Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 5 Jul 1997 00:16:24 +0000 (GMT)
From:      Adam David <adam@veda.is>
To:        ache@nagual.pp.ru (=?KOI8-R?B?4c7E0sXKIP7F0s7P1w==?=)
Cc:        cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG, cvs-sbin@FreeBSD.ORG
Subject:   Re: cvs commit: src/sbin/init init.c
Message-ID:  <199707050016.AAA01667@ubiq.veda.is>
In-Reply-To: <Pine.BSF.3.96.970705033929.22827A-100000@lsd.relcom.eu.net> from "[______ ______]" at "Jul 5, 97 03:47:28 am"

next in thread | previous in thread | raw e-mail | index | archive | help
> > Of course. I mean there might be active ttys which are no longer listed
> > because the file changed.
> 
> If you add/remove getty lines, you suppose to send HUP to init to handle
> updates, if you change non-getty lines, it is very rare case.

Typically, getty spawns login and you end up with a login shell as the
session leader with controlling tty. 'kill -HUP 1' does not affect the
session, but when init is shutting down a HUP should be delivered.

> It not worth
> sending HUP to _all_ processes as done previously.  Such connetions will
> be killed by TERM in any case. 

Agreed, except I'd say it had negative worth.

> To handle such rare cases init must use kvm mechanism to get all process
> with controlling ttys which will be unnecessary bloat for init. 

There are other ways, but they might not be desirable either.

--
Adam David <adam@veda.is>



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199707050016.AAA01667>