I am following debian testing. In the course of a system update I have recently pulled in a change in the kde or an underlying component that causes the plasma panel (only the panel) to lock up (become unrespsonsive to mouse clicks, clock and animated icons (nm) do not update). I 'updated' to unstable with caused the freeze to become more systematic. Typical packet versions: ibkf5plasma5 at 5.28.0, plasma components (-nm, -desktop, ...) at 5.8.4. Kernel is 4.8.11, Qt 5.7.1~20161021, intel graphics. I observe similar freezes currently also at home on a PC with AMD/Radeon graphics. Freezes do not seem to depend on compositing. I went from GL 3.1 to Xrender without changing the behavior noticably. I cannot identify anything in journalctl, .xsession-errors, dmesg and the like. Symptoms: sddm comes up, login, animated splash screen comes up, then animation stops. Switching virtual console CTL-ALT-F1 and back to CTL-ALT-F7 gets rid of the splash screen and exposes the desktop. At this point, the panel seems to live (reacts to the mouse). Kwallet asks me for login info, after closing, the panel is frozen. This has happened three times in a row which makes me confident that I should be able to reproduce the problem. Before going to 'unstable', freezes occured at random times during work. Some applications (Mathematica) would then freeze with the panel, then get unlocked after the 'killall plasmashell, plasmashell &' spell from a terminal. I could help debugging (a little) if you tell me what to watch out for... TIA, Stefan
This is a section of the .xsession-errors which is associated to one of the 'random' freezes during a session (which would freeze also some applications where some = Mathematica which would then start working again after a restart of plasmashell. The SEARCH is from akonadi/kdepim which I know to have taken place before the first freeze. Maybe it helps. SEARCH: Query: "{\n \"limit\": 1,\n \"negated\": false,\n \"rel\": 1,\n \"subTerms\": [\n {\n \"cond\": 0,\n \"key\": \"email\",\n \"negated\": false,\n \"value\": \"bbbbbb@basdfsdfaf.org\"\n }\n ]\n}\n" MimeTypes: ("text/directory") Collections: QVector(0, 408, 397) Remote: false Recursive true Executing search "akonadi_newmailnotifier_agent-SearchSession" Search done "akonadi_newmailnotifier_agent-SearchSession" (without remote search) Result: 0 matches Executing search "searchUpdate-1481840038" searchUpdateResultsAvailable 278 0 results Got 0 results, out of which 0 are already in the collection Added 0 Search done "searchUpdate-1481840038" (without remote search) Search update finished All results: 0 Removed results: 0 Executing search "searchUpdate-1481840038" searchUpdateResultsAvailable 277 0 results Got 0 results, out of which 0 are already in the collection Added 0 Search done "searchUpdate-1481840038" (without remote search) Search update finished All results: 0 Removed results: 0 Service "org.kde.StatusNotifierHost-5019" unregistered org.kde.knotifications: env says KDE is running but SNI unavailable -- check KDE_FULL_SESSION and XDG_CURRENT_DESKTOP org.kde.knotifications: env says KDE is running but SNI unavailable -- check KDE_FULL_SESSION and XDG_CURRENT_DESKTOP org.kde.knotifications: env says KDE is running but SNI unavailable -- check KDE_FULL_SESSION and XDG_CURRENT_DESKTOP Registering "org.kde.StatusNotifierItem-1855-2/StatusNotifierItem" to system tray QXcbConnection: XCB error: 3 (BadWindow), sequence: 21785, resource id: 48234507, major code: 15 (QueryTree), minor code: 0 QXcbConnection: XCB error: 3 (BadWindow), sequence: 21833, resource id: 48234518, major code: 15 (QueryTree), minor code: 0 QXcbConnection: XCB error: 3 (BadWindow), sequence: 21848, resource id: 48234527, major code: 25 (SendEvent), minor code: 0 QXcbConnection: XCB error: 3 (BadWindow), sequence: 21895, resource id: 48234527, major code: 15 (QueryTree), minor code: 0 Registering "org.kde.StatusNotifierHost-13999" as system tray kmix(1855)/kdeui (KNotification) KStatusNotifierItemPrivate::registerToDaemon: Registering a client interface to the KSta tusNotifierWatcher Service "org.kde.StatusNotifierItem-1855-2" unregistered powerdevil: Screen brightness value: 256 powerdevil: Screen brightness value max: 852 powerdevil: Kbd backlight brightness value: 1 powerdevil: Kbd backlight brightness value max: 2 powerdevil: Screen brightness value: 256 powerdevil: Screen brightness value max: 852 powerdevil: Kbd backlight brightness value: 1 powerdevil: Kbd backlight brightness value max: 2 Service "org.kde.StatusNotifierHost-13999" unregistered org.kde.knotifications: env says KDE is running but SNI