Dear Maintainer, I have not used konqueror for probably more than a year. I ran it from the Xfce menu and it would not start, with a missing file. I ran 'apt reinstall konqueror', and six additional new packages were installed. Konqueror now starts but immediately segfaults. From konqueror error box: Executable: konqueror PID: 5083 Signal: Segmentation fault (11) Time: 18/02/2021 11:19:41 GMT From terminal: ~$ konqueror KCrash: crashing... crashRecursionCounter = 2 KCrash: Application Name = konqueror path = /usr/bin pid = 5167 KCrash: Arguments: /usr/bin/konqueror KCrash: Attempting to start /usr/lib/x86_64-linux-gnu/libexec/drkonqi [1]+ Stopped konqueror ~$ QSocketNotifier: Invalid socket 11 and type 'Read', disabling... [1]+ Exit 253 konqueror The file /usr/lib/x86_64-linux-gnu/libexec/drkonqi exists, root:root 755 The same error occurs if run with sudo.
Hello Joe,
this information might not be sufficient for the maintainer.
Could you try following to collect some more informations?
- install packages:
gdb psmisc konqueror-dbgsym
- stop all running instances of konqueror:
killall konqueror
- then start konqueror inside a debugger:
gdb -q -ex 'set pagination off' -ex 'run' -ex 'bt' -ex 'detach' -ex 'quit' --args konqueror
When the crash happens, a backtrace should be printed.
Please forward this output to this bug.
More debugging hints can be found here [1].
Kind regards,
Bernhard
[1] https://wiki.debian.org/HowToGetABacktrace
Reading symbols from konqueror... Reading symbols from /usr/lib/debug/.build-id/e4/6ba37330371215b40e01c7c01397ce28f6aa3c.debug... Starting program: /usr/bin/konqueror [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". [New Thread 0x7ffff0faf700 (LWP 8878)] [New Thread 0x7fffe42c3700 (LWP 8885)] [New Thread 0x7fffe3981700 (LWP 8886)] [New Thread 0x7fffe3180700 (LWP 8887)] [New Thread 0x7fffe297f700 (LWP 8888)] [New Thread 0x7fffe217e700 (LWP 8889)] [New Thread 0x7fffe197d700 (LWP 8890)] Thread 1 "konqueror" received signal SIGSEGV, Segmentation fault. 0x0000000000000000 in ?? () #0 0x0000000000000000 in () #1 0x00007fffeafcf23f in () at /usr/lib/x86_64-linux-gnu/dri/r300_dri.so #2 0x00007fffeaaeb704 in () at /usr/lib/x86_64-linux-gnu/dri/r300_dri.so #3 0x00007fffeaaeca94 in () at /usr/lib/x86_64-linux-gnu/dri/r300_dri.so #4 0x00007fffeaace95e in () at /usr/lib/x86_64-linux-gnu/dri/r300_dri.so #5 0x00007fffeafcd903 in () at /usr/lib/x86_64-linux-gnu/dri/r300_dri.so #6 0x00007ffff07175af in () at /usr/lib/x86_64-linux-gnu/libGLX_mesa.so.0 #7 0x00007ffff0705e49 in () at /usr/lib/x86_64-linux-gnu/libGLX_mesa.so.0 #8 0x00007ffff15db1b5 in () at /usr/lib/x86_64-linux-gnu/qt5/plugins/xcbglintegrations/libqxcb-glx-integration.so #9 0x00007ffff63fa4fb in QOpenGLContext::destroy() () at /usr/lib/x86_64-linux-gnu/libQt5Gui.so.5 #10 0x00007ffff63fa7a7 in QOpenGLContext::~QOpenGLContext() () at /usr/lib/x86_64-linux-gnu/libQt5Gui.so.5 #11 0x00007ffff15dd2bf in () at /usr/lib/x86_64-linux-gnu/qt5/plugins/xcbglintegrations/libqxcb-glx-integration.so #12 0x00007ffff15dd909 in () at /usr/lib/x86_64-linux-gnu/qt5/plugins/xcbglintegrations/libqxcb-glx-integration.so #13 0x00007fffbd4d9845 in () at /usr/lib/x86_64-linux-gnu/libQt5WebEngineCore.so.5 #14 0x00007fffffffd210 in () #15 0x00007fffbd4ddec1 in () at /usr/lib/x86_64-linux-gnu/libQt5WebEngineCore.so.5 #16 0x00007fffffffd610 in () #17 0x00007fffbd4dc4e5 in () at /usr/lib/x86_64-linux-gnu/libQt5WebEngineCore.so.5 #18 0x00007fffffffd230 in () #19 0x0000000000000005 in () #20 0x00007fffffffd530 in () #21 0x00007fffffffd540 in () #22 0x00007fffffffd590 in () #23 0x00007fffffffd380 in () #24 0x00007fffffffd360 in () #25 0x00007fffffffd2a8 in () #26 0x00007fffffffd5a0 in () #27 0x0000000000000000 in () Detaching from program: /usr/bin/konqueror, process 8873 [Inferior 1 (process 8873) detached]
Le samedi 6 mars 2021, 17:43:11 CET J Rowan a écrit : Dear Joe, thanks for taking the time to report the bug and investigate. What the stack tells us is that the crash is in the R300 OpenGL graphics driver, as called by Qt Web Engine that Konqueror uses for the HTML rendering. I cannot reproduce the crash myself on various more recent pieces of hardware, and neither did the team get similar bug reports. So I can see 3 possible causes and (im)possible fixes from the most probable to the least : 1/ The OpenGL driver has a broken code path that is triggered here and should be fixed. 2/ Qt Web Engine calls the OpenGL stack incorrectly in this case and could be fixed. Alternatively it could be made to work around that particular OpenGL driver bug even if we’re in case 1/. 3/ That piece of graphics hardware is (now) broken and the code path makes it visible, nothing we can do about it. If that was the case I would expect the crash not to be reproducible 100% of the time, and here it looks like you can reproduce it so maybe not. Looking the hardware support for that driver you must have a graphics card from the Radeon 9500-9800 series, X300-X600 series or X1050/X1150 from the 2003-2005 era. This driver is in maintenance mode to say the least so I wouldn’t get my hopes too high about a possible fix. However if you want to help diagnose the issue further you could install the dbgsym packages [1] for at least libgl1-mesa-dri, libglx-mesa0, libqt5webenginecore5 and konqueror packages, and rerun the same gdb call so we get the crash stack with the function names and line numbers. If you can confirm whether you get crashes with other programs using 3D rendering besides konqueror, that would be a useful data point. Then we can decide if we want to forward the bug to mesa or Qt. Again be ready to face refusals from both upstreams to work on an issue for such an old piece of hardware. Most opensource projects are understaffed and I have the feeling that reproducing let alone fixing this issue is not going to be an easy task. Also I seem to remember that the R300 driver was made out of reverse engineering or with documentation provided to selected developers under NDA, which is not going to help either… If you don’t have the time or motivation to do the above and dig further I’d still appreciate that you say so, and I’ll just mark this bug as « won’t fix ». [1] https://wiki.debian.org/AutomaticDebugPackages?highlight=%28dbgsym%29 Thanks & happy hacking ! -- Aurélien