| > Now you and me are in the exact same position.  the board shows up during
| > boot (and we even have the same address and kernels now sincy you have
| > 1.3.2 generic also), we can see gettys running for all eight ports
| > (std.9600), and still the ports are completely dead.  I tried the exactly
| > same ttys file on my laptop 386 (4 meg ram - *shudder*), (but only two
| > entries ofcourse, only 2 ports on that thing, and devicenames changed
| > accordingly), and when getty:s were running, the com ports echoed chars
| > sent at them in 9600, but there was no sign of a login banner or nothing
| > like that. if I sent a Ctrl-D at the port, the getty for that port would
| > die and disappear until reboot. This, obviously does not happen on the
| > Vax, whose ports are completely silent and deaf.

Okay, let me see if I can understand your i386 experience.  If things are
working as I expect them to, you should have had login banner(s) appear
on any enabled tty line with a getty running.  That is what getty lives
for, i.e. to notice state changes on tty ports and provide a login banner
(or so I am of the understanding).

Do you ever seen anything "interesting" written into /var/log/messages
WRT getty?  E.g. that init is seeing problems with a given line?

So when you say cat'ing data at the tty port on the laptop with a terminal
attached - does that mean those characters appeared on the remote terminal?

Where did you "sent a Ctrl-D" from?  The laptop redirected at the tty port,
or via the terminal attached to the port?  The fact that "getty" died and
did not come back is not right - look in /var/log/messages for hints as
to what may have happened.

This sort of "partly working" behavior makes me wonder whether something
is hanging on Rx waiting for say DTR to become active or something.  The
modem control signals are often used and this is often why the old saying
goes there are a million different serial cables necessary for all the odd
combination of signals you need to either loop-back, leave open, or supply
any particular signal, e.g. CD, DTR, RTS etc.

This is why having one of those serial snoopers/break-out boxes can be
of serious usefulness.  I mean from as simple as just knowing the line
has life (i.e. some LEDs light up, indicating it is at least "plugged in"
so to speak) to noticing which had Tx or Rx up so you know that your
"other end" needs to "light up" the one that isn't to be able to talk
(what straight vs null-modem cables are about - or whatever name folks use).

-- Ken