Xref: utzoo alt.sys.sun:3779 comp.dcom.modems:9618
Path: utzoo!utgpu!cs.utexas.edu!wuarchive!uwm.edu!linac!palkovic
From: palkovic@linac.fnal.gov (John Palkovic)
Newsgroups: alt.sys.sun,comp.dcom.modems
Subject: Hardware flow control on dial in
Keywords: uucico RTS CTS SunOS 4.1
Message-ID: <TXEQ6..@linac.fnal.gov>
Date: 8 May 91 04:04:06 GMT
Sender: palkovic@linac.fnal.gov (John A. Palkovic)
Organization: Fermilab, Batavia, IL
Lines: 28

I have a 2400 bd MNP5 modem connected to serial port b of a Sun 4/260
running SunOS 4.1.  Bidirectional hardware flow control (HFC) *is*
enabled in the modem (a Microcom AX/2400c) with AT\Q3.

When I dial *out*, I can enable HFC on the serial port connected to
the modem by sending STTY=crtscts and expecting "" in the chat script.
All works well with the serial port running at 9600 and using the "e"
protocol to connect to an MNP capable modem. I get 280-300 char/s
talking to a T2500 this way.

However, I would also like to have HFC when systems call me on the
same modem, so I can run the serial port and the getty at 9600. The
gettytab(5) man page does not mention anything about forcing HFC. Does
uucico(8) handle setting HFC when run as a login shell? Seems like it
should.

There are two systems that poll me at 2400 MNP4 with the "e" protocol.
I find that the file transfers *to the polling system* always fail if
the serial interface to the modem is running at 9600 bd. Backing off
to 2400 bd they are *much* more reliable, leading me to believe I have
a flow control problem. Any thoughts?

Thanks,

-John

-- 
palkovic@linac.fnal.gov || {royko,tellab5,simon}!linac!palkovic