Bug 1303643

Summary: qt5-qtdeclarative-5.6.0-0.7.beta.fc24 broken
Product: [Fedora] Fedora Reporter: Sandro Mani <manisandro>
Component: qt5-qtdeclarativeAssignee: Rex Dieter <rdieter>
Status: CLOSED RAWHIDE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: rawhideCC: eddy.pilon, jgrulich, jreznik, laurent.rineau__fedora, ltinkl, manisandro, rdieter, rnovacek, than
Target Milestone: ---   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: qt5-qtdeclarative-5.6.0-0.8.beta.fc24 Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2016-02-03 23:29:06 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:

Description Sandro Mani 2016-02-01 14:51:26 UTC 
Description of problem:
qt5-qtdeclarative-5.6.0-0.7.beta.fc24 causes any Qt5+QML application to crash immediately on start:

$ valgrind qtcreator

==7839== Memcheck, a memory error detector
==7839== Copyright (C) 2002-2015, and GNU GPL'd, by Julian Seward et al.
==7839== Using Valgrind-3.11.0 and LibVEX; rerun with -h for copyright info
==7839== Command: qtcreator
==7839== 
==7839== Invalid read of size 8
==7839==    at 0x7BEFDDB: QV4::Object::defineDefaultProperty(QString const&, QV4::Value const&) (in /usr/lib64/libQt5Qml.so.5.6.0)
==7839==    by 0x7BF5AC0: ??? (in /usr/lib64/libQt5Qml.so.5.6.0)
==7839==    by 0x7BB3698: QV4::ExecutionEngine::ExecutionEngine(QV4::EvalISelFactory*) (in /usr/lib64/libQt5Qml.so.5.6.0)
==7839==    by 0x7CE6CA0: QV8Engine::QV8Engine(QJSEngine*) (in /usr/lib64/libQt5Qml.so.5.6.0)
==7839==    by 0x7B4203D: QJSEngine::QJSEngine() (in /usr/lib64/libQt5Qml.so.5.6.0)
==7839==    by 0x238395AF: Core::JsExpander::JsExpander() (in /usr/lib64/qtcreator/plugins/libCore.so)
==7839==    by 0x2381EB93: ??? (in /usr/lib64/qtcreator/plugins/libCore.so)
==7839==    by 0x2389F3D4: ??? (in /usr/lib64/qtcreator/plugins/libCore.so)
==7839==    by 0x2389F6B3: ??? (in /usr/lib64/qtcreator/plugins/libCore.so)
==7839==    by 0x4E5F83D: ExtensionSystem::Internal::PluginSpecPrivate::initializePlugin() (in /usr/lib64/qtcreator/libExtensionSystem.so.1.0.0)
==7839==    by 0x4E59869: ExtensionSystem::Internal::PluginManagerPrivate::loadPlugin(ExtensionSystem::PluginSpec*, ExtensionSystem::PluginSpec::State) (in /usr/lib64/qtcreator/libExtensionSystem.so.1.0.0)
==7839==    by 0x4E5A24C: ExtensionSystem::Internal::PluginManagerPrivate::loadPlugins() (in /usr/lib64/qtcreator/libExtensionSystem.so.1.0.0)
==7839==  Address 0x0 is not stack'd, malloc'd or (recently) free'd
==7839== 
==7839== 
==7839== Process terminating with default action of signal 11 (SIGSEGV)
==7839==  Access not within mapped region at address 0x0
==7839==    at 0x7BEFDDB: QV4::Object::defineDefaultProperty(QString const&, QV4::Value const&) (in /usr/lib64/libQt5Qml.so.5.6.0)
==7839==    by 0x7BF5AC0: ??? (in /usr/lib64/libQt5Qml.so.5.6.0)
==7839==    by 0x7BB3698: QV4::ExecutionEngine::ExecutionEngine(QV4::EvalISelFactory*) (in /usr/lib64/libQt5Qml.so.5.6.0)
==7839==    by 0x7CE6CA0: QV8Engine::QV8Engine(QJSEngine*) (in /usr/lib64/libQt5Qml.so.5.6.0)
==7839==    by 0x7B4203D: QJSEngine::QJSEngine() (in /usr/lib64/libQt5Qml.so.5.6.0)
==7839==    by 0x238395AF: Core::JsExpander::JsExpander() (in /usr/lib64/qtcreator/plugins/libCore.so)
==7839==    by 0x2381EB93: ??? (in /usr/lib64/qtcreator/plugins/libCore.so)
==7839==    by 0x2389F3D4: ??? (in /usr/lib64/qtcreator/plugins/libCore.so)
==7839==    by 0x2389F6B3: ??? (in /usr/lib64/qtcreator/plugins/libCore.so)
==7839==    by 0x4E5F83D: ExtensionSystem::Internal::PluginSpecPrivate::initializePlugin() (in /usr/lib64/qtcreator/libExtensionSystem.so.1.0.0)
==7839==    by 0x4E59869: ExtensionSystem::Internal::PluginManagerPrivate::loadPlugin(ExtensionSystem::PluginSpec*, ExtensionSystem::PluginSpec::State) (in /usr/lib64/qtcreator/libExtensionSystem.so.1.0.0)
==7839==    by 0x4E5A24C: ExtensionSystem::Internal::PluginManagerPrivate::loadPlugins() (in /usr/lib64/qtcreator/libExtensionSystem.so.1.0.0)



