Bug 67203 - LibreOffice Impress 4.1 3D OpenGL transitions induce a crash on R600g
Summary: LibreOffice Impress 4.1 3D OpenGL transitions induce a crash on R600g
Status: RESOLVED DUPLICATE of bug 67908
Alias: None
Product: LibreOffice
Classification: Unclassified
Component: UI (show other bugs)
Version:
(earliest affected)
unspecified
Hardware: x86-64 (AMD64) Linux (All)
: medium major
Assignee: Not Assigned
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2013-07-23 04:15 UTC by princetonf
Modified: 2013-08-09 11:26 UTC (History)
4 users (show)

See Also:
Crash report or crash signature:


Attachments
Debug info (stack trace) for LibreOffice Impress. (28.96 KB, text/x-log)
2013-07-23 12:01 UTC, princetonf
Details
GLX Info (56.44 KB, text/plain)
2013-07-26 00:37 UTC, princetonf
Details

Note You need to log in before you can comment on or make changes to this bug.
Description princetonf 2013-07-23 04:15:43 UTC
With LibreOffice 4.1 and Mesa 9.2 (on Fedora 19), the following error occurs when trying to demo any 3D (has to be from the libreoffice-ogltrans package) transition.

This bug seems to exist for users of the R600g driver, as I cannot seem to confirm this at the moment on any other cards that are available to me.


The following output was generated in gdb:
====================================================
[New Thread 0x7ffff1af4700 (LWP 4058)]
[New Thread 0x7fffe880f700 (LWP 4059)]
[New Thread 0x7fffe3fff700 (LWP 4060)]
[New Thread 0x7fffe2ff6700 (LWP 4061)]
[New Thread 0x7fffdb2af700 (LWP 4062)]
Detaching after fork from child process 4063.
[New Thread 0x7fffd90fc700 (LWP 4065)]
[Thread 0x7fffd90fc700 (LWP 4065) exited]
[Thread 0x7fffe880f700 (LWP 4059) exited]
[New Thread 0x7fffe880f700 (LWP 4066)]
[New Thread 0x7fffd90fc700 (LWP 4067)]
[Thread 0x7fffd90fc700 (LWP 4067) exited]
[New Thread 0x7fffceb75700 (LWP 4068)]
[New Thread 0x7fffd90fc700 (LWP 4069)]
[Thread 0x7fffe880f700 (LWP 4066) exited]
[New Thread 0x7fffe880f700 (LWP 4070)]
[New Thread 0x7fffcd285700 (LWP 4071)]
[New Thread 0x7fffc3afc700 (LWP 4072)]
[Thread 0x7fffc3afc700 (LWP 4072) exited]
[Thread 0x7fffe880f700 (LWP 4070) exited]
[New Thread 0x7fffe880f700 (LWP 4073)]
[New Thread 0x7fffc3afc700 (LWP 4074)]
[Thread 0x7fffc3afc700 (LWP 4074) exited]
[Thread 0x7fffe880f700 (LWP 4073) exited]
[New Thread 0x7fffe880f700 (LWP 4075)]

Program received signal SIGSEGV, Segmentation fault.
0x0000003c67c1a167 in GetGLXScreenConfigs () from /lib64/libGL.so.1
====================================================
Comment 1 Julien Nabet 2013-07-23 06:07:37 UTC
princeton: would it be possible you provide a backtrace by following this link? https://wiki.documentfoundation.org/QA/BugReport/Debug_Information#GNU.2FLinux:_How_to_get_a_backtrace
Comment 2 princetonf 2013-07-23 12:01:03 UTC
Created attachment 82862 [details]
Debug info (stack trace) for LibreOffice Impress.
Comment 3 Julien Nabet 2013-07-23 17:11:07 UTC
princeton: interesting bt! We can see scrn=1702129263 so obviously, screens[scrn] fails.
What's your glx version? (just type "glxinfo|grep version") (mine is GLX version: 1.4)

Michael: could ee4569efa8322497112e8e951a67b9adca4ac55e may help here?

BTW, i noticed there were 2 OGLTrans_TransitionerImpl.cxx files:
- slideshow/source/engine/OGLTrans/win/
- slideshow/source/engine/OGLTrans/generic/
the weird thing is in both we can see things like:
#if defined( WNT ) or #elif defined( UNX ) etc.
Comment 4 princetonf 2013-07-24 01:52:01 UTC
GLX version is 1.4 as well.
Comment 5 Julien Nabet 2013-07-24 05:41:15 UTC
princeton: ok. Since i've got no more questions, I put it back to unconfirmed.
Comment 6 Michael Meeks 2013-07-24 10:32:27 UTC
Prolly worth digging at the vi instantiation in:

bool OGLTransitionerImpl::createWindow( Window* pPWindow )

Looks to me as if we can XFree (vi) and then go on to assign it to GLWin.vi - but - perhaps I'm confused wrt. the code-flow there.

it'd be nice to have some 'glxinfo' output for your system attached (I guess) :-)

