You are not logged in.
@eriol
[ 410.820] (EE) Backtrace:
[ 410.820] (EE) 0: /usr/lib/Xorg (xorg_backtrace+0x53) [0x5654d111fbd3]
[ 410.820] (EE) 1: /usr/lib/Xorg (0x5654d0fd9000+0x151a15) [0x5654d112aa15]
[ 410.820] (EE) 2: /usr/lib/libc.so.6 (0x7f5547266000+0x3d6a0) [0x7f55472a36a0]
[ 410.820] (EE) 3: /usr/lib/Xorg (0x5654d0fd9000+0x7ef03) [0x5654d1057f03]
[ 410.820] (EE) 4: /usr/lib/Xorg (ChangeWindowAttributes+0xd7f) [0x5654d107afaf]
[ 410.820] (EE) 5: /usr/lib/Xorg (0x5654d0fd9000+0x6fa58) [0x5654d1048a58]
[ 410.820] (EE) 6: /usr/lib/Xorg (0x5654d0fd9000+0x3a165) [0x5654d1013165]
[ 410.820] (EE) 7: /usr/lib/libc.so.6 (__libc_start_main+0xf2) [0x7f554728e152]
[ 410.820] (EE) 8: /usr/lib/Xorg (_start+0x2e) [0x5654d10135ae]
[ 410.820] (EE)
[ 410.820] (EE) Segmentation fault at address 0x18
[ 410.820] (EE)
That looks more like https://bbs.archlinux.org/viewtopic.php?id=258590 but the logs don't cover the BadValue details.
Same goes for https://gitlab.freedesktop.org/xorg/xse … ssues/1073 from #25
W/o details on the X11 error predating this, we don't know whether the two are related.
The BadValue error in https://bbs.archlinux.org/viewtopic.php?id=258590 crashes xfwm4 and then it seems in turn the Xorg server crashes - the patterns are distinctly different (window attributes and input handling) and consistent, so I don't think it's the same underlying bug in Xorg.
@avallac_h, that "only" covers the "mieqProcessInputEvents" segfault originally discussed in this thread before mugamma brought up the secondary ChangeWindowAttributes bug.
Offline
I've just updated my office machine to to the latest Xorg 1.20.9-1 - and the same problem occurs with light-locker and Xorg crashing - I've tried xfce4-screensaver as an alternative locker and it works fine.
Offline
@seth Perhaps I can explain why the name of the function in the stack trace can vary. I guess it depends on the last action performed by the user before the crash:
If you performed a mouse click right before the crash you'll most likely end with mieqProcessDeviceEvent() error.
If you have sloppy mouse focus enabled, like I do, you just need to move the mouse: cursor leaves the window, window state changes, bang, ChangeWindowAttributes() error.
I even got a ChangeWindowDeviceCursor() error once. Not sure what caused it, highlighting a hyperlink in a web-browser, maybe. Or moving the cursor to the window border (those resize arrows).
авг 30 19:32:20 nuc systemd-coredump[8889]: Process 4716 (Xorg) of user 0 dumped core.
Stack trace of thread 4716:
#0 0x00007f4b5136e615 raise (libc.so.6 + 0x3d615)
#1 0x00007f4b51357862 abort (libc.so.6 + 0x26862)
#2 0x0000561aa903035a OsAbort (Xorg + 0x14a35a)
#3 0x0000561aa9031e21 FatalError (Xorg + 0x14be21)
#4 0x0000561aa9037a79 n/a (Xorg + 0x151a79)
#5 0x00007f4b5136e6a0 __restore_rt (libc.so.6 + 0x3d6a0)
#6 0x0000561aa8f64f03 n/a (Xorg + 0x7ef03)
#7 0x0000561aa8f8c03b ChangeWindowDeviceCursor (Xorg + 0xa603b)
#8 0x0000561aa8ff25cd n/a (Xorg + 0x10c5cd)
#9 0x0000561aa8f20165 n/a (Xorg + 0x3a165)
#10 0x00007f4b51359152 __libc_start_main (libc.so.6 + 0x28152)
#11 0x0000561aa8f205ae _start (Xorg + 0x3a5ae)
Could it be that lightdm (light-locker?) somehow kills Xorg and as a result we see the name of last function used?
If you are doing it, I'm doing it. If I don't do it, will you? If you don't do it, I'll do it. I'm doing it so that you do it too!! (c) Kamina
Offline
I can (as many previous) confirm that downgrading xorg-server does eliminates these crashes. If anyone needs more logs, you find a couple in the other thread https://bbs.archlinux.org/viewtopic.php … 7#p1923497 including those from lightdm if any one is interested (as seth predicted, I can't find any interesting in them, but on the other side, my knowledge in interpreting logs is low at best).
Offline
Xorg devs found a regression in 1.20.9 that applies our case:
https://gitlab.freedesktop.org/xorg/xse … ssues/1067
Patch is already available, maybe someone wants to test it:
https://gitlab.freedesktop.org/xorg/xse … quests/504
Update:
I confirm that the patch works around this issue.
PKGBUILD.diff for those who want to try it out (use with caution!).
Last edited by avallac_h (2020-08-31 16:23:14)
If you are doing it, I'm doing it. If I don't do it, will you? If you don't do it, I'll do it. I'm doing it so that you do it too!! (c) Kamina
Offline
Okay - I'm going to mark this thread as solved. avallac_h has tested a patch, that works, which i assume will be added to the next iteration of xorg-server. Also, a downgrade will fix the issue until it has been officially corrected. Finally, using a different locker such as XFCE4-screensaver will work around the problem. Many thanks to all those who have contributed to this thread.
Offline