[comp.unix.questions] Call-in & call-out on the same line

dwm@fibercom.COM (David W. Minnich) (09/28/89)

My system: Sun 3/160, 3.5, w/ Telebit TB+

I need to use the same line for both call-in and call-out.  According to 
everything I read in TFM, this should be possible, but no luck yet.  I have
seen references to uugetty, which may be what I need.  Is uugetty PD?  Where
is it available?  If you have solved this problem on a Sun (or something
similar), or if you have suggestions on where I should look, I'd appreciate
the help.

Please reply by email to the address below; followups have been directed to
comp.unix.questions.

--dave
-- 
David W. Minnich               INTERNET: dwm@fibercom.com
FiberCom, Inc.                 UUCP: ...!uunet!fibercom!dwm   
P.O. Box 11966                 FAX: (703) 342-5961
Roanoke, VA  24022-1966        PHONE: (703) 342-6700, (800) 423-1183

news@fps.com (NetNews Administrator) (09/29/89)

In article <1912@fibercom.COM> dwm@fibercom.COM (David W. Minnich) writes:
>
>My system: Sun 3/160, 3.5, w/ Telebit TB+
>
>I need to use the same line for both call-in and call-out.  According to 
>everything I read in TFM, this should be possible, but no luck yet.  I have
>seen references to uugetty, which may be what I need.  Is uugetty PD?  Where
>is it available?  If you have solved this problem on a Sun (or something
>similar), or if you have suggestions on where I should look, I'd appreciate
>the help.
>
>Please reply by email to the address below; followups have been directed to
>comp.unix.questions.
>
>--dave

Actually, it's pretty easy on a Sun.  You have to build some cu devices.

If you have:

	crw-rw-rw-  1 root      12,   0 May 10 00:13 /dev/ttya

Just use mknod to make:

	crw-rw-rw-  1 root      12, 128 Sep 24 18:04 /dev/cua0

These refer to the same physical port.  Run your getty on ttya and
do you dialouts on cua0.  The acu device is always has a minor
device number, 128 greater than the normal device.  So you could also
have:

	crw-rw-rw-  1 root      12,   1 May 10 00:13 /dev/ttyb
	crw-rw-rw-  1 root      12, 129 Sep 24 18:04 /dev/cua1 

Etc.

Make sure that your modem has echo and result codes turned off or
it will blow your getty's mind.  When a user opens the acu device
the getty, which was blocked on carrier, just keeps on blocking.

When you close the acu device, it can take login's again.

If only it worked so easily on a VAX.

--Bill

guy@auspex.auspex.com (Guy Harris) (10/05/89)

>Actually, it's pretty easy on a Sun.  You have to build some cu devices.

*And* you might have to reconfigure your kernel to turn "soft carrier"
off for the port(s) in question; that's a common cause of failure when
people try to use the dial-in/dial-out device.  (In a future release, I
expect this to be done by a flag in "/etc/ttytab", so you don't have to
reconfigure your kernel.  Of course, you don't want to run a generic
kernel any longer than you have to, anyway....)

williams@cs.umass.edu (12/01/89)

I saved a discussion in the group from Sept/Oct concerning trying to use the
same terminal line for dailin/dialout on a Sun.  I am now trying to get this
working myself and am having problems.  I wonder if anyone can suggest a
solution from the following description? 

Sun 3/60 running SunOS V4

The problem is that if ttyd0 is turned on in /etc/ttytab, a getty is
sitting on the line and tip complains that all ports are busy even
though tip is using device cua0 and getty is using ttyd0.

ps shows:

 7099 d0 I     0:00 - Fast-Dial-1200 ttyd0 (getty)

Now, the manual says "You should not see a getty on the port, except
momentarily as the port receives a call".  This is not true in my
case, but there are a number of mistakes in this section of the manual
and another posting on the subject said that the getty running on
ttyd0 would be "blocked" when cua0 was in use.  Which is correct?
Should the getty always be there or not?  If I turn line ttyd0 off in
/etc/ttytab then I can dialout with tip umass.  I have checked for
lock files in /var/spool and /var/spool/uucp.  For testing purposes, I
do not now have a modem plugged into the port so there are no
characters being sent to the port nor is carrier detect being asserted.

A description of pertinent files and changes to the system follows:

The kernal has been reconfigured for hardware carrier detect via:

device		zs0 at obio ? csr 0x20000 flags 0x2 priority 3

and installed.

ttya was moved to ttyd0 and a node was made for cua0:

crw-------  1 uucp      12, 128 Nov 29 20:16 cua0
crw--w--w-  1 root      12,   0 Nov 29 20:16 ttyd0

The following changes were made to /etc/ttytab:

ttya	"/usr/etc/getty std.1200"	unknown		off
ttyd0	"/usr/etc/getty Fast-Dial-1200"	dialup		on

followed by kill -1 1.

The file /etc/ttytype did not exist on my system so I created it:

dialup ttyd0

My /etc/remote is as follows:

umass:\
	:pn=5453090%:tc=VMS-1200:
UNIX-1200:\
	:el=@:du:at=hayes:ie=:oe=:br#1200:tc=dialers:
VMS-1200:\
	:el=@:du:at=hayes:ie=#$%:oe=:br#1200:tc=dialers:
dialers:\
	:dv=/dev/cua0:
tip0:\
	:tc=UNIX-1200:

Thanks for any help!
Leo Pinard c/o williams@cs.umass.edu

williams@cs.umass.edu (12/01/89)

Thanks to guy@auspex.com, my problem is solved!

In article <7179@dime.cs.umass.edu>, I write...
"The problem is that if ttyd0 is turned on in /etc/ttytab, a getty is
"sitting on the line and tip complains that all ports are busy even
"though tip is using device cua0 and getty is using ttyd0.
" 
"ps shows:
" 
" 7099 d0 I     0:00 - Fast-Dial-1200 ttyd0 (getty)
" 
"Now, the manual says "You should not see a getty on the port, except
"momentarily as the port receives a call".  This is not true in my
"case, but there are a number of mistakes in this section of the manual
"and another posting on the subject said that the getty running on
"ttyd0 would be "blocked" when cua0 was in use.  Which is correct?
"Should the getty always be there or not?  If I turn line ttyd0 off in
"/etc/ttytab then I can dialout with tip umass.  I have checked for
"lock files in /var/spool and /var/spool/uucp.  For testing purposes, I
"do not now have a modem plugged into the port so there are no
"characters being sent to the port nor is carrier detect being asserted.
" 
It turns out that on the Sun, carrier detect floats high so with no modem
connected the getty unblocks and ties up the line.  When I plugged in the
modem it worked fine.  I had tried it before some time ago with the modem
and it didn't work.  Guess I fixed something else along the way since I
last tried it with the modem.

The correct state is indicated by ps as

  7099 ? I     0:00 - Fast-Dial-1200 ttyd0 (getty)

instead of what I had previously

  7099 d0 I     0:00 - Fast-Dial-1200 ttyd0 (getty)

The question mark in the tty column indicates that the getty isn't actually
on a port yet.  When the modem answers a call, Carrier Detect gets asserted
and the getty gets attached to the line.

Thanks again, Guy!
Leo Pinard c/o williams@cs.umass.edu