Re: X ignores keyboard and mouse input, but shows cursor movement (etch)

[Ralph Katz <ralph.katz@rcn.com>]

On 09/14/2007 03:19 PM, Kelly Clowers wrote:
> On 9/14/07, Ralph Katz <ralph.katz@rcn.com> wrote:
>> A curious and rare (2 times since Etch became stable) X freeze or hang
>> has me wondering what to do.
>>
>> Problem: X ignores keyboard and mouse input, but shows cursor movement
>> and running apps update normally on-screen in visible windows (gkrellm,
>> gaim, etc).
>>
>> What happened: Viewing email message in icedove, right-click to save
>> image from the email, no response to mouse clicks, mouse wheel or
>> keyboard input, but mouse cursor movement appears normal.  icedove
>> continues to fetch new mail on schedule.  The identical symptoms
>> happened once before while composing in icedove.
>>
>> Ssh from another box to examine:  Nothing unusual in /var/log.  Kill the
>> icedove processes.  Screen shows icedove closing, then switches to VT1
>> apparently from a queued earlier attempt.  Keyboard now accepts
>> ctrl-alt-F7 back to X, but again is frozen to all input in X!  Just as
>> it was before killing icedove.
>>
>> >From ssh: Restarted gdm.  All is normal.  This last line in
>> /var/log/Xorg.0.log.old is the only thing I could find that looked
>> different:
>> FreeFontPath: FPE "/usr/share/fonts/X11/misc" refcount is 2, should be
>> 1; fixing.
>>
>> Window manager is openbox.
>>
>> Has anyone experienced this behavior?  Any ideas?  (Next time I'll
>> remember to check .xsession-errors.)
>>
>> Thanks,
>>
>> Ralph Katz
> 
> I have experienced somewhat similar freezes in Unstable. The
> differences are that I am using SeaMonkey 2.0a1pre (Nightlies)
> for web browsing, I have never managed even a VT switch and
> I am using Fluxbox. Oh, and I don't use a session manager
> (GDM, etc). This has happened at least six times.
> 
> I know the Font refcount thing is not related because it always
> says that when I exit X.
> 
> With such a hard freeze, that happens so sporadically, I feel
> (as a non-programmer) helpless to try to figure it out.
> 
> Sorry I can't help.
> 
> 
> Cheers,
> Kelly

Kelly, thanks for responding.  Looking further back in my notes, in
April of 2004, I had the same problem running sid.  In that case, ssh'd
in to kill off processes one at a time to isolate the culprit, which was
gaim.  Problem never re-occurred until what I described above.

And let me clarify that the VT switch was /after/ killing icedove.  So
during the freeze, the box was inaccessible locally.

Regards,
Ralph (also a non-programmer)