[fixed] KDE failure in X mid-session

cfr · 2022-12-04 02:28:25

seth wrote:

Rha 03 01:36:54 FyEnw kwin_x11[1026]: The X11 connection broke: I/O error (code 1)
I have a strange message to the effect that the ethernet connection is activated (and then deactivated)
Is the hostname locally set or received via dhcp?

Locally. At least, I set it in /etc/hostname. I think I set it there because it said to in the installation guide and everything seemed to work (but that was a while ago, so I might be misremembering).

seth wrote:

(Though because of "*much* earlier", that'd probably neither be relevant)

At least an hour. I think a couple. I can check if it matters whether it was 1 hour or 2 or 3, but I'm guessing not.

seth wrote:

Is this a hybrid graphics system?
https://bbs.archlinux.org/viewtopic.php?id=257313

Absolutely not. Not being hybrid graphics was one of my selection criteria . Specs in my signature are accurate (as far as I know), including the graphics.

Last edited by cfr (2022-12-04 02:32:45)

seth · 2022-12-04 08:32:11

Looking at the journal in #10 again:

Tach 12 01:51:23 MySelf plasmashell[72307]: 18 -- exe=/usr/bin/kile
Tach 12 01:51:23 MySelf plasmashell[72307]: 13 -- platform=xcb
Tach 12 01:51:23 MySelf plasmashell[72307]: 11 -- display=:0
Tach 12 01:51:23 MySelf plasmashell[72307]: 13 -- appname=kile
Tach 12 01:51:23 MySelf plasmashell[72307]: 17 -- apppath=/usr/bin
Tach 12 01:51:23 MySelf plasmashell[72307]: 10 -- signal=11
Tach 12 01:51:23 MySelf plasmashell[72307]: 10 -- pid=72307
Tach 12 01:51:23 MySelf plasmashell[72307]: 18 -- appversion=2.9.93
Tach 12 01:51:23 MySelf plasmashell[72307]: 17 -- programname=Kile
Tach 12 01:51:23 MySelf plasmashell[72307]: 31 -- bugaddress=submit@bugs.kde.org
Tach 12 01:51:23 MySelf plasmashell[72307]: 17 -- productname=kile
Tach 12 01:51:23 MySelf plasmashell[72307]: 12 -- startupid=0

kile segfaults - do you have coredumpctl entries for that preserved?

Tach 12 01:51:23 MySelf plasmashell[72307]: KCrash: crashing... crashRecursionCounter = 2
Tach 12 01:51:23 MySelf plasmashell[72307]: KCrash: Application Name = kile path = /usr/bin pid = 72307
Tach 12 01:51:23 MySelf plasmashell[72307]: KCrash: Arguments: /usr/bin/kile
Tach 12 01:51:23 MySelf plasmashell[72307]: KCrash: Attempting to start /usr/lib/drkonqi

plasmas wants to start drkonqi.

Tach 12 01:51:23 MySelf kwin_x11[71625]: qt.qpa.xcb: QXcbConnection: XCB error: 3 (BadWindow), sequence: 38020, resource id: 71303185, major code: 18 (ChangeProperty), minor code: 0
Tach 12 01:51:23 MySelf kwin_x11[71625]: qt.qpa.xcb: QXcbConnection: XCB error: 3 (BadWindow), sequence: 38027, resource id: 4194323, major code: 15 (QueryTree), minor code: 0

Kwin isn't happy about something, could be the gone kile window or the drkonqi window.

Tach 12 01:51:24 MySelf kded5[71624]: Registering ":1.285/StatusNotifierItem" to system tray
Tach 12 01:51:24 MySelf plasmashell[71678]: Could not find the Plasmoid for Plasma::FrameSvgItem(0x55790d2d4aa0) QQmlContext(0x55790bcd0920) QUrl("file:///usr/share/plasma/plasmoids/org.kde.plasma.notifications/contents/ui/global/Globals.qml")
Tach 12 01:51:24 MySelf plasmashell[71678]: Could not find the Plasmoid for Plasma::FrameSvgItem(0x55790d2d4aa0) QQmlContext(0x55790bcd0920) QUrl("file:///usr/share/plasma/plasmoids/org.kde.plasma.notifications/contents/ui/global/Globals.qml")
Tach 12 01:51:24 MySelf plasmashell[71678]: file:///usr/share/plasma/plasmoids/org.kde.plasma.notifications/contents/ui/NotificationItem.qml:221:21: QML SelectableLabel: Binding loop detected for property "implicitWidth"
Tach 12 01:51:24 MySelf plasmashell[71678]: file:///usr/share/plasma/plasmoids/org.kde.plasma.notifications/contents/ui/NotificationItem.qml:221:21: QML SelectableLabel: Binding loop detected for property "implicitWidth"
Tach 12 01:51:24 MySelf plasmashell[71678]: file:///usr/share/plasma/plasmoids/org.kde.plasma.notifications/contents/ui/NotificationItem.qml:221:21: QML SelectableLabel: Binding loop detected for property "implicitHeight"

Plasma adds (?) the SNI to the systray, QML continues to be a joke.

