*BSD News Article 19553


Return to BSD News archive

Newsgroups: comp.os.386bsd.misc
Path: sserve!newshost.anu.edu.au!munnari.oz.au!constellation!convex!convex!convex!cs.utexas.edu!geraldo.cc.utexas.edu!portal.austin.ibm.com!willcox!bob
From: bob@willcox.uucp (Bob Willcox)
Subject: Re: Using the sio ports for terminals w/o modem ctl signals
Message-ID: <CBv07B.Hzy@willcox.uucp>
Date: Mon, 16 Aug 1993 15:54:47 GMT
References: <CBJnzu.6nG@obiwan.uucp> <1993Aug10.193616.16846@fcom.cc.utah.edu> <24grraINNl92@kralizec.zeta.org.au> <1993Aug15.232601.17314@fcom.cc.utah.edu>
Organization: /etc/organization
Lines: 42

In article <1993Aug15.232601.17314@fcom.cc.utah.edu>,
A Wizard of Earth C <terry@cs.weber.edu> wrote:
>In article <24grraINNl92@kralizec.zeta.org.au> bde@kralizec.zeta.org.au (Bruce Evans) writes:
>>In <1993Aug10.193616.16846@fcom.cc.utah.edu> terry@cs.weber.edu (A Wizard of Earth C) writes:
>>There might still be problems from broken programs such as getty blindly
>>turning CLOCAL off.
>
>Why would getty leave CLOCAL set on a modem?  I don't understand.  If DCD
>drops, my applications are supposed to get SIGHUP.  You are suggesting that
>it's OK to leave my root shell up there for the next caller.  It's possible
>that the DCD dropped not because I was stupid and didn't log out first, but
>because I had a power failure on my end but not the host end.  Is it still
>implied to be "my fault" that the next caller gets root access to the machine?

For my situation, getty is not talking to a modem.  My original objective
in starting this slightly diverged thread was to obtain advice on
connecting my *local* terminals up using sio via a board that is
brain-damaged w/o modem control signal support.  (I knew this at the
time I bought it, but driving local terminals is all I wanted it for.)
It sounds like (from what Bruce is saying) that if I cooerce getty to
not turn off CLOCAL that it should work.  I am going to give this a try.

>>>The correct way to handle this is to have a different driver for the board
>>>so that it ignores it before making requirements on the board at all; if the
>>>board provides these signals as "floating", it's a broken board.  From
>>>what I remember of the thing, it doesn't, and you *do* need a different
>>>(or at least indifferent 8-)) driver.

I don't really want a different driver.  It simplifies my life to use
the same one (if possible).

>>It is sufficient to have a non-broken board and/or cables with DCD wired on.

Unfortunately, my board is broken and there doesn't appear to be a way
to do this with it.


-- 
Bob Willcox                                Internet: bob@willcox.austin.ibm.com
IBM Advanced Workstations & Systems        IBM RSCS: WILLCOX at AUSTIN
Austin, TX 78758                           Phone:    (512) 838-3914
Since I do NOT speak for IBM, any opinions IBM T/L:  678-3914