Re: [gdm-list] handled=false xservers broken in 2.20.0 vs. 2.18.1

[Brian Cameron <Brian Cameron Sun COM>]

Michael:

What process corresponds to 6774?  Any ideas why process 6774 is dying
with a SEGV (signal 11)?  GDM is probably doing the right thing by
trying to restart if it notices a child is segfaulting.

Brian


> I can't tell what is going on from the debugging information.  It
> forks the x server, and then it looks like it dies 1 second later:
> Nov  2 16:00:30 ubuntu gdm[6781]: DEBUG: gdm_server_spawn: '/usr/bin/X
> :0 -br -audit 0 -auth /var/lib/gdm/:0.Xauth -nolisten tcp vt14'
> ...
> Nov  2 16:00:31 ubuntu gdm[6773]: DEBUG: mainloop_sig_callback: Got signal 17
> Nov  2 16:00:31 ubuntu gdm[6773]: WARNING: gdm_cleanup_children: child
> 6774 crashed of signal 11
>
> Here are files with the default gdm.conf (Ubuntu 7.10) except for handled=false
> http://open-sense.com/downloads/gdm/syslog20.defaultconf
> http://open-sense.com/downloads/gdm/gdm.conf
>
> And just in case anyone is curious here are the original multi-user files:
> http://open-sense.com/downloads/gdm/syslog20.xephyr
> http://open-sense.com/downloads/gdm/syslog18.xephyr
> http://open-sense.com/downloads/gdm/gdm_xephyr.conf
>
> If any other information would be useful please let me know.
> Thanks,
> Michael Pardee
> Open Sense Solutions LLC
> http://open-sense.com
> 888-323-1742
>
>
> On 11/2/07, Brian Cameron <Brian Cameron sun com> wrote:
> > Michael:
> >
> > This might well be a bug.  I'd recommend turning on debug in the
> > GDM configuration, restarting GDM and allow it to fail, then
> > send along the GDM-related debug output that gets sent to your syslog
> > (/var/log/messages or /var/adm/messages depending on your OS).
> >
> > This might highlight why GDM thinks it needs to remanage the display
> > in this case, and assist in debugging your situation.
> >
> > Brian
> >
> >
> > > We've been using gdm 2.18.1 for a while now, and we just tried 2.20.0
> > > in ubuntu 7.10 and our setup breaks horribly.   I went through the
> > > changelog between 2.18 and 2.20, but it looks like A LOT has changed.
> > > If someone could give any ideas on which of the changes may be causing
> > > this problem that would be very helpful instead of trying lots of gdm
> > > version in between 2.18 and 2.20.  Maybe we need different parameters
> > > in the gdm.conf now?
> > >
> > > Basically, if you add handled=false to an x server in 2.20, gdm tries
> > > to start it but it does not leave it up, and eventually you get the
> > > "screen has shut down 6 times in 90 seconds" message.  Taking the
> > > exact same setup and installing gdm 2.18, the handled=false xserver is
> > > started and left up as desired.
> > >
> > > With 2.20 you can see the x server running briefly with ps, and
> > > /var/log/gdm/* and /var/log/Xorg* look normal, it just seems that gdm
> > > is "reaping" the xserver for some reason?
> > >
> > > example:
> > > [server-Standard]
> > > name=Standard server
> > > command=/usr/bin/X -br -audit 0
> > > flexible=false
> > > handled=false
> > >
> > > You're probably wondering why are we trying to start an unhandled x
> > > server?  We run one unhandled x server across many displays, and then
> > > run xephyr on each display for multi-user capability.  While that does
> > > introduce many other variables that could cause problems, the simplest
> > > case of handled=false doesn't work with gdm 2.20.0.
> > >
> > > Any suggestions would be appreciated.
> > >
> > > Thank You,
> > > Michael Pardee
> > > Open Sense Solutions LLC
> > > http://open-sense.com
> > > 888-323-1742
> > > _______________________________________________
> > > gdm-list mailing list
> > > gdm-list gnome org
> > > http://mail.gnome.org/mailman/listinfo/gdm-list