Tach 12 01:51:28 MySelf kwin_x11[71625]: qt.qpa.xcb: QXcbConnection: XCB error: 2 (BadValue), sequence: 40937, resource id: 17385233, major code: 149 (Unknown), minor code: 4
Tach 12 01:51:28 MySelf kwin_x11[71625]: qt.qpa.xcb: QXcbConnection: XCB error: 136 (Unknown), sequence: 40938, resource id: 17385234, major code: 148 (Unknown), minor code: 1
Tach 12 01:51:28 MySelf kwin_x11[71625]: The X11 connection broke: I/O error (code 1)
Tach 12 01:51:28 MySelf kwin_x11[71625]: XIO:  fatal IO error 0 (Success) on X server ":0"
Tach 12 01:51:28 MySelf kwin_x11[71625]:       after 5414946 requests (5414946 known processed) with 0 events remaining.

KWin goes down.

1. Let's see whether we can figure why kile segfaults: https://wiki.archlinux.org/title/Core_d … _core_dump
2. Do you have https://archlinux.org/packages/extra/x86_64/drkonqi/ installed ?
3. Do you usually have the SNI present? What happens for

notify-send 'Fel rhech mewn pot jam'

Sidebar, there's a suspicious amount of sheeps involved here.
https://www.walesonline.co.uk/lifestyle … ch-9557813

cfr · 2022-12-04 17:07:41

seth wrote:

kile segfaults - do you have coredumpctl entries for that preserved?
...
1. Let's see whether we can figure why kile segfaults: https://wiki.archlinux.org/title/Core_d … _core_dump

No. The last entry is for kile, but it is from 2022-11-12 and so already deleted. I don't seem to have got any coredumpctl entries from the last cascade of crashes.

seth wrote:

plasmas wants to start drkonqi.
...
2. Do you have https://archlinux.org/packages/extra/x86_64/drkonqi/ installed ?

Yes. And it does work i.e. I sometimes get KDE crash report windows following something going down.

seth wrote:

Kwin isn't happy about something, could be the gone kile window or the drkonqi window.

kwin is always unhappy about something.

seth wrote:

Plasma adds (?) the SNI to the systray, QML continues to be a joke.

??

seth wrote:

3. Do you usually have the SNI present? What happens for
notify-send 'Fel rhech mewn pot jam'

I don't really know what SNI is. I get a desktop notification if I execute that command in a terminal emulator. I also use notify-send in at least one script and get desktop notifications as expected. Does that mean SNI is usually present? Does that imply something about QML? (I have even less idea what QML is than SNI.)

seth wrote:

Sidebar, there's a suspicious amount of sheeps involved here.
https://www.walesonline.co.uk/lifestyle … ch-9557813

What's suspicious about sheeps?

seth · 2022-12-04 18:12:39

"StatusNotifierItem" the thing that shows messages eg. from notify-send. It's also (likely) dbus activated, ie. gets added when you need it.
https://en.wikipedia.org/wiki/QML is a wonky implementation of an idiotic idea to make writing code mode like "wEbDeVeLoPmEnT"

Apparently kwin isn't upset about the notification itself, maybe about the first message of the session (though i doubt it)
So it's more about kile crashing/drkonqi showing?
If kile crashes and drkonqi shows up and kwin goes south, make sure to preserve the kile backtrace in the drkonqi window.
Maybe we can address the trigger this way.

Edit: the rest of the world holds a bunch of disgusting and shocking tropes about British nations w/ more sheep then men

Last edited by seth (2022-12-04 18:14:12)

cfr · 2022-12-04 19:01:12

seth wrote:

"StatusNotifierItem" the thing that shows messages eg. from notify-send. It's also (likely) dbus activated, ie. gets added when you need it.
https://en.wikipedia.org/wiki/QML is a wonky implementation of an idiotic idea to make writing code mode like "wEbDeVeLoPmEnT"

Thanks.

seth wrote:

So it's more about kile crashing/drkonqi showing?
If kile crashes and drkonqi shows up and kwin goes south, make sure to preserve the kile backtrace in the drkonqi window.
Maybe we can address the trigger this way.

