From owner-cvs-all Tue Jul 4 17: 2:45 2000 Delivered-To: cvs-all@freebsd.org Received: from localhost (localhost [127.0.0.1]) by hub.freebsd.org (Postfix) with ESMTP id 37AFC37BE67; Tue, 4 Jul 2000 17:01:57 -0700 (PDT) (envelope-from green@FreeBSD.org) Date: Tue, 4 Jul 2000 20:01:55 -0400 (EDT) From: Brian Fundakowski Feldman X-Sender: green@green.dyndns.org To: Robert Watson Cc: Wes Morgan , cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG Subject: Re: cvs commit: src/crypto/openssh sshd.c In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Tue, 4 Jul 2000, Robert Watson wrote: > > I must have missed that commit while catching up from travel. I don't > like that technique for restarting daemons -- for one thing, it relies on > /proc being mounted. I have a continuing strong objection to the presence > of "file" in /proc in the first place, as I don't feel it substantially > contributes to useful functionality in /proc. While the switch from vnode > reference to symlink is an improvement, I'm still not sure any of the > associated complexity is a good idea. Does anything actually rely on this > behavior, other than sshd now? No, no, you two are mistaken here; this is a fallback if the execv() fails for the saved argv. It works like this: try to exec what we were invoked with. try to exec what we were invoked with _determined by the file system_, using /proc give up. The middle step is what I added. There's no reliance on procfs. procfs exists as a convenience, and in this case it makes it possible for you to HUP sshd even when you invoked it using a non-canonical name. This is a big convenience when you start daemons interactively. It's Good (TM) :) -- Brian Fundakowski Feldman \ FreeBSD: The Power to Serve! / green@FreeBSD.org `------------------------------' To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message