Kubuntu 9.04 live CD: "shutdown" doesn't

Bug #366199 reported by Alan Jenkins
2
Affects Status Importance Assigned to Milestone
kdebase-workspace (Ubuntu)
Invalid
Undecided
Unassigned

Bug Description

Shutdown or restart from Kubuntu live CD doesn't work. It logs you out, but doesn't get any further.

I have to switch to a text console and press ctrl+alt+delete. I see the following message:

kdm not responding to TERM signal

Revision history for this message
Alan Jenkins (aj504) wrote :

Aha! This only happens if I enable desktop effects before trying to shut down. I have Intel 965 graphics. So this may be related to the other Intel issues, e.g. Bug 359392.

Revision history for this message
Alan Jenkins (aj504) wrote :

$ lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 9.04
Release: 9.04
Codename: jaunty

$ apt-cache policy kdm
kdm:
  Installed: 4:4.2.2-0ubuntu2
  Candidate: 4:4.2.2-0ubuntu2
  Version table:
 *** 4:4.2.2-0ubuntu2 0
        500 http://archive.ubuntu.com jaunty/main Packages
        100 /var/lib/dpkg/status
$ apt-cache policy xorg
xorg:
  Installed: 1:7.4~5ubuntu18
  Candidate: 1:7.4~5ubuntu18
  Version table:
 *** 1:7.4~5ubuntu18 0
        500 http://archive.ubuntu.com jaunty/main Packages
        100 /var/lib/dpkg/status

$ lspci -nn | grep VGA
00:02.0 VGA compatible controller [0300]: Intel Corporation 82946GZ/GL Integrated Graphics Controller [8086:2972] (rev 02)

Revision history for this message
Alan Jenkins (aj504) wrote :

Xorg.log, prior to enabling desktop effects

Revision history for this message
Alan Jenkins (aj504) wrote :

Damn, I can't reproduce it now. I'll leave it for the time being; I have a load of other issues to sort through and report.

I have some ideas as to what I can try when (if) it happens again. I think X has actually exited at this point, and it's just kdm which is hanging. Because it gets as far as logging me out, and switching to a text console (but doesn't actually restart/shutdown). So I'll try to confirm that with "ps", to see that X is no longer running, and see what process state kdm is in (D-state?). And obviously I'll check Xorg.log and kdm.log for errors.

Changed in kdebase-workspace (Ubuntu):
status: New → Incomplete
Revision history for this message
Alan Jenkins (aj504) wrote :

Gotcha! Ok, this does _not_ depend on me enabling desktop effects. I'm not sure what causes it :-). It could be it's just random.

Changed in kdebase-workspace (Ubuntu):
status: Incomplete → New
Revision history for this message
Alan Jenkins (aj504) wrote :

Xorg.log after trying to reboot. Probably not interesting.

Revision history for this message
Alan Jenkins (aj504) wrote :

kdm.log after trying to reboot. Nothing obvious here, at least to my eye.

Revision history for this message
Alan Jenkins (aj504) wrote :

Output of "ps -ax" after trying to reboot.

There is no X process. There is a "-:0" process, I'm not sure what that means exactly, except that it's related to ":0", my X display / session.

Both kdm and kdm_greet are still running, in state "S" (interuptable sleep). As described before "killall kdm" does nothing; either the TERM signal is blocked or the signal handler isn't able to do it's job.

I checked with "strace -p" just in case it showed anything interesting about kdm, but it just showed a select() call with a number of fds, just as you would expect in a standard event loop.

Revision history for this message
Alan Jenkins (aj504) wrote :

It happened again and I gathered some more data, this will probably make more sense if you have the source code in front of you.

This was after maybe another 5 reboots, so it doesn't seem to happen too often.

select(12, [4 6 7 9 11], NULL, NULL, NULL) = ? ERESTARTNOHAND (To be restarted)
--- SIGTERM (Terminated) @ 0 (0) ---
write(5, "\17"..., 1) = 1
rt_sigreturn(0x5) = -1 EINTR (Interrupted system call)
kill(4123, SIGTERM) = 0
kill(4123, SIGCONT) = 0
select(12, [4 6 7 9 11], NULL, NULL, NULL) = 1 (in [4])
read(4, "\17"..., 1) = 1
kill(4123, SIGTERM) = 0
kill(4123, SIGCONT) = 0
kill(4123, SIGTERM) = 0
kill(4123, SIGCONT) = 0
select(12, [4 6 7 9 11], NULL, NULL, NULL

$ ps -ax | grep 4123
 4123 ? S 0:00 -:0

So "-:0" _is_ significant. Rats, I should have used strace on that as well. I'll just have to wait for next time again.

Revision history for this message
Alan Jenkins (aj504) wrote :

Here's the strace for "-:0", which is actually an instance of kdm (according to /proc/exe)

Process 4129 attached - interrupt to quit
rt_sigsuspend([]) = ? ERESTARTNOHAND (To be restarted)
--- SIGTERM (Terminated) @ 0 (0) ---
rt_sigsuspend([]

I.e. it's waiting in sigsuspend() and hasn't blocked TERM, but it refuses to die.

Revision history for this message
Jonathan Thomas (echidnaman) wrote :

Is this still an issue with 9.10?

Changed in kdebase-workspace (Ubuntu):
status: New → Incomplete
Revision history for this message
Jonathan Thomas (echidnaman) wrote :

Hi there!

Thanks for reporting this bug! Your bug seems to be a problem with the KDE program itself, and not with our KDE packages. While we appreciate your issue, it would be better if it was tracked at https://bugs.kde.org, so that the KDE developers can deal with this speedily and have direct communication with you as the reporter for more effective debugging.

Thanks!

Changed in kdebase-workspace (Ubuntu):
status: Incomplete → Invalid
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.