I don't know how likely this is, but I can look for a coredump, too. Evidently kile produced one in November, so maybe it will again. drkonqi doesn't always seem to get activated (or maybe it does, but something happens to it) but I'll save the report if I get one.

seth wrote:

Edit: the rest of the world holds a bunch of disgusting and shocking tropes about British nations w/ more sheep then men

Actually, the English (at least) hold similar views about Welshmen and sheep. Not being a Welshman (or, indeed, a British man), I'm presumably less suspect.

cfr · 2022-12-08 18:33:05

Crashed again.

Finally managed to set the display correctly: http://0x0.st/od25.txt

coredumpctl shows a coredump for chromium, but that was yesterday (and I didn't get a crash yesterday). drkonqi did not show up.

Wed 2022-12-07 02:51:10 GMT  48967 1000 1000 SIGILL  present  /usr/lib/chromium/chromium                                        40.5M

Here's a complete journal: http://0x0.st/od_z.txt. The crash shows up at Rha 08 16:35:22:

Rha 08 16:33:03 MySelf kwin_x11[1004]: kwin_core: XCB error: 3 (BadWindow), sequence: 23642, resource id: 46137818, major code: 129 (SHAPE), minor code: 6 (Input)
...
Rha 08 16:35:22 MySelf kwin_x11[1004]: The X11 connection broke: I/O error (code 1)
Rha 08 16:35:22 MySelf kwin_x11[1004]: XIO:  fatal IO error 62 (Timer expired) on X server ":0"
Rha 08 16:35:22 MySelf kwin_x11[1004]:       after 9211743 requests (9211743 known processed) with 0 events remaining.
Rha 08 16:35:22 MySelf systemd[883]: plasma-kwin_x11.service: Consumed 47min 39.268s CPU time.
...
Rha 08 18:00:57 MySelf systemd-coredump[83901]: Failed to connect to coredump service: Connection refused
Rha 08 18:00:57 MySelf plasmashell[1759]: [1759:1759:1208/180057.682645:ERROR:gl_surface_presentation_helper.cc(260)] GetVSyncParametersIfAvailable() failed for 1 times!
Rha 08 18:00:57 MySelf kwalletd5[1766]: The X11 connection broke (error 1). Did the X11 server die?
Rha 08 18:00:57 MySelf plasmashell[1759]: [1759:1787:1208/180057.683764:ERROR:connection.cc(46)] X connection error received.
Rha 08 18:00:57 MySelf kile[1301]: The X11 connection broke (error 1). Did the X11 server die?
Rha 08 18:00:57 MySelf plasmashell[1759]: X connection to :0 broken (explicit kill or server shutdown).
Rha 08 18:00:57 MySelf systemd[1]: udisks2.service: Main process exited, code=dumped, status=6/ABRT
Rha 08 18:00:57 MySelf systemd[1]: udisks2.service: Failed with result 'core-dump'.
Rha 08 18:00:57 MySelf systemd[1]: Stopped Disk Manager.
Rha 08 18:00:57 MySelf systemd[1]: udisks2.service: Consumed 1.329s CPU time.

Kile looks like victim rather than cause here? But why is the coredump connection refused?

cfr · 2022-12-17 01:46:54

I managed to get an 'inaccessible' coredump, but I guess that's just permissions. Here's the info

           PID: 1228 (udisksd)
           UID: 0 (root)
           GID: 0 (root)
        Signal: 6 (ABRT)
     Timestamp: Sat 2022-12-17 01:19:37 GMT (20min ago)
  Command Line: /usr/lib/udisks2/udisksd
    Executable: /usr/lib/udisks2/udisksd
 Control Group: /system.slice/udisks2.service
          Unit: udisks2.service
         Slice: system.slice
       Boot ID: cb649a20106c41b2aa0a2edfd81743ae
    Machine ID: 0f074fbf02c348ec82b98d1c49e98e74
      Hostname: CyfrifiadurCFR
       Storage: /var/lib/systemd/coredump/core.udisksd.0.cb649a20106c41b2aa0a2edfd81743ae.1228.1671239977000000.zst (present)
  Size on Disk: 505.2K
       Message: Process 1228 (udisksd) of user 0 dumped core.
                
