| Summary: | Faces Detection crashes digiKam | ||
|---|---|---|---|
| Product: | [Applications] digikam | Reporter: | Cliff Albert <cliff> |
| Component: | Faces-Detection | Assignee: | Digikam Developers <digikam-bugs-null> |
| Status: | RESOLVED FIXED | ||
| Severity: | crash | CC: | caulier.gilles, kdebugs, metzpinguin, paazmaya |
| Priority: | NOR | ||
| Version: | 7.0.0 | ||
| Target Milestone: | --- | ||
| Platform: | Appimage | ||
| OS: | Linux | ||
| Latest Commit: | Version Fixed In: | 7.2.0 | |
| Sentry Crash Report: | |||
|
Description
Cliff Albert
2020-05-05 07:49:22 UTC
Yes, the appropriate Ubuntu version and let me guess, you probably have an Intel Core 5. Maik No this is a Xeon X5670 (6 core 2.93Ghz), from the Gulfton/Westmere EP architecture. Equivalent to the Core i7 970-990 series. I indeed read some bug reports concerning the Core i5 having something to do with the running kernel. Currently running: Linux porkstation 5.3.0-46-generic #38-Ubuntu SMP Fri Mar 27 17:37:05 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux I've planned to upgrade the machine to 20.04 today, i'll give digikam a new test afterwards Upgraded to Ubuntu 20.04 LTS (with kernel Linux porkstation 5.4.0-29-generic #33-Ubuntu SMP Wed Apr 29 14:32:27 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux) The following error appears after which a segfault occurs: Digikam::DImg::getUniqueHashV2: DImg::getUniqueHash called without originalFilePath property set! --Type <RET> for more, q to quit, c to continue without paging-- Thread 27 "Thread (pooled)" received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fff765ed700 (LWP 4655)] 0x0000000003520f90 in ?? () (gdb) bt #0 0x0000000003520f90 in ?? () #1 0x00007ffff747413d in Digikam::LoadingTask::progressInfo (this=0x3ed77b0, progress=0.64755553) at /b/dktemp/digikam-master/core/libs/threadimageio/fileio/loadsavetask.cpp:63 #2 0x00007fffc8aba5dd in DigikamJPEGDImgPlugin::DImgJPEGLoader::load (this=0x7fff6c0d6d80, filePath=..., observer=0x3ed77c0) at /b/dktemp/digikam-master/core/dplugins/dimg/jpeg/dimgjpegloader_load.cpp:434 #3 0x00007ffff738869a in Digikam::DImg::load (this=this@entry=0x3ed7998, filePath=..., loadFlagsInt=loadFlagsInt@entry=63, observer=observer@entry=0x3ed77c0, rawDecodingSettings=...) at /b/dktemp/digikam-master/core/libs/dimg/dimg_fileio.cpp:132 #4 0x00007ffff73896f0 in Digikam::DImg::load (this=this@entry=0x3ed7998, filePath=..., observer=observer@entry=0x3ed77c0, rawDecodingSettings=...) at /b/dktemp/digikam-master/core/libs/dimg/dimg_fileio.cpp:62 #5 0x00007ffff74528c1 in Digikam::PreviewLoadingTask::execute (this=0x3ed77b0) at /b/dktemp/digikam-master/core/libs/threadimageio/preview/previewtask.cpp:314 #6 0x00007ffff746831e in Digikam::LoadSaveThread::run (this=0x21c9f70) at /b/dktemp/digikam-master/core/libs/threadimageio/fileio/loadsavethread.cpp:134 #7 0x00007ffff748d5ce in Digikam::DynamicThread::Private::run (this=0x7fffcc00bcb0) at /b/dktemp/digikam-master/core/libs/threads/dynamicthread.cpp:191 #8 0x00007ffff41dbd87 in ?? () from /tmp/.mount_digikaAeveBn/usr/lib/libQt5Core.so.5 #9 0x00007ffff41d8335 in ?? () from /tmp/.mount_digikaAeveBn/usr/lib/libQt5Core.so.5 #10 0x00007ffff3b84609 in start_thread (arg=<optimized out>) at pthread_create.c:477 #11 0x00007ffff3aab103 in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95 This looks the same as https://bugs.kde.org/show_bug.cgi?id=415867#c22, after which Maik Qualmann indicated it was core i5 related. We might probably add Xeon X56xx also to the