)         copy
[     7.840] (II)         composite (RENDER acceleration)
[     7.840] (II)         put_image
[     7.840] (II)         get_image

do you have that output too? i think this should mean that uxa was loaded.

fjvinal · 2015-07-24 09:39:23

I also have UXA and the issue still happens:

Section "Device"
        Identifier      "HP"
        Driver          "intel"
        Option  "AccelMethod"   "uxa"
EndSection

arojas · 2015-07-24 09:49:32

This thread has become a mess of different unrelated issues, which makes it little helpful. The crash in #5 was introduced in a recent update to the Intel graphics stack, there's no known fix yet and switching to UXA should work around it. The crash in #12 is totally unrelated, and it's already fixed for Plasma 5.4 (https://bugs.kde.org/show_bug.cgi?id=343058). Switching to UXA should have no effect on that one, since it has nothing to to with the X driver.

pirogornian · 2015-07-24 15:09:37

I have also plasma crashes at shutdown, and after resume there is desktop visible, but soon hard disk start working and system response become very slow, so slow that I must restart by power button. Only few times I got clean esume, with login screen.

I often have also hung up of plasmashell process. It stop responding (dekstop freezes, but not applications), it begins consume ~50% CPU and I must kill it and run again.

Graphics: nVidia GeForce 8600GT, driver: nouveau (I tried recently legacy nvidia driver, but desktop was nearly unusable). Login manager: sddm (unusable on PC with intel card, btw).

loginctl session-status did't show any errors, even after plasmashell process manual restart.

Arch Linux

#1 2015-07-02 18:57:13

Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

#2 2015-07-03 18:54:07

Re: Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

#3 2015-07-03 21:15:06

Re: Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

#4 2015-07-04 13:45:30

Re: Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

#5 2015-07-04 23:10:06

Re: Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

#6 2015-07-05 07:56:21

Re: Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

#7 2015-07-06 16:36:18

Re: Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

#8 2015-07-07 04:52:35

Re: Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

#9 2015-07-10 14:53:35

Re: Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

#10 2015-07-10 15:33:44

Re: Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

#11 2015-07-10 15:41:52

Re: Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

#12 2015-07-10 18:38:53

Re: Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

#13 2015-07-10 19:07:17

Re: Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

#14 2015-07-23 00:12:47

Re: Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

#15 2015-07-23 10:47:10

Re: Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

#16 2015-07-23 19:34:45

Re: Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

#17 2015-07-24 09:39:23

Re: Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

#18 2015-07-24 09:49:32

Re: Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

#19 2015-07-24 15:09:37

Re: Plasma Desktop 5 segfaults on shutdown, wake from sleep never worked

Board footer