> Rather than *ignoring* carrier detect, the better way is to *reverse*
> the test when opening for dialout:
>
>         DCD low: no one is using the port, go ahead and open it
>
>         DCD high: someone is using the port, block

Uh, so, what happens to that getty process that was waiting for
carrier after tip dials the modem and connects (and thus DCD goes
high)?  Somewhere, you still have to remember that the device is being
used for dialout, so you can prevent the getty from finishing its
open.  (You can't just block all simultaneous opens, because, as I
said before, a dialin session *must* be able to reopen the tty.)  So
the logic is no simpler than for the pure SunOS-style devices, and it
has the serious disadvantage that it's not actually compatible with
any existing system.

That said, the issue is now more or less settled.  I've implemented
and committed SunOS-style dialout devices, with the appropriate hooks
in the com and zs drivers.  People responsible for other drivers
should DTRT, using them as a reference.