list. As I said, these crashes are very strange. We only watch them on Ubuntu based distributions related to Intel Core 5 or 7. This message would only appear if there was a code error, under normal circumstances it can never be issued. An original file path attribute is always set. ---- Digikam::DImg::getUniqueHashV2: DImg::getUniqueHash called without originalFilePath property set! ---- And here we are again in the wrong overloaded function, in principle impossible. ---- #1 0x00007ffff747413d in Digikam::LoadingTask::progressInfo (this=0x3ed77b0, progress=0.64755553) at /b/dktemp/digikam-master/core/libs/threadimageio/fileio/loadsavetask.cpp:63 ---- Maik Having the same(?) issue when scanning for faces on some folders. - digikam-7.0.0-rc-20200516T134714-x86-64.appimage - Ubuntu 20.04 LTS - Intel(R) Core(TM) i7-8550U CPU @ 1.80GHz Backtrace: Thread 48 "Thread (pooled)" received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fff2d7fa700 (LWP 742579)] 0x00007ffff744cc2b in ?? () from /tmp/.mount_digikaI33g9h/usr/lib/libdigikamcore.so.7.0.0 (gdb) bt #0 0x00007ffff744cc2b in ?? () from /tmp/.mount_digikaI33g9h/usr/lib/libdigikamcore.so.7.0.0 #1 0x00007ffff7462a3e in Digikam::LoadSaveThread::run() () from /tmp/.mount_digikaI33g9h/usr/lib/libdigikamcore.so.7.0.0 #2 0x00007ffff7487c5e in ?? () from /tmp/.mount_digikaI33g9h/usr/lib/libdigikamcore.so.7.0.0 #3 0x00007ffff41d7d87 in ?? () from /tmp/.mount_digikaI33g9h/usr/lib/libQt5Core.so.5 #4 0x00007ffff41d4335 in ?? () from /tmp/.mount_digikaI33g9h/usr/lib/libQt5Core.so.5 #5 0x00007ffff3b81609 in start_thread (arg=<optimized out>) at pthread_create.c:477 #6 0x00007ffff3aa8103 in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95 digikam-7.0.0-rc-20200516T134714-x86-64.appimage ==> use the "-debug" bundle version and reproduce the crash to have a suitable backtrace. Gilles Caulier Git commit 880b4c0faa511364135deed61e68e031b8ddfe2e by Maik Qualmann. Committed on 14/06/2020 at 09:32. Pushed by mqualmann into branch 'master'. do not use static for this DImg function Related: bug 422735, bug 418939 M +3 -3 core/libs/dimg/dimg.h M +7 -8 core/libs/dimg/dimg_metadata.cpp https://invent.kde.org/graphics/digikam/commit/880b4c0faa511364135deed61e68e031b8ddfe2e Git commit b66eb0a0556388fc532fcc5bba7a40ed6eda2fc1 by Maik Qualmann. Committed on 12/07/2020 at 06:58. Pushed by mqualmann into branch 'master'. disable OpenCL for everyone An analysis with Heaptrack shows that we have a big memory leak with active OpenCL. Related: bug 423632, bug 416266, bug 420411 M +10 -0 core/app/main/main.cpp M +0 -5 project/bundles/appimage/data/AppRun https://invent.kde.org/graphics/digikam/commit/b66eb0a0556388fc532fcc5bba7a40ed6eda2fc1 *** Bug 424133 has been marked as a duplicate of this bug. *** Hi, Can you check if this crash still exist with last weekly AppImage build available here : https://files.kde.org/digikam/ Thanks in advance After testing for 2 days I haven't encountered the problem. I'll do some more this weekend. But so far it's all good. Thanks for the feedback. I close this file now Gilles Caulier Unfortunately the same error is popping up again with digikam-7.2.0-beta1-20201013T065929-x86-64-debug.appimage running on Ubuntu 20.04.1 (Linux porkstation 5.4.0-48-generic #52-Ubuntu SMP Thu Sep 10 10:58:49 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux) [New Thread 0x7ffefdffb700 (LWP 19929)] [Thread 0x7fffb508a700 (LWP 19928) exited] [Thread 0x7ffefdffb700 (LWP 19929) exited] Digikam::DImg::getUniqueHashV2: DImg::getUniqueHash