Weirdness in gdm shutdown

Ken Moffat zarniwhoop at
Thu Nov 1 15:36:17 PDT 2007

On Thu, Nov 01, 2007 at 02:53:59PM -0700, Dan Nicholson wrote:
> On 11/1/07, Ken Moffat <zarniwhoop at> wrote:
> > 3. X restarts (in my opinion, this is not the expected behaviour).
> >  (in the log, the kernel agp messages appear)
> That's not supposed to happen. Do you see anything interesting at the
> end of Xorg.0.log or Xorg.0.log.old?
 I don't know what I'm looking for, plus the entries don't have
timestamps, and when X restarts it generates a log.  So, after
reboot I have Xorg.0.log.old from the 'bogus' run of X (ends with
RADEONRestoreMemMapRegisters informational message, finished PLL2,
finished PLL1, Restore TV stuff, [drm] messages for the removal of
the context, unmapping, closure, and a refcount message from
FreeFontPath (funny, I thought that message had gone with 7.3, at
least I did't notice it when I was using 'startx').

 The regular Xorg.0.log has everything similar apart from what I've
listed above.

 Unfortunately, I don't understand _how_ X gets re-instantiated (is
that a word?) from a normal logout.
> I'm using gdm-2.18, but I don't think gdm has changed that behavior in
> a while. The real question is why Xorg is restarting and not just
> shutting down. One thing to note is that xorg-server-1.4.0 has quite a
> few bugs that are supposed to be fixed in 1.4.1 in a couple weeks.
> Don't know if that's the issue.
 Like all the xorg development, it seems hard to follow - I found a
post suggesting the problem I had with the 6.7 ati driver was fixed
in the current development tree, so I'm waiting for a new release
there, and I found a message that 1.4.1 was targetted to the 10th, I
think, but no sign of what is happening.

 At the moment, I seem to have 2 options for this - revert gdm to
2.18.4 just in case it's an accidental breakage (but, doing that on a
system already using a newer gdm would leave me only 98% convinced
if the behaviour is unchanged), or building an older xorg-server
with an appropriate Mesa (and hope the ati 6.7.19n driver on this
box works with it, and that not too much else gets broken).
Whichever, I'm unlikely to get started before next week.

das eine Mal als Tragödie, das andere Mal als Farce

More information about the blfs-support mailing list