[Akonadi] [Bug 363881] akonadi_control fails with memory corruption
https://bugs.kde.org/show_bug.cgi?id=363881 Andreas Schneider changed: What|Removed |Added Resolution|--- |FIXED Status|REPORTED|RESOLVED -- You are receiving this mail because: You are watching all bug changes.
[Akonadi] [Bug 363881] akonadi_control fails with memory corruption
https://bugs.kde.org/show_bug.cgi?id=363881 --- Comment #7 from Andreas Schneider --- András, you might want to look at the QT bug I created. Hopefully this happend for you with QT 5.7 :) https://bugreports.qt.io/browse/QTBUG-53957 -- You are receiving this mail because: You are watching all bug changes.
[Akonadi] [Bug 363881] akonadi_control fails with memory corruption
https://bugs.kde.org/show_bug.cgi?id=363881 András Manţia changed: What|Removed |Added CC||[email protected] --- Comment #6 from András Manţia --- akonadiconsole also crashes in a similar way after a while (can't run in valgrind so far as it crashes in webengine that way on startup): #0 0x7fffeb6f20c7 in __GI_raise (sig=sig@entry=6) at ../nptl/sysdeps/unix/sysv/linux/raise.c:56 #1 0x7fffeb6f3478 in __GI_abort () at abort.c:78 #2 0x7fffeb72f784 in __libc_message (do_abort=do_abort@entry=2, fmt=fmt@entry=0x7fffeb821b78 "*** Error in `%s': %s: 0x%s ***\n") at ../sysdeps/posix/libc_fatal.c:175 #3 0x7fffeb735026 in malloc_printerr (action=3, str=0x7fffeb822240 "malloc(): memory corruption (fast)", ptr=, ar_ptr=) at malloc.c:5037 #4 0x7fffeb7375d5 in _int_malloc (av=av@entry=0x7fffbc20, bytes=bytes@entry=70) at malloc.c:3392 #5 0x7fffeb73895c in __GI___libc_malloc (bytes=70) at malloc.c:2908 #6 0x7fffec3b286f in QArrayData::allocate(unsigned long, unsigned long, unsigned long, QFlags) (objectSize=objectSize@entry=2, alignment=alignment@entry=8, capacity=capacity@entry=23, options=..., options@entry=...) at tools/qarraydata.cpp:114 #7 0x7fffec434b23 in QString::QString(int, Qt::Initialization) (options=..., capacity=23) at ../../src/corelib/tools/qarraydata.h:222 #8 0x7fffec434b23 in QString::QString(int, Qt::Initialization) (this=0x7fffc3ffe570, size=22) at tools/qstring.cpp:1565 #9 0x7fffec5be29d in QUtf8::convertToUnicode(char const*, int) (chars=chars@entry=0x137ad0c "Akonadi::ItemCreateJob", len=22) at codecs/qutfcodec.cpp:264 #10 0x7fffec43b3ce in QString::fromUtf8_helper(char const*, int) (str=str@entry=0x137ad0c "Akonadi::ItemCreateJob", size=) at tools/qstring.cpp:4785 #11 0x7fffeeafbf2d in QDBusDemarshaller::toVariantInternal() (size=-1, str=0x137ad0c "Akonadi::ItemCreateJob") at ../../src/corelib/tools/qstring.h:543 #12 0x7fffeeafbf2d in QDBusDemarshaller::toVariantInternal() (this=0x7fffc3ffe610) at qdbusdemarshaller.cpp:133 #13 0x7fffeeafbf2d in QDBusDemarshaller::toVariantInternal() (this=this@entry=0x7fffc3ffe620) at qdbusdemarshaller.cpp:261 #14 0x7fffeead21db in QDBusMessagePrivate::fromDBusMessage(DBusMessage*, QFlags) (dmsg=, capabilities=...) at qdbusmessage.cpp:251 #15 0x7fffeeaca8ce in qDBusSignalFilter(DBusConnection*, DBusMessage*, void*) (connection=, message=, data=0x7fffbc0030f0) at qdbusintegrator.cpp:510 #16 0x7fffdcd9e6a6 in dbus_connection_dispatch (connection=0x7fffbc00e140) at dbus-connection.c:4677 #17 0x7fffeeaca4f1 in QDBusConnectionPrivate::doDispatch() (connection=) at qdbus_symbols_p.h:195 #18 0x7fffeeaca4f1 in QDBusConnectionPrivate::doDispatch() (this=this@entry=0x7fffbc0030f0) at qdbusintegrator.cpp:1177 #19 0x7fffeeaca7e9 in QDBusConnectionPrivate::socketRead(int) (this=0x7fffbc0030f0, fd=17) at qdbusintegrator.cpp:1203 #20 0x7fffec590cfc in QMetaObject::activate(QObject*, int, int, void**) (a=0x7fffc3ffea50, r=0x7fffbc0030f0, this=0x7fffbc010300) at ../../src/corelib/kernel/qobject_impl.h:130 #21 0x7fffec590cfc in QMetaObject::activate(QObject*, int, int, void**) (sender=sender@entry=0x7fffbc00deb0, signalOffset=, local_signal_index=local_signal_index@entry=0, argv=argv@entry=0x7fffc3ffea50) at kernel/qobject.cpp:3723 #22 0x7fffec591247 in QMetaObject::activate(QObject*, QMetaObject const*, int, void**) (sender=sender@entry=0x7fffbc00deb0, m=m@entry=0x7fffec9ad620 , local_signal_index=local_signal_index@entry=0, argv=argv@entry=0x7fffc3ffea50) at kernel/qobject.cpp:3602 #23 0x7fffec60816e in QSocketNotifier::activated(int, QSocketNotifier::QPrivateSignal) (this=this@entry=0x7fffbc00deb0, _t1=17) at .moc/moc_qsocketnotifier.cpp:135 #24 0x7fffec59d7f9 in QSocketNotifier::event(QEvent*) (this=0x7fffbc00deb0, e=0x7fffc3ffeb90) at kernel/qsocketnotifier.cpp:266 #25 0x7fffec5685f1 in QCoreApplication::notifyInternal2(QObject*, QEvent*) (event=0x7fffc3ffeb90, receiver=0x7fffbc00deb0) at kernel/qcoreapplication.cpp:1063 #26 0x7fffec5685f1 in QCoreApplication::notifyInternal2(QObject*, QEvent*) (receiver=0x7fffbc00deb0, event=event@entry=0x7fffc3ffeb90) at kernel/qcoreapplication.cpp:987 #27 0x7fffec5b7b3e in socketNotifierSourceDispatch(GSource*, GSourceFunc, gpo
[Akonadi] [Bug 363881] akonadi_control fails with memory corruption
https://bugs.kde.org/show_bug.cgi?id=363881 Erick Osorio changed: What|Removed |Added CC||[email protected] -- You are receiving this mail because: You are watching all bug changes.
