Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!nrl-cmf!ames!oliveb!apple!rutgers!bellcore!faline!thumper!ulysses!att!lzaz!hcj
From: hcj@lzaz.ATT.COM (HC Johnson)
Newsgroups: comp.os.minix
Subject: Re: problems with rs232 on minix ST 1.0
Summary: Heres what going on
Keywords: ST minix rs232 help
Message-ID: <459@lzaz.ATT.COM>
Date: 8 Mar 89 16:17:01 GMT
Organization: AT&T ISL Lincroft NJ USA
Lines: 37

In article <122@dftsrv.gsfc.nasa.gov>, stailey@iris613.gsfc.nasa.gov (Ken Stailey) writes:
> 
> I have run into some snags with rs232.  When the system boots it works fine.
> I can type "cat -u </dev/tty1" and get normal results.  But after I run my
> terminal program (which I'm using now)  it can get very messed up and won't
> let me read from /dev/tty1 (writes always work fine).  I think that it had
> to do with not closing /dev/tty1 after the program terminated.
> 
> Is there an ioctl (or something) that will reset this?
> 
> 							Concerned,
> 							 Ken

Two undocumented ioctls affect the rs232; TIOCSTART, and TIOCSTOP.
This is what happened.  Minix lacks the concept of how many times a device
is opened, so there is no way to due a 'final' close.  For example, 
cat -u < /dev/tty1 will do 4 opens and 3 closes before reading a single byte.

I added TIOCSTART and TIOCSTOP when i would run a separate program to 
send these.  This seemed awkward, so I attempted to get rs232 to start up
better on open.

Also, cu has problems, especially if you do not terminate it with ~. .

If all else fails, and you need to run cat -u < /dev/tty1 after cu, I
would think that
	~%local /bin/cat -u
executed from within cu will do the job.  As the shell executes this line
I would also expect 
	~%local /bin/cat -u > myfile
to work.  But don't hold me to it.


Howard C. Johnson
ATT Bell Labs
att!lzaz!hcj
hcj@lzaz.att.com