Date: Wed, 5 Feb 1997 15:00:20 -0700 (MST) From: Terry Lambert <terry@lambert.org> To: davidn@labs.usn.blaze.net.au (David Nugent) Cc: terry@lambert.org, freebsd-current@FreeBSD.org Subject: Re: getty patches Message-ID: <199702052200.PAA15711@phaeton.artisoft.com> In-Reply-To: <19970205153554.HM65449@labs.usn.blaze.net.au> from "David Nugent" at Feb 5, 97 03:35:54 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> > > If the init chat fails, it logs via syslog and exits. There is no > > > other sane thing to do. > > > > So is the initd expected to say "getty repeating too quickly on > > port %s, sleeping" for this? > > If GETTY_SPACING in init.c was to be increased from 5 seconds, yes. > But this may have undesirable side-effects, and I'm not sure if I'd > want this behaviour in any case. It's 10 seconds according to the man page. 5 seconds id the average for a process coming in a rndom interval of 0-10 seconds following a sleep event. So 5 seconds is average, but the worst case is double that. > > Waiting an average of 5 minutes for a login when this happens (while > > initd is busy sleeping it off) seems a bad thing. > > ? I meant "seconds". [ ... /etc/issue ... ] > > I thought different transports could be given different classes > > in the BSDI implementation. > > No. This is seen as a "service" to the authenticator. Login classes > may have different combinations of service/authenticator values, > but which ones are valid are still driven by the user's login > class. OK, objection withdrawn. > > > Any way you like, Terry. If you don't wish to use the inbound call > > > features, then don't. That's why it is optional. :-) A blocking > > > open() is still done unless you enable the "ac" capability. > > > > Ah. I see. So I can use the same modem inbound vs. outbound, > > as long as I turn this feature off. > > "Unless you enable it in the first place" is a more accurate > picture. Right, but it's a desirable feature, so I'll probably have it on. 8-) I guess the problem is that I can't have both desirable features at the same time, unless I go to the "open-completes-on-RI-signal" model, or unless I fully virtualize the modem and use DCD inbound to lock outbound access (EWOULDBLOCK instead of hanging for the chat to get out of the way). Regards, Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199702052200.PAA15711>