| Summary: | artsdsp crashes | ||
|---|---|---|---|
| Product: | [Unmaintained] arts | Reporter: | Erik Schnetter <schnetter> |
| Component: | general | Assignee: | Stefan Westerfeld <stefan> |
| Status: | RESOLVED WORKSFORME | ||
| Severity: | critical | CC: | ana, brenno, casteyde.christian, dav, d_baron, ervalerio, esigra, greg_g, helio, kaval, kdebugs, kegary, linuxmail, luke-jr+kdebugs, mariausol, Psimer, sean, tma.klein |
| Priority: | VHI | ||
| Version: | unspecified | ||
| Target Milestone: | --- | ||
| Platform: | unspecified | ||
| OS: | Linux | ||
| Latest Commit: | Version Fixed In: | ||
| Sentry Crash Report: | |||
| Attachments: |
test program for double-init.
Experimental patch |
||
|
Description
Erik Schnetter
2003-08-13 16:48:57 UTC
I can't reproduce this, but I am definitely not surprised. *** Bug 64098 has been marked as a duplicate of this bug. *** *** Bug 64469 has been marked as a duplicate of this bug. *** *** Bug 65036 has been marked as a duplicate of this bug. *** I've seen a crash + backtrace like this in an SDL app. Platform is Gentoo/Athlon-tbird/gcc-3.3/-O3. It went away with recompiling arts in debug mode (-O3 -g3), so I suspect it is a compiler/optimizer bug. Both, crashing and now working version are KDE-3.2beta1 (arts-1.1.93). *** Bug 73779 has been marked as a duplicate of this bug. *** I got this bug after upgrading Mandrake 9.1 to 9.2 and using thac's multimedia kernel and other sound system apps ( http://rpm.nyvalls.se/index9.2.html). If I enable artsd then sure enough the crash goes away. I could reproduce this bug with xine, and with the little test case attached below. Compile it and run it in quick succession, half the time it will output a nice "warning: leaving MCOP Dispatcher and still 146 types alive.", the other half it will core dump. Valgrind tells me there's something dodgy going on with the init: 32173== Use of uninitialised value of size 16 ==32173== at 0x58C5D53: Arts::CpuInfoStartup::startup() (in /usr/lib/libartsflow.so.1.0.0) ==32173== by 0x13783C2: Arts::StartupManager::startup() (in /usr/lib/libmcop.so.1.0.0) ==32173== by 0x13577E7: Arts::Dispatcher::Dispatcher(Arts::IOManager*, Arts::Dispatcher::StartServer) (in /usr/lib/libmcop.so.1.0.0) ==32173== by 0x140E974: arts_backend_init (in /usr/lib/libartscbackend.so.0.0.0) ==32173== ==32173== Invalid read of size 4 ==32173== at 0x13783BC: Arts::StartupManager::startup() (in /usr/lib/libmcop.so.1.0.0) ==32173== by 0x13577E7: Arts::Dispatcher::Dispatcher(Arts::IOManager*, Arts::Dispatcher::StartServer) (in /usr/lib/libmcop.so.1.0.0) ==32173== by 0x140E974: arts_backend_init (in /usr/lib/libartscbackend.so.0.0.0) ==32173== by 0xD1B545: arts_init (in /usr/lib/libartsc.so.0.0.0) ==32173== Address 0x30202D4 is not stack'd, malloc'd or free'd Created attachment 5547 [details]
test program for double-init.
*** Bug 77830 has been marked as a duplicate of this bug. *** *** Bug 80322 has been marked as a duplicate of this bug. *** *** Bug 81537 has been marked as a duplicate of this bug. *** *** Bug 81795 has been marked as a duplicate of this bug. *** This bug is causing quite a few duplicates, both on the KDE and the GNOME bugzilla: http://bugzilla.gnome.org/show_bug.cgi?id=138830 *** Bug 86395 has been marked as a duplicate of this bug. *** *** Bug 94087 has been marked as a duplicate of this bug. *** Please report used compilers and versions Created attachment 8566 [details]
Experimental patch
Don't know if it makes any difference, but it might. Please try the patch.
