Subject: Re: Downloading code to IO processor
To: Jonathan Stone <jonathan@DSG.Stanford.EDU>
From: Martin Husemann <martin@rumolt.teuto.de>
List: tech-kern
Date: 06/18/1997 07:21:28
Jonathan wrote:

> Everything you say about devices being unimportant `outside of [isdn]
> management' could apply equally well to IP/TCP. No? 

Yes, you are right. Its very similar and indeed there are names for the
cards. It comes down to the question discussed on current-users some
time ago: is it a good idea to special case network interfaces and don't
have device nodes for them. Seems like this is the next special case here
and the same decision has to be made.

And of course the "card" is important in one way in the ISDN case: multiple
cards may be attached to the same or different ISDN's, and thus share or not
share the available B channels for each BRI/PRI. A fact that should be known
to the ISDN management. Similar to ethernet interfaces on the same wire, also
the network code doesn't care much (but probably could, say for load
balancing).

So, one could argue twofold:

 1) The ISDN cards are very simmilar to network interfaces and could be
    handled similar, without creating device nodes for them.
 2) The ISDN cards are very simmilar to network interfaces and the whole
    "we don't need device nodes for network interfaces" should be cleaned
    up before another instance of the same design comes up

Another point that comes to mind here: it has been proposed to treat
ISDN just as another link level network interface, create a socket address
variant carying ISDN phone no's and use the full socket interface as the
API of choice. This would fit nicely here...


Martin