Version-Release number of selected component (if applicable):
qt5-qtdeclarative-5.6.0-0.7.beta.fc24

How reproducible:
Always


Additional info:
Possibly GCC6 related.
Comment 1 Sandro Mani 2016-02-01 15:44:48 UTC 
Rebuilding qt5-qtdeclarative-5.6.0-0.6.beta.fc24 with GCC6 also breaks it, so yes, it is related to GCC6.
Comment 2 Than Ngo 2016-02-01 15:46:55 UTC 
it looks like gcc6 miscompiled the codes again. I'm trying to build it with -O0 and checks if it works
Comment 3 Than Ngo 2016-02-01 21:10:37 UTC 
which platform did you test?
Comment 4 Sandro Mani 2016-02-01 21:12:03 UTC 
on x86_64
Comment 5 Than Ngo 2016-02-02 13:39:31 UTC 
*** Bug 1303881 has been marked as a duplicate of this bug. ***
Comment 6 Rex Dieter 2016-02-02 15:47:16 UTC 
We're guessing this may be due to gcc6 new behavior in, see
https://gcc.gnu.org/ml/gcc/2016-01/msg00184.html

In which case, we'll try (re)building Qt5 with -fno-delete-null-pointer-checks
Comment 7 Rex Dieter 2016-02-02 16:04:38 UTC 
%changelog
* Tue Feb 02 2016 Rex Dieter <rdieter> 5.6.0-0.8.beta
- build with -fno-delete-null-pointer-checks to workaround gcc6-related runtime crashes (#1303643)
Comment 8 Than Ngo 2016-02-02 17:09:20 UTC 
it seems the compiler flag -fno-delete-null-pointer-checks will fix the miscompiled codes. We can use it as the workaround.

Sandro, could you please try the new build with this flag and give us the feebacks

http://koji.fedoraproject.org/koji/buildinfo?buildID=715479

Thanks
Comment 9 Sandro Mani 2016-02-02 17:11:29 UTC 
Yep, works, thanks!
Comment 10 poma 2016-02-03 05:24:58 UTC 
[NVA8/NV98] KDE Plasma locks up: Nouveau reports error "resource sanity check" "unable to handle kernel paging request"
https://bugs.freedesktop.org/show_bug.cgi?id=92892
Comment 11 Kevin Kofler 2016-02-03 23:27:56 UTC 
That upstream bug is a completely different issue.
Comment 12 poma 2016-02-04 14:34:43 UTC 
You are some kind of the Nouveau "expert"? :)

BTW Rawhide's KDE Live is B.R.O.K.E.N.
Who maintains that crap?
Comment 13 Rex Dieter 2016-02-04 14:40:54 UTC 
Please adhere to the fedora code of conduct.

Again, if you're trying interested in your issue being fixed, file a separate bug (nouveau bugs are not relevant to this report)