*** Bug 96581 has been marked as a duplicate of this bug. *** *** Bug 93910 has been marked as a duplicate of this bug. *** It's just stupid that we still have this problem. It's very widespread, and all of my attempts to debug it resulted in valgrind/gdb crashes or the bug disappearing. Sounds like a stability issue to me. So let's do something. I got a stack backtrace similar to that in bug #93910 on my gentoo (AMD64) machine. It popped up after recompilation of kdelibs, but disappeared after I recompiled libsdl. Between the compilation of libsdl and the kdelibs compilation that broke arts gcc was upgraded from 3.3.? to 3.4.3. It looks like some incompatability between gcc 3.3 and gcc 3.4. Compiling all with gcc 3.3 works, just as compiling all with gcc 3.4. *** Bug 103470 has been marked as a duplicate of this bug. *** *** Bug 78764 has been marked as a duplicate of this bug. *** *** Bug 104226 has been marked as a duplicate of this bug. *** *** Bug 106047 has been marked as a duplicate of this bug. *** Using host libthread_db library "/lib/tls/libthread_db.so.1". [Thread debugging using libthread_db enabled] [New Thread -1238441856 (LWP 4174)] [KCrash handler] #3 0xb46cfa64 in X509_get_subject_name () from /usr/lib/libcrypto.so.0.9.7 #4 0xb768f9f8 in KOpenSSLProxy::X509_get_subject_name () from /usr/lib/libkio.so.4 #5 0xb767c768 in KSSLCertificate::getSubject () from /usr/lib/libkio.so.4 #6 0xb7682eb1 in KSSLPeerInfo::certMatchesAddress () from /usr/lib/libkio.so.4 #7 0xb6012ada in KSSLSocket::verifyCertificate () from /usr/lib/kde3/kopete_irc.so #8 0xb60119b3 in KSSLSocket::slotConnected () from /usr/lib/kde3/kopete_irc.so #9 0xb6014073 in KSSLSocket::qt_invoke () from /usr/lib/kde3/kopete_irc.so #10 0xb6b404a6 in QObject::activate_signal () from /usr/lib/libqt-mt.so.3 #11 0xb6b40354 in QObject::activate_signal () from /usr/lib/libqt-mt.so.3 #12 0xb71e7a8c in KExtendedSocket::connectionSuccess () from /usr/lib/libkdecore.so.4 #13 0xb71e6ce7 in KExtendedSocket::connectionEvent () from /usr/lib/libkdecore.so.4 #14 0xb71e5f9b in KExtendedSocket::socketActivityRead () from /usr/lib/libkdecore.so.4 #15 0xb71e7b68 in KExtendedSocket::qt_invoke () from /usr/lib/libkdecore.so.4 #16 0xb601403b in KSSLSocket::qt_invoke () from /usr/lib/kde3/kopete_irc.so #17 0xb6b4052c in QObject::activate_signal () from /usr/lib/libqt-mt.so.3 #18 0xb6b4068d in QObject::activate_signal () from /usr/lib/libqt-mt.so.3 #19 0xb6e7ef82 in QSocketNotifier::activated () from /usr/lib/libqt-mt.so.3 #20 0xb6b5ccd0 in QSocketNotifier::event () from /usr/lib/libqt-mt.so.3 #21 0xb6ae3bff in QApplication::internalNotify () from /usr/lib/libqt-mt.so.3 #22 0xb6ae31fe in QApplication::notify () from /usr/lib/libqt-mt.so.3 #23 0xb711df93 in KApplication::notify () from /usr/lib/libkdecore.so.4 #24 0xb6ad360a in QEventLoop::activateSocketNotifiers () from /usr/lib/libqt-mt.so.3 #25 0xb6a8cc33 in QEventLoop::processEvents () from /usr/lib/libqt-mt.so.3 #26 0xb6af5f58 in QEventLoop::enterLoop () from /usr/lib/libqt-mt.so.3 #27 0xb6af5e08 in QEventLoop::exec () from /usr/lib/libqt-mt.so.3 #28 0xb6ae3e51 in QApplication::exec () from /usr/lib/libqt-mt.so.3 #29 0x080699fd in ?? () #30 0xbffff9c0 in ?? () #31 0xbffff9a0 in ?? () #32 0xbffff9a0 in ?? () #33 0x00000000 in ?? () #34 0x080a815e in _IO_stdin_used () #35 0x00000001 in ?? () #36 0x080a94c0 in _IO_stdin_used () #37 0x080a7d76 in _IO_stdin_used () #38 0x080a7d8b in _IO_stdin_used () #39 0x080a7d5c in _IO_stdin_used () #40 0x00000000 in ?? () #41 0xbffff9a0 in ?? () #42 0xb6f86680 in vtable for QGArray () from /usr/lib/libqt-mt.so.3 #43 0x082339c8 in ?? () #44 0x0805ec15 in ?? () #45 0xb6646e54 in ?? () from /usr/lib/libstdc++.so.5 #46 0xb6f86680 in vtable for QGArray () from /usr/lib/libqt-mt.so.3 #47 0x080f0b58 in ?? () #48 0xbffff9e8 in ?? () #49 0xb66252ee in operator new () from /usr/lib/libstdc++.so.5 #50 0xb64417f8 in __libc_start_main () from /lib/tls/libc.so.6 #51 0xb6563fcc in ?? () from /lib/tls/libc.so.6 Not much activity on this for a while. Can anyone still reproduce it with KDE 3.5? I've tried to load that website on KDE 3.5.4. I've had no crash. Maybe it was a problem related to some plugin? OK, then. Presumably it's fixed. Please reopen if anyone is still seeing this problem in 3.5 |