Thanks ! - Thorsten, whats' your take on that ?
Comment 7 princetonf 2013-07-26 00:37:39 UTC
Created attachment 83014 [details]
GLX Info

> it'd be nice to have some 'glxinfo' output for your system attached (I
> guess) :-)

Sorry about that, and about getting back to you on such a delayed notice.
Comment 8 rogern 2013-07-30 19:00:36 UTC
I appear to be hitting the same or a very similar bug using the nvidia proprietary drivers in Fedora 19, GLX 1.4. LibreOffice 4.0.4.2 works OK. Traceback below:

Program received signal SIGSEGV, Segmentation fault.
0x0000003990c822d4 in ?? () from /usr/lib64/nvidia/libGL.so.1
(gdb) where
#0  0x0000003990c822d4 in ?? () from /usr/lib64/nvidia/libGL.so.1
#1  0x0000003990c8559a in ?? () from /usr/lib64/nvidia/libGL.so.1
#2  0x0000003990c7d331 in glXCreateContext () from /usr/lib64/nvidia/libGL.so.1
#3  0x00007f97637e86e4 in (anonymous namespace)::OGLTransitionerImpl::initWindowFromSlideShowView(com::sun::star::uno::Reference<com::sun::star::presentation::XSlideShowView> const&) ()
   from /usr/lib64/libreoffice/program/../program/libOGLTranslo.so
#4  0x00007f97637ec7dc in (anonymous namespace)::OGLTransitionFactoryImpl::createTransition(short, short, com::sun::star::uno::Reference<com::sun::star::presentation::XSlideShowView> const&, com::sun::star::uno::Reference<com::sun::star::rendering::XBitmap> const&, com::sun::star::uno::Reference<com::sun::star::rendering::XBitmap> const&) ()
   from /usr/lib64/libreoffice/program/../program/libOGLTranslo.so
#5  0x00007f9763b601d6 in slideshow::internal::(anonymous namespace)::PluginSlideChange::addTransition(boost::shared_ptr<slideshow::internal::UnoView> const&) () from /usr/lib64/libreoffice/program/../program/libslideshowlo.so
#6  0x00007f9763b618bb in slideshow::internal::TransitionFactory::createSlideTransition(boost::shared_ptr<slideshow::internal::Slide> const&, boost::shared_ptr<slideshow::internal::Slide> const&, slideshow::internal::UnoViewContainer const&, slideshow::internal::ScreenUpdater&, slideshow::internal::EventMultiplexer&, com::sun::star::uno::Reference<com::sun::star::presentation::XTransitionFactory> const&, short, short, bool, slideshow::internal::RGBColor const&, boost::shared_ptr<slideshow::internal::SoundPlayer> const&) ()
   from /usr/lib64/libreoffice/program/../program/libslideshowlo.so
#7  0x00007f9763b322eb in (anonymous namespace)::SlideShowImpl::displaySlide(com::sun::star::uno::Reference<com::sun::star::drawing::XDrawPage> const&, com::sun::star::uno::Reference<com::sun::star::drawing::XDrawPagesSupplier> const&, com::sun::star::uno::Reference<com::sun::star::animations::XAnimationNode> const&, com::sun::star::uno::Sequence<com::sun::star::beans::PropertyValue> const&) () from /usr/lib64/libreoffice/program/../program/libslideshowlo.so
#8  0x00007f9783a9fcd9 in sd::AnimationSlideController::displayCurrentSlide(com::sun::star::uno::Reference<com::sun::star::presentation::XSlideShow> const&, com::sun::star::uno::Reference<com::sun::star::drawing::XDrawPagesSupplier> const&, bool) () from /usr/lib64/libreoffice/program/../program/libsdlo.so
#9  0x00007f9783aa1b8e in sd::SlideshowImpl::displayCurrentSlide(bool) ()
   from /usr/lib64/libreoffice/program/../program/libsdlo.so
#10 0x00007f9783aa2892 in sd::SlideshowImpl::startShowImpl(com::sun::star::uno::Sequence<com::sun::star::beans::PropertyValue> const&) () from /usr/lib64/libreoffice/program/../program/libsdlo.so
#11 0x00007f9783aa6491 in sd::SlideshowImpl::startPreview(com::sun::star::uno::Reference<com::sun::star::drawing::XDrawPage> const&, com::sun::star::uno::Reference<com::sun::star::animations::XAnimationNode> const&, Window*) ()
   from /usr/lib64/libreoffice/program/../program/libsdlo.so
#12 0x00007f9783a95111 in sd::SlideShow::StartInPlacePresentation() ()
   from /usr/lib64/libreoffice/program/../program/libsdlo.so
#13 0x00007f9783a96808 in sd::SlideShow::startWithArguments(com::sun::star::uno::Sequence<com::sun::star::beans::PropertyValue> const&) () from /usr/lib64/libreoffice/program/../program/libsdlo.so
Comment 9 David Tardon 2013-08-09 11:26:32 UTC

*** This bug has been marked as a duplicate of bug 67908 ***