Re: rfc: gnome-inetd

[Mark McLoughlin <markmc redhat com>]

Hi,
	The idea does sounds reasonable FWIW.

On Tue, 2004-11-30 at 22:16 -0500, Sean Middleditch wrote:
> On Wed, 2004-12-01 at 01:29 +0000, Alan Cox wrote:
> > > around is the potential performance problem.  Spawning new processes for
> > > each new connection can be fairly heavyweight.  Fortunately, desktop
> > > services don't usually serve a large number of users, and many of the
> > > interesting protocols don't need a lot of reconnecting (even for things
> > > like web sharing, browsers tend to reuse existing connections these
> > > days).
> > 
> > xinetd gets around this a variety of ways. For TCP it allows a daemon to
> > start up and be passed the socket. In essence for a given port it does
> 
> Just to make sure we're on the same page here, you mean that the daemon
> takes over the *listening* socket, yes?  I hadn't thought of that.  Yes
> sir, I like it.  ^_^

	See the "wait" option in the xinetd.conf manpage. Also, Steven's UNPv1
has a good description on how inetd does all this.

> One still needs a separate library so that you can continue to abstract
> away things like access control, and turning the daemon off without just
> blatantly SIGTERMing it (although, I guess, assuming the daemon authors
> did things right, that would of course work properly anyhow) - i.e.,
> when the user disables the service in GConf, the daemon knows to stop
> listening for new connections.

	I don't think you want to get too fancy - tcp wrappers abstracts host
based access control already and its not hard to shut down a daemon when
its unused.

Cheers,
Mark.