UUCP and RTS/CTS handshaking

[asfy at vax5.cit.cornell.edu]

In article <1991Jun22.180055.13559 at jpradley.jpr.com>,
jpr at jpradley.jpr.com (Jean-Pierre Radley) writes: 
> In article <11680 at hub.ucsb.edu> 6600joef at ucsbuxa.ucsb.edu (Joe Foster) writes:
>>	Help! I'm trying to get uucp, cu, etc. to talk to my high-speed
>>modem using hardware (RTS/CTS) flow control instead of using XON/XOFF.
[...]
>>
>>P.S. The modem is a ViVa model 9642e 9600 baud V.42bis, external, and is
>>connected to a Pioneer Vantage 486/25 through a serial card using NS16450
>>UART's. uucico is talking to the modem through the tty2A modem-control port.
>>My XENIX version is 386GT 2.3.3, with SLS's xnx129b (updated UUCP binaries)
>>and xnx155b (lots o' screwy changes) installed.
> 
> 
> I've checked into this same problem with uucp on SCO Xenix and SCO Unix.
> My dialTBIT.c source clearly sets RTS and CTS. If I do an stty on the dial port
> just as uucico has called dialTBIT, I can see this setting. It also promptly
> goes away, so I suspect that SCO's uucico, once addressing the connection,
> resets RTS and CTS off.

Compare the maximum window size in UUCP (7 * 64 bytes + protocol
overhead) to the size of io buffers in MNP or V.42 modems, and
you won't see much to worry about.  The g protocol itself does
"flow control" that is adequate for most situations.

  --jh