Mailinglist Archive: opensuse-factory (396 mails)

[Kay Sievers <kay.sievers@xxxxxxx>]

On Mon, 2011-01-10 at 16:55 +0100, Vincent Untz wrote:
> Le dimanche 09 janvier 2011, à 23:08 +0100, Bernhard M. Wiedemann a écrit :
> > The extra-repo testing so far had turned up a segfault in
> > KDE-Playground's amarok that already got fixed
> > and a more complicated problem in GNOME-devel's gdm concerning
> > VT-allocation and Xorg.
>
> Part of this is a bug in Xorg that was eating 100% of the CPU, and after
> some more investigation, the fix is now submitted (sr#57731).
>
> But that's not the full fix. Since gdm lets Xorg decides on which vt to
> start, Xorg chooses the first free one. And since Xorg is started before
> the end of the boot process, and before init starts the mingetty
> processes, we end up with Xorg on vt2. And a bit later, a mingetty is
> started on tty2 (following what's in /etc/inittab).
>
> The issue is that whatever you type in X is then also sent to mingetty,
> and you can see failed attempts to login in /var/log/messages.
>
> (Note: it works fine with systemd)
>
> I'm not sure what's the proper fix here: should init check there's
> nothing on tty2 before starting mingetty there? Or should we just force
> Xorg to start on vt7 anyway?

I guess we should always force X/gdm to be on the active tty, regardless
of any splash logic, and along with that just never start getty1 in
runlevel 5.

Fedora officially does not support booting without plymouth, hence they
have such logic always applied.

Kay

-- 
To unsubscribe, e-mail: opensuse-factory+unsubscribe@xxxxxxxxxxxx
For additional commands, e-mail: opensuse-factory+help@xxxxxxxxxxxx