modems on terminal servers?

• Messages sorted by:[ date ][ thread ][ subject ][ author ]

Hi,

Paul Fox wrote:
> > I can't say for sure, but I think it should work -- we run about 8 modems
> > on a terminal server in a company here, connected over TCP/IP to an AIX
> > box. This is by fare more reliable than directly connected modems
> > (AIX...!).
> > 
> > Mgetty doesn't need to know about this, all is handled by the terminal
> > server software. Thus, there is no support for it in mgetty either...
> 
> so just to help complete my understanding -- 
>  1) does mgetty need full modem control (via the hardware signals) over
> 	the modem?

No. Mgetty assumes that the kernel device drives "do the right thing"
concerning hardware or software flow control. It is a "nice to have" if
you can see and control the hardware lines, but it's not a must (after
all, there are some Unix variants where you don't have access to them
either).

>  2) does mgetty, when it opens the "device", need to be able to do
> 	tty ioctl commands to it?

Yes. At least for setting baud rate, parity, "raw mode". This is a must.

>  3) or is the AT command set sufficient to do everything it needs to do? 

Yes, but the ioctl() commands are necessary to get the serial port into a
state where you can *talk* AT commands and reliably get the responses.

> if the answer to 1 or 2 is yes, then i need to run the vendor's pty-creator,
> which makes a pty that "shadows" the real port on the terminal server. if
> the answer to 3 is yes, i can run a much simpler daemon that transfers data
> but not ioctl information (and that i have source for :-).

I think you will need the pty software...

gert
-- 
USENET is *not* the non-clickable part of WWW!
           //www.muc.de/~gert/
Gert Doering - Munich, Germany      gert@greenie.muc.de
fax: +49-89-3545980     gert.doering@physik.tu-muenchen.de
.