called without originalFilePath property set! Thread 67 "Thread (pooled)" received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fff5f015700 (LWP 17252)] 0x00007ffff41e13e1 in QMapDataBase::rebalance(QMapNodeBase*) () from /tmp/.mount_digikay5n2NY/usr/lib/libQt5Core.so.5 (gdb) bt #0 0x00007ffff41e13e1 in QMapDataBase::rebalance(QMapNodeBase*) () from /tmp/.mount_digikay5n2NY/usr/lib/libQt5Core.so.5 #1 0x00007ffff41e19c4 in QMapDataBase::createNode(int, int, QMapNodeBase*, bool) () from /tmp/.mount_digikay5n2NY/usr/lib/libQt5Core.so.5 #2 0x00007ffff73c6183 in QMapData<QString, QVariant>::createNode (left=<optimized out>, parent=0x7fff6191f4e0, v=..., k=..., this=<optimized out>) at ././/include/QtCore/qmap.h:223 #3 QMap<QString, QVariant>::insert (avalue=..., akey=..., this=0x7fff5401b540) at ././/include/QtCore/qmap.h:719 #4 Digikam::DImg::setAttribute (this=<optimized out>, key=..., value=...) at /b/dktemp/digikam-master/core/libs/dimg/dimg_props.cpp:235 #5 0x00007ffff725a519 in Digikam::DImgLoader::imageSetAttribute (this=<optimized out>, key=..., value=...) at /b/dktemp/digikam-master/core/libs/dimg/loaders/dimgloader.cpp:142 #6 0x00007ffff725ba42 in Digikam::DImgLoader::readMetadata (this=0x7fff619261c0, filePath=...) at /b/dktemp/digikam-master/core/libs/dimg/loaders/dimgloader.cpp:233 #7 0x00007fffc90c3ff9 in DigikamJPEGDImgPlugin::DImgJPEGLoader::load (this=0x7fff619261c0, filePath=..., observer=0xaa53840) at /b/dktemp/digikam-master/core/dplugins/dimg/jpeg/dimgjpegloader_load.cpp:56 #8 0x00007ffff73bd66a in Digikam::DImg::load (this=this@entry=0xaa53a10, filePath=..., loadFlagsInt=loadFlagsInt@entry=63, observer=observer@entry=0xaa53840, rawDecodingSettings=...) at /b/dktemp/digikam-master/core/libs/dimg/dimg_fileio.cpp:132 #9 0x00007ffff73be6c0 in Digikam::DImg::load (this=this@entry=0xaa53a10, filePath=..., observer=observer@entry=0xaa53840, rawDecodingSettings=...) at /b/dktemp/digikam-master/core/libs/dimg/dimg_fileio.cpp:62 #10 0x00007ffff74890ae in Digikam::PreviewLoadingTask::execute (this=0xaa53830) at /b/dktemp/digikam-master/core/libs/threadimageio/preview/previewtask.cpp:310 #11 0x00007ffff749f07e in Digikam::LoadSaveThread::run (this=0x2398ab0) at /b/dktemp/digikam-master/core/libs/threadimageio/fileio/loadsavethread.cpp:154 #12 0x00007ffff74c370e in Digikam::DynamicThread::Private::run (this=0xa4acef0) at /b/dktemp/digikam-master/core/libs/threads/dynamicthread.cpp:191 #13 0x00007ffff41a4d87 in ?? () from /tmp/.mount_digikay5n2NY/usr/lib/libQt5Core.so.5 #14 0x00007ffff41a1335 in ?? () from /tmp/.mount_digikay5n2NY/usr/lib/libQt5Core.so.5 #15 0x00007ffff3b31609 in start_thread (arg=<optimized out>) at pthread_create.c:477 #16 0x00007ffff3a58293 in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95 Git commit d37b969a2fcd09ce97707c48c5c23730d97e069c by Maik Qualmann. Committed on 15/10/2020 at 17:12. Pushed by mqualmann into branch 'master'. this bug after the code revert could be the problem M +1 -1 core/libs/threadimageio/fileio/loadsavetask.cpp M +1 -1 core/libs/threadimageio/preview/previewtask.cpp M +1 -1 core/libs/threadimageio/thumb/thumbnailtask.cpp https://invent.kde.org/graphics/digikam/commit/d37b969a2fcd09ce97707c48c5c23730d97e069c I don't know if the patch already made it to digikam-7.2.0-beta1-20201016T092735-x86-64-debug.appimage, but here is another slightly different backtrace: Thread 29 "Thread (pooled)" received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fff74dea700 (LWP 519207)] 0x00007ffff423e9d4 in ?? () from /tmp/.mount_digikaj6ug8P/usr/lib/libQt5Core.so.5 (gdb) bt #0 0x00007ffff423e9d4 in ?? () from /tmp/.mount_digikaj6ug8P/usr/lib/libQt5Core.so.5 #1 0x00007ffff42431d0 in operator<(QString const&, QString const&) () from /tmp/.mount_digikaj6ug8P/usr/lib/libQt5Core.so.5 #2 0x00007ffff73c8700 in qMapLessThanKey<QString> (key2=..., key1=...) at ././/include/QtCore/qmap.h:68 #3 QMapNode<QString, QVariant>::lowerBound (akey=..., this=<optimized out>) at ././/include/QtCore/qmap.h:152 #4 QMapData<QString, QVariant>::findNode (this=<optimized out>, akey=...) at ././/include/QtCore/qmap.h:284 #5 0x00007ffff73c4a13 in QMap<QString, QVariant>::contains (akey=..., this=<optimized out>) at ././/include/QtCore/qmap.h:693 #6 Digikam::DImg::hasAttribute (this=this@entry=0xa71c9c0, key=...) at /b/dktemp/digikam-master/core/libs/dimg/dimg_props.cpp:254 #7 0x00007ffff73c08a9 in Digikam::DImg::getUniqueHashV2 (this=0xa71c9c0) at /b/dktemp/digikam-master/core/libs/dimg/dimg_metadata.cpp:76 #8 0x00007ffff73c2135 in Digikam::DImg::createHistoryImageId (this=<optimized out>, filePath=..., type=type@entry=Digikam::HistoryImageId::Current) at /b/dktemp/digikam-master/core/libs/dimg/dimg_metadata.cpp:328 #9 0x00007ffff725b9dd in Digikam::DImgLoader::readMetadata (this=0x7fff5807d7e0, filePath=...) at /b/dktemp/digikam-master/core/libs/dimg/loaders/dimgloader.cpp:229 #10 0x00007fffc90c3ff9 in DigikamJPEGDImgPlugin::DImgJPEGLoader::load (this=0x7fff5807d7e0, filePath=..., observer=0xa71c7f0) at /b/dktemp/digikam-master/core/dplugins/dimg/jpeg/dimgjpegloader_load.cpp:56 #11 0x00007ffff73bd66a in Digikam::DImg::load (this=this@entry=0xa71c9c0, filePath=..., loadFlagsInt=loadFlagsInt@entry=63, observer=observer@entry=0xa71c7f0, rawDecodingSettings=...) at /b/dktemp/digikam-master/core/libs/dimg/dimg_fileio.cpp:132 #12 0x00007ffff73be6c0 in Digikam::DImg::load (this=this@entry=0xa71c9c0, filePath=..., observer=observer@entry=0xa71c7f0, rawDecodingSettings=...) at /b/dktemp/digikam-master/core/libs/dimg/dimg_fileio.cpp:62 #13 0x00007ffff74890ae in Digikam::PreviewLoadingTask::execute (this=0xa71c7e0) at /b/dktemp/digikam-master/core/libs/threadimageio/preview/previewtask.cpp:310 #14 0x00007ffff749f07e in Digikam::LoadSaveThread::run (this=0xa670050) at /b/dktemp/digikam-master/core/libs/threadimageio/fileio/loadsavethread.cpp:154 #15 0x00007ffff74c36fe in Digikam::DynamicThread::Private::run (this=0xa6678c0) at /b/dktemp/digikam-master/core/libs/threads/dynamicthread.cpp:191 #16 0x00007ffff41a4d87 in ?? () from /tmp/.mount_digikaj6ug8P/usr/lib/libQt5Core.so.5 #17 0x00007ffff41a1335 in ?? () from /tmp/.mount_digikaj6ug8P/usr/lib/libQt5Core.so.5 #18 0x00007ffff3b31609 in start_thread (arg=<optimized out>) at pthread_create.c:477 #19 0x00007ffff3a58293 in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95 The libjpeg is not 100% thread safe, at least the error handling can lead to a crash. Tonight I will test the DImg Jpeg loader with a mutex. Maik Git commit 4cf9c413ee1834b9c199f5e762a8fca5b3f46c5e by Maik Qualmann. Committed on 20/10/2020 at 05:52. Pushed by mqualmann into branch 'master'. only for a test protect jpeg loader with a mutex M +4 -0 core/dplugins/dimg/jpeg/dimgjpegloader_load.cpp M +2 -0 core/dplugins/dimg/jpeg/dimgjpegplugin.cpp M +3 -0 core/dplugins/dimg/jpeg/dimgjpegplugin.h https://invent.kde.org/graphics/digikam/commit/4cf9c413ee1834b9c199f5e762a8fca5b3f46c5e Here is another backtrace. This one has been made with digikam-7.2.0-beta1-20201025T154101-x86-64-debug.appimage The machine i'm