unavailable -- check KDE_FULL_SESSION and XDG_CURRENT_DESKTOP org.kde.knotifications: env says KDE is running but SNI unavailable -- check KDE_FULL_SESSION and XDG_CURRENT_DESKTOP Registering "org.kde.StatusNotifierItem-1855-3/StatusNotifierItem" to system tray QXcbConnection: XCB error: 3 (BadWindow), sequence: 3865, resource id: 48234507, major code: 15 (QueryTree), minor code: 0 QXcbConnection: XCB error: 3 (BadWindow), sequence: 3914, resource id: 48234518, major code: 15 (QueryTree), minor code: 0 QXcbConnection: XCB error: 3 (BadWindow), sequence: 3929, resource id: 48234527, major code: 25 (SendEvent), minor code: 0 QXcbConnection: XCB error: 3 (BadWindow), sequence: 3979, resource id: 48234527, major code: 15 (QueryTree), minor code: 0 org.kde.knotifications: env says KDE is running but SNI unavailable -- check KDE_FULL_SESSION and XDG_CURRENT_DESKTOP Registering "org.kde.StatusNotifierHost-14044" as system tray kmix(1855)/kdeui (KNotification) KStatusNotifierItemPrivate::registerToDaemon: Registering a client interface to the KSta tusNotifierWatcher Service "org.kde.StatusNotifierItem-1855-3" unregistered powerdevil: Screen brightness value: 256 powerdevil: Screen brightness value max: 852 powerdevil: Kbd backlight brightness value: 1 powerdevil: Kbd backlight brightness value max: 2 powerdevil: Screen brightness value: 256 powerdevil: Screen brightness value max: 852 powerdevil: Kbd backlight brightness value: 1 powerdevil: Kbd backlight brightness value max: 2 () (" ---- foobar.nb -------- ") ()
here another trace in .xsession-errors of a 'random' freeze 3 file:///usr/lib/x86_64-linux-gnu/qt5/qml/QtQuick/Controls/Button.qml:99: TypeError: Cannot read property of null trying to show an empty dialog QFileInfo::absolutePath: Constructed with empty filename QFileInfo::absolutePath: Constructed with empty filename QFileInfo::absolutePath: Constructed with empty filename QFileInfo::absolutePath: Constructed with empty filename QFileInfo::absolutePath: Constructed with empty filename QFileInfo::absolutePath: Constructed with empty filename QFileInfo::absolutePath: Constructed with empty filename "/org/freedesktop/UDisks2/block_devices/dm_2d1" lost interfaces: ("org.freedesktop.UDisks2.Filesystem", "org.freedesktop.UDisks2.Block") QFileInfo::absolutePath: Constructed with empty filename
I attach a screenshot when moving the mouse across the taskmanager component in the panel triggered a freeze. In the Konsole output are the last messages that the running plasmashell emitted before freezing. The xcb errors correspond to the mouse moving acorss the task manager 'buttons', the nm stuff accumulated over the lifetime of this 'plasmashell - the startup messages disappeared further up.
Created attachment 102809 [details] screenshot of plasmashell output after freeze see previous comment for explanation
Created attachment 102810 [details] stack trace of freeze OK, here is again a freeze that I triggered by moving the mouse over the taskmanager component. This time I started plasmashell within gdb, CTL_C and 'where' in gdb after freeze. plasmashell/libxcb apparently waits for Godot.
Created attachment 102812 [details] another stack trace after freeze, now for all threads another freeze - trace was taken about two hours after freeze (deduced from frozen clock reading), trigger reason unclear
Created attachment 102882 [details] stacktrace after freeze, more complete symbols another stacktrace which has more complete symbol resolution
None of those traces show anything abnormal, thread 0 is just waiting for the graphic card to reply to a request to do something, which is something we should be doing. Can you see if this is: https://bugs.kde.org/show_bug.cgi?id=373131 and report back
On Tuesday, 20 December 2016 00:39:34 CET you wrote: > https://bugs.kde.org/show_bug.cgi?id=373697 > > David Edmundson <kde@davidedmundson.co.uk> changed: > > What |Removed |Added > ---------------------------------------------------------------------------- > CC| |kde@davidedmundson.co.uk > > --- Comment #8 from David Edmundson <kde@davidedmundson.co.uk> --- > None of those traces show anything abnormal, thread 0 is just waiting for > the graphic card to reply to a request to do something, which is something > we should be doing. > > Can you see if this is: > https://bugs.kde.org/show_bug.cgi?id=373131 > > and report back Hello David, yes, what is described in 373131 sounds very much like the problem I see. Sorry that I overlooked that bug when I posted and thanks for your hint. Happy Christmas, Stefan
thanks *** This bug has been marked as a duplicate of bug 373131 ***