NOTICE: /dev/cuaa%d -> /dev/cuad%d renaming

[Poul-Henning Kamp]

In message <16750.46203.217249.805105 at ran.psg.com>, Randy Bush writes:
>> Think of DCD as indicating that a "connection is established" with the
>> other side. The ttyd device is _supposed_ to block until a remote
>> connection is established. If you can't provide a DCD signal, then
>> just use "cuad" instead of "ttyd":
>> 
>> The intention is that's for making outgoing calls you use cuad, so a
>> getty waiting to open ttyd won't stop the open on cuad. In your case,
>> it's probably ok for the getty to just hog the modem itself, so using 
>> /dev/cuad0 sounds like the right thing to do.
>
>it's the getty that is the problem.  this is out of band access to
>the console, i.e. getty.

It is a rather hairy issue.

It is true that /dev/cua* should be used when DCD should be ignored,
but in addition to that we also lock CLOCAL on consoleports in order
to make /dev/tty* for them act essentially the same way.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.