using has been updated to Ubuntu 20.10: Linux porkstation 5.8.0-25-generic #26-Ubuntu SMP Thu Oct 15 10:30:38 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux I'm guessing something is also wrong with the debugging image of the 20201025 build as gdb gives a file does not exist message: Thread 42 "Thread (pooled)" received signal SIGSEGV, Segmentation fault. [Switching to Thread 0x7fff9f7fe640 (LWP 25961)] 0x00007ffff749dc17 in Digikam::LoadingTask::progressInfo (this=0xe2fcf70, progress=0.300000012) at /b/dktemp/digikam-master/core/libs/threadimageio/fileio/loadsavetask.cpp:100 100 /b/dktemp/digikam-master/core/libs/threadimageio/fileio/loadsavetask.cpp: No such file or directory. (gdb) bt #0 0x00007ffff749dc17 in Digikam::LoadingTask::progressInfo (this=0xe2fcf70, progress=0.300000012) at /b/dktemp/digikam-master/core/libs/threadimageio/fileio/loadsavetask.cpp:100 #1 0x00007fffc8eac945 in DigikamJPEGDImgPlugin::DImgJPEGLoader::load (this=0x7fff882da0f0, filePath=..., observer=0xe2fcf80) at /b/dktemp/digikam-master/core/dplugins/dimg/jpeg/dimgjpegloader_load.cpp:443 #2 0x00007ffff73af3aa in Digikam::DImg::load (this=this@entry=0xe2fd150, filePath=..., loadFlagsInt=loadFlagsInt@entry=63, observer=observer@entry=0xe2fcf80, rawDecodingSettings=...) at /b/dktemp/digikam-master/core/libs/dimg/dimg_fileio.cpp:132 #3 0x00007ffff73b0400 in Digikam::DImg::load (this=this@entry=0xe2fd150, filePath=..., observer=observer@entry=0xe2fcf80, rawDecodingSettings=...) at /b/dktemp/digikam-master/core/libs/dimg/dimg_fileio.cpp:62 #4 0x00007ffff747d02e in Digikam::PreviewLoadingTask::execute (this=0xe2fcf70) at /b/dktemp/digikam-master/core/libs/threadimageio/preview/previewtask.cpp:310 #5 0x00007ffff7492ffe in Digikam::LoadSaveThread::run (this=0xe2796c0) at /b/dktemp/digikam-master/core/libs/threadimageio/fileio/loadsavethread.cpp:154 #6 0x00007ffff74b796e in Digikam::DynamicThread::Private::run (this=0xde6e7b0) at /b/dktemp/digikam-master/core/libs/threads/dynamicthread.cpp:191 #7 0x00007ffff4180d87 in ?? () from /tmp/.mount_digikaUIYI90/usr/lib/libQt5Core.so.5 #8 0x00007ffff417d335 in ?? () from /tmp/.mount_digikaUIYI90/usr/lib/libQt5Core.so.5 #9 0x00007ffff3b0e590 in start_thread (arg=0x7fff9f7fe640) at pthread_create.c:463 #10 0x00007ffff3a34223 in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95 Git commit cdef325da3796effb0699ed9654d58c20eda7108 by Maik Qualmann. Committed on 26/10/2020 at 11:35. Pushed by mqualmann into branch 'master'. add extra debug M +2 -0 core/libs/threadimageio/fileio/loadsavetask.cpp M +1 -0 core/libs/threadimageio/preview/previewtask.cpp https://invent.kde.org/graphics/digikam/commit/cdef325da3796effb0699ed9654d58c20eda7108 It has been a while, but i've hit the same issue again. Release: digikam-7.2.0-rc-20210103T180117-x86-64-debug.appimage This time the issue does not appear during normal face detection / face recognition. But only appears when running it from the maintenance menu with "clear unconfirmed results and scan again" active. It is reproducible in this way on my system everytime I run it that way. Digikam::PreviewLoadingTask::~PreviewLoadingTask: PreviewLoadingTask destructor was called 0x110beaf0 Digikam::FacePreviewLoader(0xcc6b230) unknown: ASSERT: "!m_LoadingPreview" in file /b/dktemp/digikam-master/core/libs/threadimageio/preview/previewtask.cpp, line 61 --Type <RET> for more, q to quit, c to continue without paging-- Thread 52 "Thread (pooled)" received signal SIGABRT, Aborted. [Switching to Thread 0x7fff9e7fc640 (LWP 