                Stack trace of thread 1228:
                #0  0x00007f3030d0764c n/a (libc.so.6 + 0x8864c)
                #1  0x00007f3030cb7958 raise (libc.so.6 + 0x38958)
                #2  0x00007f3030ca153d abort (libc.so.6 + 0x2253d)
                #3  0x00007f3030e8806a n/a (libglib-2.0.so.0 + 0x1d06a)
                #4  0x00007f3030ee3ae0 g_assertion_message_expr (libglib-2.0.so.0 + 0x78ae0)
                #5  0x00007f3031096877 g_static_resource_fini (libgio-2.0.so.0 + 0x89877)
                #6  0x00007f3031479aae n/a (ld-linux-x86-64.so.2 + 0x4aae)
                #7  0x00007f3030cb9fa5 n/a (libc.so.6 + 0x3afa5)
                #8  0x00007f3030cba120 exit (libc.so.6 + 0x3b120)
                #9  0x00007f3030ca2297 n/a (libc.so.6 + 0x23297)
                #10 0x00007f3030ca234a __libc_start_main (libc.so.6 + 0x2334a)
                #11 0x00005602733b049e _start (udisksd + 0x1949e)
                
                Stack trace of thread 1230:
                #0  0x00007f3030d7a0bf __poll (libc.so.6 + 0xfb0bf)
                #1  0x00007f3030f171ff n/a (libglib-2.0.so.0 + 0xac1ff)
                #2  0x00007f3030ebf132 g_main_context_iteration (libglib-2.0.so.0 + 0x54132)
                #3  0x00007f3030ebf182 n/a (libglib-2.0.so.0 + 0x54182)
                #4  0x00007f3030eede05 n/a (libglib-2.0.so.0 + 0x82e05)
                #5  0x00007f3030d058fd n/a (libc.so.6 + 0x868fd)
                #6  0x00007f3030d87a60 n/a (libc.so.6 + 0x108a60)
                
                Stack trace of thread 1232:
                #0  0x00007f3030d7a0bf __poll (libc.so.6 + 0xfb0bf)
                #1  0x00007f3030f171ff n/a (libglib-2.0.so.0 + 0xac1ff)
                #2  0x00007f3030ebfddf g_main_loop_run (libglib-2.0.so.0 + 0x54ddf)
                #3  0x00007f303111737c n/a (libgio-2.0.so.0 + 0x10a37c)
                #4  0x00007f3030eede05 n/a (libglib-2.0.so.0 + 0x82e05)
                #5  0x00007f3030d058fd n/a (libc.so.6 + 0x868fd)
                #6  0x00007f3030d87a60 n/a (libc.so.6 + 0x108a60)
                ELF object binary architecture: AMD x86-64
More than one entry matches, ignoring rest.

I dumped it to a file with

sudo coredumpctl dump -o coredump-2022-12-17

which I hope is correct.

What do I do with it now? If I need to upload it somewhere, how can I check if it contains confidential data? Because of what I was working on at the time, it might contain confidential data which is not even mine.

Edit: https://gitlab.freedesktop.org/mesa/mes … ests/19972 and the final posts in https://bugs.kde.org/show_bug.cgi?id=461316 look hopeful, if I'm not being overly optimistic?

Last edited by cfr (2022-12-17 05:16:19)

seth · 2022-12-17 08:09:11

Process 1228 (udisksd) of user 0 dumped core.

Not sure whether that's directly related to your issue.

Edit: https://gitlab.freedesktop.org/mesa/mes … ests/19972 and the final posts in https://bugs.kde.org/show_bug.cgi?id=461316 look hopeful, if I'm not being overly optimistic?

https://bbs.archlinux.org/viewtopic.php … 7#p2067567 - but it's been meanwhile indeed reverted, so w/ a little luck, a MESA update will indeed fix it.

Arch Linux

#26 2022-12-04 02:28:25

Re: [fixed] KDE failure in X mid-session

#27 2022-12-04 08:32:11

Re: [fixed] KDE failure in X mid-session

#28 2022-12-04 17:07:41

Re: [fixed] KDE failure in X mid-session

#29 2022-12-04 18:12:39

Re: [fixed] KDE failure in X mid-session

#30 2022-12-04 19:01:12

Re: [fixed] KDE failure in X mid-session

#31 2022-12-08 18:33:05

Re: [fixed] KDE failure in X mid-session

#32 2022-12-17 01:46:54

Re: [fixed] KDE failure in X mid-session

#33 2022-12-17 08:09:11

Re: [fixed] KDE failure in X mid-session

Board footer