[Akonadi] [Bug 363881] akonadi_control fails with memory corruption
https://bugs.kde.org/show_bug.cgi?id=363881 --- Comment #5 from Eric MSP Veith --- It took me a bit of fiddeling and much patience to encouter this bug with GDB. But here it is. I've got a backtrace and a core dump. The core dump file is quite big, so I've uploaded it to an external source (6,7 MB > 4000kB; sorry, self-signed cert...): - https://oc.veith-m.de/index.php/s/fK7kcbw7mYQORPg - https://oc.veith-m.de/index.php/s/pLRB5I0VBPE2rdl I'm not sure whether this is actually a KDE-specific bug? Perhaps only something in KDE triggers it, I don't see any KDE libs in the backtrace. But I'm by no means an expert here; I hope this helps somebody. Here are the last few lines before emitting a SIGSEGV: [...] QDBusObjectPath Akonadi::Server::NotificationManager::subscribe(const QString&, bool) Akonadi::Server::NotificationManager(0x85b830) "akonadi_followupreminder_agent_28779_PP6Fkp" false QDBusObjectPath Akonadi::Server::NotificationManager::subscribe(const QString&, bool) Akonadi::Server::NotificationManager(0x85b830) "akonadi_ical_resource_3_28781_zVEkXR" false QDBusObjectPath Akonadi::Server::NotificationManager::subscribe(const QString&, bool) Akonadi::Server::NotificationManager(0x85b830) "akonadi_davgroupware_resource_2_28778_uaMY6p" false QDBusObjectPath Akonadi::Server::NotificationManager::subscribe(const QString&, bool) Akonadi::Server::NotificationManager(0x85b830) "akonadi_archivemail_agent_28775_ubx5xw" false QDBusObjectPath Akonadi::Server::NotificationManager::subscribe(const QString&, bool) Akonadi::Server::NotificationManager(0x85b830) "akonadi_mailfilter_agent_28801_MpmVCD" false QDBusObjectPath Akonadi::Server::NotificationManager::subscribe(const QString&, bool) Akonadi::Server::NotificationManager(0x85b830) "akonadi_maildir_resource_5_28794_K6eQWS" false QDBusObjectPath Akonadi::Server::NotificationManager::subscribe(const QString&, bool) Akonadi::Server::NotificationManager(0x85b830) "akonadi_maildispatcher_agent_28799_xR5u02" false QDBusObjectPath Akonadi::Server::NotificationManager::subscribe(const QString&, bool) Akonadi::Server::NotificationManager(0x85b830) "akonadi_imap_resource_6_28790_7aAYVF" false QDBusObjectPath Akonadi::Server::NotificationManager::subscribe(const QString&, bool) Akonadi::Server::NotificationManager(0x85b830) "akonadi_sendlater_agent_28806_FPSn8U" false Cannot pause an inactive timer Cannot pause an inactive timer QDBusObjectPath Akonadi::Server::NotificationManager::subscribe(const QString&, bool) Akonadi::Server::NotificationManager(0x85b830) "akonadi_imap_resource_5_28786_7wZJ5V" false QDBusObjectPath Akonadi::Server::NotificationManager::subscribe(const QString&, bool) Akonadi::Server::NotificationManager(0x85b830) "akonadi_newmailnotifier_agent_28804_e4MqgF" false QDBusObjectPath Akonadi::Server::NotificationManager::subscribe(const QString&, bool) Akonadi::Server::NotificationManager(0x85b830) "akonadi_notes_agent_28805_lXx6g1" false QDBusObjectPath Akonadi::Server::NotificationManager::subscribe(const QString&, bool) Akonadi::Server::NotificationManager(0x85b830) "akonadi_invitations_agent_0_28792_q6189K" false QDBusObjectPath Akonadi::Server::NotificationManager::subscribe(const QString&, bool) Akonadi::Server::NotificationManager(0x85b830) "akonadi_indexing_agent_28791_GC7RfJ" false QDBusObjectPath Akonadi::Server::NotificationManager::subscribe(const QString&, bool) Akonadi::Server::NotificationManager(0x85b830) "akonadi_migration_agent_28803_YOhJyk" false QDBusObjectPath Akonadi::Server::NotificationManager::subscribe(const QString&, bool) Akonadi::Server::NotificationManager(0x85b830) "akonadi_ical_resource_4_28783_srS5ST" false *** Error in `/usr/bin/akonadi_control': malloc(): memory corruption (fast): 0x7fffd8018e3c *** [New Thread 0x7fffde0b0700 (LWP 28765)] [New Thread 0x7fffecb6d700 (LWP 28764)] Program received signal SIGSEGV, Segmentation fault. 