bblue@crash.cts.com (Bill Blue) (04/08/90)
In article <36@mnopltd.UUCP> gatech!stiatl!mnopltd!neal writes: >tty1A has modem control; tty1a does not. You want to enable tty1A in the >ttys file. You DO want to specify the terminal type in ttytype as "modem". >This causes various bits of magic to occur. What bits of magic, specifically? Unix only or Xenix too? --Bill
neal@mnopltd.UUCP (04/08/90)
->In article <36@mnopltd.UUCP> gatech!stiatl!mnopltd!neal writes:
->>tty1A has modem control; tty1a does not. You want to enable tty1A in the
->>ttys file. You DO want to specify the terminal type in ttytype as "modem".
->>This causes various bits of magic to occur.
->
->What bits of magic, specifically? Unix only or Xenix too?
Oops! I get to correct my own posting! After this got loose I realized that
the terminal type should be "dialup", NOT "modem". As far as magic, I noticed
that when the port is enabled or disabled, the revelant dialer program is
called to poke the modem into or out of auto-answer state. Beyond that, I
can't comment except to say the whole thing works properly in a bi-directional
persuasion. Since ttys/ttytype is a BSD/xenix aberration, this would not
be the case in more formal AT&T (feh!) versions of unix and I can't comment
on SCO Unix.
------------------------------------------------------------------------------
Neal Rhodes MNOP Ltd (404)- 972-5430
President Lilburn (atlanta) GA 30247 Fax: 978-4741
uunet!emory!jdyx!mnopltd!neal Or uunet!gatech!stiatl!mnopltd!neal
------------------------------------------------------------------------------bblue@crash.cts.com (Bill Blue) (04/09/90)
In article <40@mnopltd.UUCP> gatech!stiatl!mnopltd!neal writes: > >->In article <36@mnopltd.UUCP> gatech!stiatl!mnopltd!neal writes: >->>tty1A has modem control; tty1a does not. You want to enable tty1A in the >->>ttys file. You DO want to specify the terminal type in ttytype as "modem". >->>This causes various bits of magic to occur. >-> >->What bits of magic, specifically? Unix only or Xenix too? > >Oops! I get to correct my own posting! After this got loose I realized that >the terminal type should be "dialup", NOT "modem". As far as magic, I noticed >that when the port is enabled or disabled, the revelant dialer program is >called to poke the modem into or out of auto-answer state. Beyond that, I >can't comment except to say the whole thing works properly in a bi-directional >persuasion. Since ttys/ttytype is a BSD/xenix aberration, this would not >be the case in more formal AT&T (feh!) versions of unix and I can't comment >on SCO Unix. I thought this might be what you were hinting at. In my experience with Xenix, the terminal type specified in ttytype has nothing to do with bi-directional operation, only with the default terminal type that may be associated with that port, and then only if you use tset creatively in your .profile or .login. True bi-directional port operation is more or less automatic when the port is enabled and then used for outgoing traffic. --- One really interesting Xenix behavior I've noticed with bi-directional ports and never seen discussed anywhere is that the modem turnaround (dial -h or equivalent) that resets the modem for dial-in operation after a call has completed, is sensitive to entry placement in the Devices file. It seems that you must place the major (modem control) device entry (ACU entry) before the minor (no modem control) entry for the same device if you have both entered in the file. For example: ACU tty1A - 2400-19200 dialHA24 Direct tty1a - 1200-19200 direct ACU tty2A - 1200-4800 dialHAHST Direct tty2a - 2400-19200 direct If the Direct entry is not there for that device, it will reset normally at call completion. Likewise if it appears AFTER the ACU entry. But if you place the Direct entry first: Direct tty1a - 1200-19200 direct ACU tty1A - 2400-19200 dialHA24 the modem turnaround will not occur. Everything else will appear to function normally. Anyone know why this is? --Bill