On Tue, Dec 29, 2015 at 1:17 PM, Genes Lists <lists@xxxxxxxxxxxx> wrote: > Fully updated from testing repo. > > Since update to plasma 5.5 I am getting periodic crashes of plasmashell > -trace back below. Seems to be connected with dri/swrast_dri.so. > > i915 graphics so not sure why swrast_dri is called - but then again not > sure what that does exactly (sounds like software rendering). > > Anyone got suggestions or seen similar. > > plasma restarts itself and things go normally for a while until the next > crash. > > gene > > > --- trace back > > Application: Plasma (plasmashell), signal: Segmentation fault > Using host libthread_db library "/usr/lib/libthread_db.so.1". > [Current thread is 1 (Thread 0x7f7c26ff0800 (LWP 2502))] > > ... > > Thread 1 (Thread 0x7f7c26ff0800 (LWP 2502)): > [KCrash Handler] > #5 0x00007f7c34c52b22 in __memmove_avx_unaligned () from > /usr/lib/libc.so.6 > #6 0x00007f7c1763dc8d in ?? () from > /usr/lib/xorg/modules/dri/swrast_dri.so > #7 0x00007f7c1763c645 in ?? () from > /usr/lib/xorg/modules/dri/swrast_dri.so > #8 0x00007f7c1db1e8fa in ?? () from > /usr/lib/qt/qml/org/kde/plasma/core/libcorebindingsplugin.so > #9 0x00007f7c1db1f3f9 in ?? () from > /usr/lib/qt/qml/org/kde/plasma/core/libcorebindingsplugin.so > #10 0x00007f7c1db1f954 in ?? () from > /usr/lib/qt/qml/org/kde/plasma/core/libcorebindingsplugin.so > #11 0x00007f7c38652ac0 in > QQuickWindowPrivate::updateDirtyNode(QQuickItem*) () from > /usr/lib/libQt5Quick.so.5 > #12 0x00007f7c3865331b in QQuickWindowPrivate::updateDirtyNodes() () from > /usr/lib/libQt5Quick.so.5 > #13 0x00007f7c38653490 in QQuickWindowPrivate::syncSceneGraph() () from > /usr/lib/libQt5Quick.so.5 > #14 0x00007f7c3862219c in ?? () from /usr/lib/libQt5Quick.so.5 > > ... > For anyone having this kind of problem some additional useful info might be obtained by running the suggested commands in the page at http://dri.freedesktop.org/wiki/glxinfo/ - and getting at some 3d rendering info specifically when the system has a problem indicated by the fact that the rendering is falling back to the swrast_dri driver. Any chance this report relates to failed graphics drivers which may originate in bugs in the mesa/plasma/X etc code - but at least some initial info showing where there is a problem may point to where to look for more problems so narrow down where the initial failure occurs? -- mike c
