mgetty spawning too fast?

• Messages sorted by:[ date ][ thread ][ subject ][ author ]

Hi,

Jeff Uphoff wrote:
> GD> Could be the modem cable, or the modem setup. Is the user shell terminating
> GD> itself after logout? If not, init won't start a new (m)getty process.
> 
> We've had problems with Zoom modems misbehaving here.  They've sometimes
> needed init strings that look like monkey puke in order to recover from
> such "odd" events as the remote end hanging up.  But then, Gert knows a
> hell of a lot more about this than I do...

Zooms are *beautiful* modems. Right now I'm chasing a bug where mgetty
dumps core when I do fax polling with my Zoom (it won't work becuase the
Zoom is too dumb, but it causes some crazy things to happen inside mgetty,
which shouldn't happen either)...

> GD> Make sure that the modem sets the DCD line properly (AT&C1) and that the
> GD> CLOCAL flag is not set on the serial line (stty -clocal).
> 
> What about "&D2" as well?  (Damn...now I can't even remember what that
> does!?!)

&D2 means "hey, modem, hang up and go to command mode when DTR goes
down". I don't think it's relevant here, but it's a good idea anyway :)

> >> Third, is there any way, at runtime, to tell mgetty where the 
> >> mgetty.config file is located?  
> 
> GD> No, because this would be a security hole (IMHO).
> 
> Unless the user doing it has a *real* uid of 0.  (It should then be OK.)

Hmmm. Nice idea. I'll think about it. It isn't *very* hard to do (though
it's a bit messy inside "conf_mg.c" [which is *very* messy anyway :) ]).

> Sendmail does this sort of trick of course, giving up any privileges
> (gained by its setuid bit) if you specify an alternate configuration
> file at runtime.  But if you're already root then it's kosher...

Hmmm, actually I do not really want mgetty to be mentioned in one
paragraph with sendmail when discussing security... :)

gert

-- 
                                                            //www.muc.de/~gert
Gert Doering - Munich, Germany                             gert@greenie.muc.de
fax: +49-89-3545980 <---new!!!              gert.doering@physik.tu-muenchen.de