0x75aebb2e in _dbus_counter_notify (counter=0x7fffd800dae0) at dbus-resources.c:187 187 dbus-resources.c: Datei oder Verzeichnis nicht gefunden. (gdb) bt #0 0x75aebb2e in _dbus_counter_notify (counter=0x7fffd800dae0) at dbus-resources.c:187 #1 0x75ae4bc2 in free_counter (element=0x7fffd800dae0, data=0x7fffd800f0e0) at dbus-message.c:639 #2 0x75ae50d2 in dbus_message_unref (message=0x7fffd800f0e0) at dbus-message.c:661 #3 0x75ae50d2 in dbus_message_unref (message=0x7fffd800f0e0) at dbus-message.c:1707 #4 0x776b72a7 in () at /usr/lib64/libQt5DBus.so.5 #5 0x776b747b in QDBusMessage::~QDBusMessage() () at /usr/lib64/libQt5DBus.so.5 #6 0x776a970c in () at /usr/lib64/libQt5DBus.so.5 #7 0x776a9779 in () at /usr/lib64/libQt5DBus.so.5 #8 0x76ae7da3 in QCoreApplicationPrivate::sendPostedEvents(QObject*, int, QThreadData*) () at /usr/lib64/libQt5Core.so.5 #9 0x76b36ca3 in () at /usr/lib64/libQt5Core.so.5 #10 0x73fd0c84 in g_main_context_dispatch () at /usr/lib64/libglib-2.0.so.0 #11 0x000
[Akonadi] [Bug 363881] akonadi_control fails with memory corruption
https://bugs.kde.org/show_bug.cgi?id=363881 [email protected] changed: What|Removed |Added CC||[email protected] -- You are receiving this mail because: You are watching all bug changes.
[Akonadi] [Bug 363881] akonadi_control fails with memory corruption
https://bugs.kde.org/show_bug.cgi?id=363881 Eric MSP Veith changed: What|Removed |Added CC||[email protected] --- Comment #4 from Eric MSP Veith --- I am also facing this bug. I have tried to make akonadi_control create a core dump when it crashes, but adding "ulimit -c 52428800" to ~/.Xprofile didn't help. Andreas, how did you attach valgrind/gdb to akonadi_control? -- You are receiving this mail because: You are watching all bug changes.
[Akonadi] [Bug 363881] akonadi_control fails with memory corruption
https://bugs.kde.org/show_bug.cgi?id=363881 --- Comment #3 from Andreas Schneider --- Yes, after further debugging it seems to be an issue in Qt serializing a buffer. I opened https://bugreports.qt.io/browse/QTBUG-53957 but I'm not able to provide any useful valgrind output till now. -- You are receiving this mail because: You are watching all bug changes.
[Akonadi] [Bug 363881] akonadi_control fails with memory corruption
https://bugs.kde.org/show_bug.cgi?id=363881 O. Puetz changed: What|Removed |Added CC||[email protected] --- Comment #2 from O. Puetz --- I can confirm this malloc crash happening in different akonadi modules at first. Running a test system without akonadi, kded crashed with the same malloc error after a while (run time varies). Test system config and tested software versions: - AMD Phenom II with 16GB RAM - openSUSE Leap 42.1 - Qt 5.6.0 and now 5.6.1 - KDE Frameworks 5.2.1 and now 5.2.2 - Plasma 5.6.3 and now 5.6.4 - KDE:Applications 16.04.1 To me it seems this is no KDE PIM bug, it should be in some of the underlying software layers. And may be related with specific hardware config. Wiping /tmp seems to help at first, but it just allowed akonadi to start successfully and after an undetermined time a malloc error appeared again (e.g. in akonadi imap resource). BTW, an identical setup in a VirtualBox machine runs without these malloc crashes. Some other users are faced with the same or similar problem, see https://forums.opensuse.org/showthread.php/516850-After-Updating-KDE-FW-to-5-20-and-Plasma-to-5-6-(with-Qt-to-5-6-as-well)-Akonadi-quot-doesn-t-work-quot -- You are receiving this mail because: You are watching all bug changes.
[Akonadi] [Bug 363881] akonadi_control fails with memory corruption
https://bugs.kde.org/show_bug.cgi?id=363881 --- Comment #1 from Andreas Schneider --- I think there is a race condition. If I start akonadi and then kontact, akonadi 8 out of 10 times doesn't work and react. I have to kill all akonadi processes. However if I run start akonadi_control with valgrind, everything starts to work. -- You are receiving this mail because: You are watching all bug changes.
