Re: Alt+F7 fails to switch to graphical console (X server) from a text console under some conditions
On Mon, 8 Feb 2010 06:26:33 -0500 (EST), Tom Furie wrote:
> I haven't filed a bug report because I'm not sure what to file it
> against. I'm not even sure it could be called a bug, I don't have
> anything in my config to tell GDM to specifically start on tty7, so it
> just grabs the first available one.
/var/log/Xorg.0.log tells you what console it started on, and the flags
indicate that the option is supplied on the command line. For example,
Markers: (--) probed, (**) from config file, (==) default setting,
(++) from command line, (!!) notice, (II) informational,
(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
...
(++) using VT number 7
The double plus sign is the marker for "from command line". So if you
find the script that starts the X server and supplies that command line
option, you'll be well on your way to finding the bug. The next step
will be to find the logic that "calculates" this command line switch.
And if you find the faulty logic you'll know which package to report a
bug against. (i.e. which package is this script file shipped with?)
And you'll even be able to supply a patch file with the bug report.
I'll start looking too, but I won't have access to that machine until
this evening. Maybe a good place to start looking would be
/etc/init.d/gdm.
> You aren't restarting X or GDM, but if GDM started on tty8, because tty7
> was unavailable for whatever reason, the first you would know about it
> is when you tried to switch back to it from another tty.
Good point. And a reboot always seems to cure it. And that makes sense,
because following a reboot, virtual terminal 7 is always available.
But if I logout and login again, that restarts the X server. Hmm.
Maybe all vestiges of the old copy of the X server have not yet been
completely removed from the system at the time the logic executes which
tries to determine the next available virtual terminal. I wonder if
this has anything to do with the new dependency-based booting? Or
are we there yet? I also wonder if I issued
/etc/init.d/gdm stop
/etc/init.d/gdm start
(with a wait time of a few second between the two commands) instead of
/etc/init.d/gdm restart
if that would make a difference? (Of course the /etc/init.d/gdm
script has to be invoked from a text console as root.)
Reply to: