com2 under 386/ix

Marco S Hyman marc at dumbcat.UUCP
Thu Sep 7 16:35:13 AEST 1989


In article <3064 at amelia.nas.nasa.gov> izen at cwru.cwru.edu (Steven H. Izen) writes:
>
> To enable COM2 I added a line to cd /etc/conf/sdevice.d/asy  to make it look
> like the following:
> 
> asy     Y       1       7       1       3       2f8     2ff     0       0
> asy     Y       1       7       1       4       3f8     3ff     0       0
> 
> (I haven't been able to get the modem control minor device to work-the driver
> just hangs for some unkonown reason when I use minor device 17).

The configuration lines, as given above, do not enable the modem control
ports.  The third field (number of sub-devices supported by driver) must be
changed to a 2 to enable the modem control ports.  Also, I've had problems
depending upon the order the lines are given.  What works (with 386/ix
2.0.2) is:

asy	Y	2	7	1	4	3f8	3ff	0	0
asy	Y	2	7	1	3	2f8	2ff	0	0

After the kernel is built and you've re-booted you should have the following
device files:

crw-rw-rw-   1 root     root       3,  0 Sep  6 22:50 /dev/tty00
crw-rw-rw-   1 root     root       3,  1 Sep  6 22:48 /dev/tty01
crw--w--w-   1 uucp     root       3, 16 Sep  6 22:50 /dev/ttyd0
crw-rw-rw-   1 root     root       3, 17 Sep  6 22:48 /dev/ttyd1

Change your inittab so it fires up a uugetty on the appropriate port.
(Caveat: I've only used ttyd0 -- ttyd1 is untested).  My entry looks like:

	00:234:respawn:/usr/lib/uucp/uugetty -r -t 60 ttyd0 DIAL96

(The DIAL96 is a home grown gettydefs entry).

> One more comment on setup of the incoming port.  I found that I needed to 
> disable the modem's messages (eg. connect 2400) to prevent the modem and 
> uugetty from going into a infinite loop upon disconnect.

This is not a problem with the above setup.  The modem is configured according
to the recommendation in /usr/lib/uucp/Dialers.  I did modify the entries to
do an AT, not an ATZ, for no particular reason.

#   Hayes Smartmodem -- new BNU modem control capabilities
#   If you do not want to strap CD on, use these switch settings:
#
#       S1 - UP		S2 - UP		S3 - DOWN	S4 - UP
#       S5 - UP		S6 - UP		S7 - ?		S8 - DOWN
#
hayes	=,-,	"" \M\dAT\r\c OK\r \EATDT\T\r\c CONNECT \m\c
direct	=,-,	"" \M\dAT\r\c OK\r \c
#
#  Furthermore, you must add a ",M" subfield to the line field (field
#  2) of the associated Devices file entries, as shown here:
#
#	ACU culd0,M - 1200 hayes \T

For the newer modems the switch settings are the equivalent of

	AT &D2 V1 Q E1 S0=1 &C1 &S1

followed by an AT&W to write the configuration to NVRAM.

> "no carrier" and uugetty thought that was an attempt to login, sent characters
> to the modem which were promptly echoed back,etc.  Many months later, while
> bugging :-) ISC tech support about some other problems I mentioned that one.
> I figured they'd tell me how to do it correctly.  They told me to do what I
> described above.  Does anyone out there know a better way?  Obviously turning
> off the messages will break any uucp communications going out on that line, and
> since I read here on the net that ISC claims to have fixed uugetty...

The above worked great on 2.0.1.  However, since upgrading to 2.0.2 (last
night) I've had problems.  Every so often the message

INIT: Command is respawning too rapidly....
id:  00 "/usr/lib/uugetty ...

shows up on the console.  I don't know what triggers it yet as it seems to
show up without any activity on the serial port.  The uugetty is gone and
must be restarted with a kill -1 1.  It will also start itself after a
period of time, but dailins are disabled in the meanwhile.  [DTR is dropped
to the modem so it does not answer the phone if programmed correctly.]

So, when you can find the (missing) documentation you can piece together an
almost working system.  It's not a broken as several people have posted, but
it's not as good as it could be either.

--marc
-- 
// Marco S. Hyman
// ...!pacbell!dumbcat!marc



More information about the Comp.unix.i386 mailing list