Red Hat Bugzilla – Bug 657816
[Arrandale] X11 crash for another user on user logoff
Last modified: 2011-05-12 08:04:26 EDT
Created attachment 463299 [details]
X11 log crash
Description of problem:
X11 crashes for another user on user logoff.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Have two users logged in.
2. Log one off.
Should not crash.
[ 52551.303] 0: /usr/bin/Xorg (xorg_backtrace+0x2d) [0x483065]
[ 52551.303] 1: /usr/bin/Xorg (0x400000+0x52cb2) [0x452cb2]
[ 52551.303] 2: /lib64/libc.so.6 (0x346d400000+0x34100) [0x346d434100]
[ 52551.303] 3: /usr/bin/Xorg (0x400000+0xfb1cf) [0x4fb1cf]
[ 52551.303] 4: /usr/bin/Xorg (0x400000+0xfba80) [0x4fba80]
[ 52551.303] 5: /usr/bin/Xorg (0x400000+0xfac47) [0x4fac47]
[ 52551.303] 6: /usr/bin/Xorg (0x400000+0xfadae) [0x4fadae]
[ 52551.303] 7: /usr/bin/Xorg (xf86Wakeup+0x552) [0x45ab38]
[ 52551.303] 8: /usr/bin/Xorg (WakeupHandler+0x42) [0x42e202]
[ 52551.303] 9: /usr/bin/Xorg (WaitForSomething+0x2d1) [0x44d7b9]
[ 52551.303] 10: /usr/bin/Xorg (0x400000+0x2adbf) [0x42adbf]
[ 52551.303] 11: /usr/bin/Xorg (0x400000+0x215da) [0x4215da]
[ 52551.303] 12: /lib64/libc.so.6 (__libc_start_main+0xfd) [0x346d41ee7d]
[ 52551.303] 13: /usr/bin/Xorg (0x400000+0x210a9) [0x4210a9]
[ 52551.304] Segmentation fault at address 0x40eb4000
Fatal server error:
[ 52551.312] Caught signal 11 (Segmentation fault). Server aborting
I am sorry for leaving this bug for so long behind. Could you please confirm us the current status of this bug with the latest updates to your distro (F14 or whatever you have) applied, please?
I haven't see this in a while (I'm on F-14 x86_64).
Actually, I'm suspecting that this happens when one of the components X11 links to or X11 itself gets updated and the machine is not rebooted, but instead one continues to use it and performs a user switch in due course.
(In reply to comment #3)
> I haven't see this in a while (I'm on F-14 x86_64).
> Actually, I'm suspecting that this happens when one of the components X11 links
> to or X11 itself gets updated and the machine is not rebooted, but instead one
> continues to use it and performs a user switch in due course.
OK, closing as WORKSFORME. Problems with upgrade without reboot is kind of questionable issue, but we still consider this operators problem not to reboot when you are told so by PackageKit (or when you would be told so, if you use plain yum).