SUMMARY *** i installed a fresh kde neon install on a fresh ssd. i started having trouble from the first run. first, the system monitor or some other applications started pinning the processor core to 100% causing full pause to the machine. then over the past week i am seeing,that when ever i search for "kon" i type "s" and plasma shell crashes. 100% of the time. i could not install debug symbols but with going to 5.26 the day before, i was able to do that now and i have 1-2 backtraces. attaching them here. i have general stability issues on the machine which is a desktop without any graphics card. oh, also, i checked x11 is running on hardware acceleration while wayland TELLS me i am running on software and that crashes within 1-2 minutes. *** STEPS TO REPRODUCE 1. press super key 2. type kons 3. crash OBSERVED RESULT plasma crash EXPECTED RESULT ADDITIONAL INFORMATION
Created attachment 152799 [details] backtrace 1
Created attachment 152800 [details] backtrace 2
Operating System: KDE neon 5.26 KDE Plasma Version: 5.26.0 KDE Frameworks Version: 5.98.0 Qt Version: 5.15.6 Kernel Version: 5.15.0-50-generic (64-bit) Graphics Platform: X11 Processors: 2 × Intel® Core™2 CPU 4300 @ 1.80GHz Memory: 3.1 GiB of RAM Graphics Processor: Mesa DRI Intel® 945G
i tried to use the dr konqi and auto upload the bug but apparently dr konqi itself crashes if i do that, 100% of the time
the machine has general unstability like, i am on desktop and if i select with the mouse and draw a selection square, the screen flickers once-twice and stops. same for doing some refresh, the entire screen goes dark fr half a second sometimes while opening kickoff or even doing f5
Pasting backtrace inline for searchability: Thread 8 (Thread 0x7fe8c3fff700 (LWP 5732)): [KCrash Handler] #4 __memcpy_ssse3 () at ../sysdeps/x86_64/multiarch/memcpy-ssse3.S:1829 #5 0x00007fe8d714bf25 in memcpy (__len=<optimized out>, __src=<optimized out>, __dest=<optimized out>) at /usr/include/x86_64-linux-gnu/bits/string_fortified.h:34 #6 elt (copy=copy@entry=0x7fe8c3ffc830, elt_idx=elt_idx@entry=37880) at ../src/mesa/tnl/t_split_copy.c:262 #7 0x00007fe8d714c4d0 in replay_elts (copy=<optimized out>) at ../src/mesa/tnl/t_split_copy.c:395 #8 _tnl_split_copy (ctx=0x559984f476a0, arrays=0x559984f99a50, prim=0x7fe8c3ffdca0, nr_prims=1, ib=0x7fe8c3ffdcc0, draw=draw@entry=0x7fe8d7149f80 <_tnl_draw_prims>, limits=0x7fe8c3ffdaf4) at ../src/mesa/tnl/t_split_copy.c:634 #9 0x00007fe8d714bda1 in _tnl_split_prims (ctx=ctx@entry=0x559984f476a0, arrays=arrays@entry=0x559984f99a50, prim=prim@entry=0x7fe8c3ffdca0, nr_prims=nr_prims@entry=1, ib=ib@entry=0x7fe8c3ffdcc0, min_index=min_index@entry=0, max_index=<optimized out>, num_instances=1, base_instance=0, draw=0x7fe8d7149f80 <_tnl_draw_prims>, limits=0x7fe8c3ffdaf4) at ../src/mesa/tnl/t_split.c:132 #10 0x00007fe8d714a345 in _tnl_draw_prims (ctx=0x559984f476a0, arrays=0x559984f99a50, prim=0x7fe8c3ffdca0, nr_prims=1, ib=0x7fe8c3ffdcc0, index_bounds_valid=<optimized out>, min_index=<optimized out>, max_index=<optimized out>, num_instances=1, base_instance=0) at ../src/mesa/tnl/t_draw.c:486 #11 0x00007fe8d7244ddf in _mesa_draw_gallium_fallback (ctx=0x559984f476a0, info=0x7fe8c3ffdd30, drawid_offset=0, draws=<optimized out>, num_draws=1) at ../src/mesa/main/draw.c:1016 #12 0x00007fe8d7243dbb in _mesa_validated_drawrangeelements (ctx=ctx@entry=0x559984f476a0, mode=mode@entry=4, index_bounds_valid=index_bounds_valid@entry=false, start=start@entry=0, end=end@entry=4294967295, count=count@entry=5682, type=5123, indices=0x127f0, basevertex=0, numInstances=1, baseInstance=0) at ../src/mesa/main/draw.c:1805 #13 0x00007fe8d7246024 in _mesa_DrawElements (mode=4, count=5682, type=5123, indices=0x127f0) at ../src/mesa/main/draw.c:1938 #14 0x00007fe8e45f4755 in QOpenGLFunctions::glDrawElements (this=0x7fe8b8005460, indices=<optimized out>, type=5123, count=<optimized out>, mode=<optimized out>) at /usr/include/x86_64-linux-gnu/qt5/QtGui/qopenglfunctions.h:760 #15 QSGBatchRenderer::Renderer::renderMergedBatch (batch=0x7fe8b85a9350, this=0x7fe8b80052f0) at scenegraph/coreapi/qsgbatchrenderer.cpp:3111 #16 QSGBatchRenderer::Renderer::renderMergedBatch (this=0x7fe8b80052f0, batch=0x7fe8b85a9350) at scenegraph/coreapi/qsgbatchrenderer.cpp:3006 #17 0x00007fe8e45fa365 in QSGBatchRenderer::Renderer::renderBatches (this=this@entry=0x7fe8b80052f0) at scenegraph/coreapi/qsgbatchrenderer.cpp:4046 #18 0x00007fe8e45fa8ba in QSGBatchRenderer::Renderer::render (this=<optimized out>) at scenegraph/coreapi/qsgbatchrenderer.cpp:4343 #19 0x00007fe8e45de784 in QSGRenderer::renderScene (bindable=..., this=0x7fe8b80052f0) at scenegraph/coreapi/qsgrenderer.cpp:264 #20 QSGRenderer::renderScene (this=0x7fe8b80052f0, bindable=...) at scenegraph/coreapi/qsgrenderer.cpp:220 #21 0x00007fe8e45dec53 in QSGRenderer::renderScene (this=<optimized out>, fboId=<optimized out>) at scenegraph/coreapi/qsgrenderer.cpp:212 #22 0x00007fe8e464d797 in QSGDefaultRenderContext::renderNextFrame (this=0x55998547b7b0, renderer=0x7fe8b80052f0, fboId=<optimized out>) at scenegraph/qsgdefaultrendercontext.cpp:228 #23 0x00007fe8e46bbd69 in QQuickWindowPrivate::renderSceneGraph (this=this@entry=0x55998363a9a0, size=..., surfaceSize=...) at items/qquickwindow.cpp:617 #24 0x00007fe8e465b339 in QSGRenderThread::syncAndRender (this=0x5599852e0290, grabImage=0x0) at /usr/include/x86_64-linux-gnu/qt5/QtGui/5.15.6/QtGui/private/qrhi_p.h:1259 #25 0x00007fe8e465f7c7 in QSGRenderThread::run (this=0x5599852e0290) at scenegraph/qsgthreadedrenderloop.cpp:1042 #26 0x00007fe8e29bc543 in QThreadPrivate::start (arg=0x5599852e0290) at thread/qthread_unix.cpp:330 #27 0x00007fe8e21ed609 in start_thread (arg=<optimized out>) at pthread_create.c:477 #28 0x00007fe8e2639133 in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95
This is a regression in your graphics drivers. Unfortunately there were a number of Mesa regressions recently. I would recommend reporting it to the Mesa folks; see https://docs.mesa3d.org/bugs.html.
(In reply to Nate Graham from comment #7) > This is a regression in your graphics drivers. Unfortunately there were a > number of Mesa regressions recently. I would recommend reporting it to the > Mesa folks; see https://docs.mesa3d.org/bugs.html. ok. i have done that.... can you tell me when will "wlfreerdp" package in kde neon wayland session be installed ? i am unable to use krdc because it does not exist currently
That's a question for the Neon devs that doesn't seem related to this bug report.
Hi folks! It should be added here that the effective solution in this topic is to switch to OpenGL ES compositing in KWin. The OP is using a really really old Intel 3th gen 945G iGPU which even don't supports OpenGL 2.0 fully. However, the awesome Mesa drivers offers stable OpenGL ES 2.0 support even for that old vintage hardware. ;-) To get there, open the following file: sudo nano /etc/profile.d/kwin.sh And then add: export KWIN_COMPOSE=O2ES Save and reboot. By the way, it was confirmed by the OP "johnathan" that this works. More information can be found at the corresponding Mesa bug report: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7499 @Nate Graham I really don't understand why the OpenGL ES renderer is not present as an official option in the UI. It is absolutely useful especially on older and weaker hardware like we have here. The only drawback is that some more complex themes / global designs may show artifacts. But that's acceptable as long as the standard Breeze theme is fully working. I can confirm this for latest Kubuntu 22.04 running at a dual-GPU based iMac 12,2 computer. Absolutely no issues there even with the enforced OpenGL ES compositing backend. :D
Removing it wasn't my decision; you'll have to talk to the KWin folks about that. IIRC they said it caused maintenance headaches.