Note: This is a public test instance of Red Hat Bugzilla. The data contained within is a snapshot of the live data so any changes you make will not be reflected in the production Bugzilla. Email is disabled so feel free to test any aspect of the site that you want. File any problems you find or give feedback at bugzilla.redhat.com.

Bug 1331593

Summary: Fedora 24 i686 KDE gets stuck at a black screen on login
Product: [Fedora] Fedora Reporter: Adam Williamson <awilliam>
Component: plasma-desktopAssignee: Rex Dieter <rdieter>
Status: CLOSED DUPLICATE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: high Docs Contact:
Priority: unspecified    
Version: 24CC: jgrulich, kde-sig, me, mrmazda, rdieter, than
Target Milestone: ---   
Target Release: ---   
Hardware: i686   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2016-05-28 16:06:37 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:
Embargoed:
Attachments:
Description Flags
/var/log/messages
none
journal none

Description Adam Williamson 2016-04-28 22:54:45 UTC
I was sure I'd already reported this, but can't find the bug: please do dupe it if you can.

Current F24 KDE 32-bit lives hang at a black screen when the desktop should appear. This has been happening ever since the 32-bit kernel bugs were resolved, and I think it was actually happening *before* that too. See e.g.:

https://openqa.fedoraproject.org/tests/14921

I can reproduce this manually both in a VM and on bare metal. Attaching journal and /var/log/messages from a VM boot.

Comment 1 Adam Williamson 2016-04-28 22:55:08 UTC
Created attachment 1152069 [details]
/var/log/messages

Comment 2 Adam Williamson 2016-04-28 22:55:30 UTC
Created attachment 1152070 [details]
journal

Comment 3 Adam Williamson 2016-04-28 23:01:59 UTC
So the system isn't really hung, it's more like it's at a desktop but just not showing most things. KDE is definitely *there*. If I hit the power button, the "Logging out" stripe appears, for e.g. If I move the mouse to the top-left corner of the screen, I see the blue glow. Clicking where the kicker should be doesn't make it visible, though.

Comment 4 Felix Miata 2016-04-29 04:18:51 UTC
This is far from a fresh problem. Earliest mention I can find that I made anywhere is https://bugzilla.redhat.com/show_bug.cgi?id=1283348#c33 but I know I didn't mention until some number of weeks after first seeing it. I just booted multiboot host gx27b's F24 to update, after last having been updated 9 weeks ago, and it did exactly as I described there. I was able to open apps via Alt-F2, but until then, and the inevitable segfault window, the only GUI activity I saw were mouse pointer and brief startup splash screen. After dnf update, black screen/missing panel behavior is unchanged, but a segfault window does not materialize.

Comment 5 Rex Dieter 2016-04-29 10:22:50 UTC
it appears plasmashell process may be deadlocked, getting a backtrace from it may be helpful

Comment 6 pavel raur 2016-05-23 11:16:32 UTC
Happened the same to me.

Fedora i386 Live KDE

Fedora-KDE-Live-i386-24-20160522.n.0.iso
After boot shows just black screen.

Comment 7 Peter Gückel 2016-05-23 17:40:42 UTC
Yup, yet another problem on my i686 (x86_64 unaffected).

Gnome works fine, but when I try to log into KDE, I get the pretty starfield and the nice progress bar, all swift and clean, and then...

nothing!

Comment 8 Rex Dieter 2016-05-23 19:07:31 UTC
OK, just reproduced it on my box.

As suspected (by me), it's dying down in /lib/sse2/libQt5ml.so.5 in QV4:: routines, apparently related to memory management

(transcribing)

#0 ... in ?? ()
#1 ... in QV4::PersistentValueStorage::mark(QV4::ExecuationEngine*) ...
#2 ... in QV4::MemoryManager::mark() ...
#3 ... in QV4::MemoryManager::runGC() ...
#4 ... in QV4::MemoryManager::allocData(unsigned int, unsigned int) ...
#5 ... in QV4::ExecuationEngine::newVariantObject(QVariant const&) ...

For giggles, moved the sse2 one out of the way to force use of /lib/libQt5Qml.so.5, and get an (apparent) identical backtrace.

Comment 9 Rex Dieter 2016-05-23 19:17:22 UTC
Looks like a dup of similar/same issue in bug #1329715 (currently assigned to sddm), but it's clearly both lower-level in qt5-qtdeclarative

Comment 10 Felix Miata 2016-05-25 06:27:04 UTC
Is this bug what I'm seeing on i686 F24 host gx280, using startx in multi-user.target?

# rpmqa kactiv
kactivities-libs-4.13.3-20.fc24.i686
kactivitymanagerd-5.6.4-1.fc24.i686
kf5-kactivities-libs-5.22.0-1.fc24.i686
kf5-kactivities-stats-5.22.0-1.fc24.i686
# rpm -qa | grep ^qt
qt-4.8.7-17.fc24.i686
...
qt5-qtdeclarative-5.6.0-5.fc24.i686
...
[  149.727392] traps: plasmashell[1328] general protection ip:ab6df4cd sp:bfcd1be8 error:502
[  362.029952] kactivitymanage[1221]: segfault at b0b52b4c ip af1c5ccc sp bfa8010c error 4 in libQt5Sql.so.5.6.0[af1b3000+48000]
[  454.716529] traps: plasmashell[1801] general protection ip:ab6694ae sp:bfb7556c error:c6a0
[  489.654786] kactivitymanage[1700]: segfault at b0b15b4c ip b01c7ccc sp bf8149bc error 4 in libQt5Sql.so.5.6.0[b01b5000+48000]
[  727.423430] kactivitymanage[2084]: segfault at b0b11b4c ip ae779ccc sp bfcd990c error 4 in libQt5Sql.so.5.6.0[ae767000+48000]
[ 4827.296710] kactivitymanage[2510]: segfault at b0b57b4c ip b01c7ccc sp bfaa3b7c error 4 in libQt5Sql.so.5.6.0[b01b5000+48000]
[ 6152.646571] systemsettings5[3173]: segfault at 0 ip   (null) sp bfa2278c error 4 in systemsettings5[80042000+1a000]

I'm able to get Konsole to open via a quick Alt-F2 after splash went away, but not if I wait.

Comment 11 Rex Dieter 2016-05-28 16:04:25 UTC
OK, after a bit of discussion, kde-sig highly suspected this is a qt5-qtdeclarative issue, most likely triggered by gcc6.

So, installed f24/i686 workstation, and installed @kde-desktop-environment on top of that to do additional testing, and found that building qt5-qtdeclarative with addtional flag: -fno-lifetime-dse , seemed to fix the crashes for me.

Here's a build that includes that for testing purposes:
http://koji.fedoraproject.org/koji/taskinfo?taskID=14291923

Comment 12 Rex Dieter 2016-05-28 16:06:37 UTC

*** This bug has been marked as a duplicate of bug 1329715 ***