rcNG problems on RELENG_5

[Mike Makonnen]

On Tue, Sep 28, 2004 at 11:14:10AM +0200, Jose M Rodriguez wrote:
> On Tue, 28 Sep 2004 11:14:16 +0300, Mike Makonnen <mtm at identd.net> wrote:
> 
> There is a last process in the rc boot: init that launch gettys via  
> /etc/ttys.
> xdm really needs gettys running when it comes up, this is a race condiction
> between xdm an init.
> 
> I've got problems with this since /usr/X11R6/etc/rc.d/gdm.sh hit the  
> tree.  With
> /etc/rc.d/xdm is even worse.
> 
> rcGN really lacks BEFORE tags that isolates a 'near last' stage where just  
> that things with gettys race problems lives. ?Why to Solve?
> 
> - By a new 'syncpoint' before LOGIN and moving to this all those things  
> that are run at LOGIN and _not_ related to gettys or similar.
> 
> - By a a new 'syncpoint' after login (LAST?) that make possible put things  
> for run really as the very last of rcorder.
> 
> After renaming xdm to 0xdm, my race problems gone away, but I thing this  
> must be take in account for rcNG developpment.

>From what I gather your problem is that you can't position the xdm script to
win the race with getty. Correct? If so, you're going about it the wrong way.
You shouldn't be looking at strategies to win the race. Rather, you should
be looking at ways to eliminate the race itself.

The first step is to understand
why xdm depends on init starting the gettys. For example I use gdm and it's
started from a startup script. My /etc/ttys has ttyv8, the tty on which gdm
starts, turned off. So, how does init starting the gettys on the other terminals
affect X? It seems to me that the problem really isn't
with init's starting of gettys but some other side-effect of that.

In short, we should understand EXACTLY what the problem is before going
off and trying to solve it.

Cheers.
-- 
Mike Makonnen  | GPG-KEY: http://www.identd.net/~mtm/mtm.asc
mtm at identd.net | Fingerprint: AC7B 5672 2D11 F4D0 EBF8  5279 5359 2B82 7CD4 1F55
mtm at FreeBSD.Org| FreeBSD - Unleash the Daemon !