154035)] __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:49 49 ../sysdeps/unix/sysv/linux/raise.c: No such file or directory. (gdb) bt #0 __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:49 #1 0x00007ffff3933864 in __GI_abort () at abort.c:79 #2 0x00007ffff414fa3c in QMessageLogger::fatal(char const*, ...) const () from /tmp/.mount_digikaY43SEi/usr/lib/libQt5Core.so.5 #3 0x00007ffff414f13b in qt_assert(char const*, char const*, int) () from /tmp/.mount_digikaY43SEi/usr/lib/libQt5Core.so.5 #4 0x00007ffff747512a in Digikam::PreviewLoadingTask::~PreviewLoadingTask (this=0x110beaf0, __in_chrg=<optimized out>) at /b/dktemp/digikam-master/core/libs/threadimageio/preview/previewtask.cpp:61 #5 0x00007ffff7475199 in Digikam::PreviewLoadingTask::~PreviewLoadingTask (this=0x110beaf0, __in_chrg=<optimized out>) at /b/dktemp/digikam-master/core/libs/threadimageio/preview/previewtask.cpp:62 #6 0x00007ffff748c156 in Digikam::LoadSaveThread::run (this=0xcc6b230) at /b/dktemp/digikam-master/core/libs/threadimageio/fileio/loadsavethread.cpp:130 #7 0x00007ffff74b0b5e in Digikam::DynamicThread::Private::run (this=0x7fffcc00b6e0) at /b/dktemp/digikam-master/core/libs/threads/dynamicthread.cpp:191 #8 0x00007ffff4173d87 in ?? () from /tmp/.mount_digikaY43SEi/usr/lib/libQt5Core.so.5 #9 0x00007ffff4170335 in ?? () from /tmp/.mount_digikaY43SEi/usr/lib/libQt5Core.so.5 #10 0x00007ffff3b00590 in start_thread (arg=0x7fff9e7fc640) at pthread_create.c:463 #11 0x00007ffff3a26223 in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95 Git commit df6e4fcf0157246f6d1d646f6d768d2b27e48f89 by Maik Qualmann. Committed on 05/01/2021 at 11:42. Pushed by mqualmann into branch 'master'. this could be a cause of the crash in the preview loader Related: bug 429307, bug 427380, bug 427333 M +4 -0 core/libs/threadimageio/fileio/loadsavetask.cpp M +5 -0 core/libs/threadimageio/preview/previewtask.cpp M +2 -0 core/libs/threadimageio/thumb/thumbnailtask.cpp https://invent.kde.org/graphics/digikam/commit/df6e4fcf0157246f6d1d646f6d768d2b27e48f89 Is this patch already in digiKam-7.2.0-rc-20210112T182159-x86-64-debug.appimage ? Because it just crashed again (Maintenance -> Face Recognition with "clear unconfirmed results and scan again" on): Digikam::DatabaseWriter::process: Removing old entries () Digikam::PreviewLoadingTask::~PreviewLoadingTask: PreviewLoadingTask destructor was called 0x2944aae0 Digikam::FacePreviewLoader(0x211b2ed0) unknown: ASSERT: "!m_LoadingPreview" in file /b/dktemp/digikam-master/core/libs/threadimageio/preview/previewtask.cpp, line 61 Digikam::FacePipeline::Private::checkFinished: Check for finish: 28 packages, 0 infos to filter, hasFinished() false Digikam::FacePipeline::Private::checkFinished: Check for finish: 27 packages, 0 infos to filter, hasFinished() false Digikam::FacePipeline::Private::checkFinished: Check for finish: 26 packages, 0 infos to filter, hasFinished() false Digikam::DMetadata::getIccProfile: Exif color-space tag is sRGB. Using default sRGB ICC profile. unknown: End Of Image ( 1 ) --Type <RET> for more, q to quit, c to continue without paging-- Thread 305 "Thread (pooled)" received signal SIGABRT, Aborted. [Switching to Thread 0x7ffeff7fe640 (LWP 163300)] __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:49 49 ../sysdeps/unix/sysv/linux/raise.c: No such file or directory. (gdb) bt #0 __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:49 #1 0x00007ffff3914864 in __GI_abort () at abort.c:79 #2 0x00007ffff4130a3c in QMessageLogger::fatal(char const*, ...) const () from /tmp/.mount_digiKaClLuwJ/usr/lib/libQt5Core.so.5 #3 0x00007ffff413013b in qt_assert(char const*, char const*, int) () from /tmp/.mount_digiKaClLuwJ/usr/lib/libQt5Core.so.5 #4 0x00007ffff746bc2a in Digikam::PreviewLoadingTask::~PreviewLoadingTask (this=0x2944aae0, __in_chrg=<optimized out>) at /b/dktemp/digikam-master/core/libs/threadimageio/preview/previewtask.cpp:61 #5 0x00007ffff746bc99 in Digikam::PreviewLoadingTask::~PreviewLoadingTask (this=0x2944aae0, __in_chrg=<optimized out>) at /b/dktemp/digikam-master/core/libs/threadimageio/preview/previewtask.cpp:62 #6 0x00007ffff7482c66 in Digikam::LoadSaveThread::run (this=0x211b2ed0) at /b/dktemp/digikam-master/core/libs/threadimageio/fileio/loadsavethread.cpp:130 #7 0x00007ffff74a767e in Digikam::DynamicThread::Private::run (this=0x7fffcc00a960) at /b/dktemp/digikam-master/core/libs/threads/dynamicthread.cpp:191 #8 0x00007ffff4154d87 in ?? () from /tmp/.mount_digiKaClLuwJ/usr/lib/libQt5Core.so.5 #9 0x00007ffff4151335 in ?? () from /tmp/.mount_digiKaClLuwJ/usr/lib/libQt5Core.so.5 #10 0x00007ffff3ae1590 in start_thread (arg=0x7ffeff7fe640) at pthread_create.c:463 #11 0x00007ffff3a07223 in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95 (gdb) Git commit d62937a9e0c3a496f8ccbdbe72835d08899d6718 by Maik Qualmann. Committed on 14/01/2021 at 06:33. Pushed by mqualmann into branch 'master'. the DynamicThread should not be a QRunnable Otherwise it is a bit strange that a QRunnable started a QRunnable, only the QThreadPool should do this. Related: bug 429307, bug 427380, bug 427333 M +2 -3 core/libs/threads/dynamicthread.cpp M +7 -3 core/libs/threads/dynamicthread.h https://invent.kde.org/graphics/digikam/commit/d62937a9e0c3a496f8ccbdbe72835d08899d6718 Git commit 46d4abd0cfcbdbdbab1f1189186e1737caec2839 by Maik Qualmann. Committed on 14/01/2021 at 06:49. Pushed by mqualmann into branch 'master'. call wait() after stop() before call start() again Otherwise a running task has not yet ended and and is deleted when the QThreadPool is started again. Related: bug 429307, bug 427380, bug 427333 M +1 -0 core/utilities/facemanagement/threads/facepreviewloader.cpp https://invent.kde.org/graphics/digikam/commit/46d4abd0cfcbdbdbab1f1189186e1737caec2839 A new version of the AppImage that contains the latest patches is available. Can you please test whether the problem still exists? Maik (In reply to Maik Qualmann from comment #25) > A new version of the AppImage that contains the latest patches is available. > Can you please test whether the problem still exists? > > Maik Looks good, normally it crashes within 20 minutes. But now it has been running for almost 6 hours. (300K+ images) I'll keep you posted. (In reply to Cliff Albert from comment #26) > (In reply to Maik Qualmann from comment #25) > > A new version of the AppImage that contains the latest patches is available. > > Can you please test whether the problem still exists? > > > > Maik > > Looks good, normally it crashes within 20 minutes. But now it has been > running for almost 6 hours. (300K+ images) I'll keep you posted. It completed the run with no incident. I'll keep testing it, but at the moment everything looks fine and I had no more crashes. Thank you for testing and your feedback. I was pretty sure now that the problem was fixed. ((:-)) We subjected the image loader to a review and code optimization, which was definitely good. In the end, it was just a missing command that led to these weird crashes. This ends a long chapter of crashes the image loader. Maik |