>No its not wrong at all. If the protocol were restricted per AX.25 connection
>then they would have put the PID into the SABM instead.
Well, the PID is actually a rather odd construct.
First, I don't think the PID was ever part of the CCITT X.25 Recommendation.
Rather, it is specified in X.29, (and mostly in a lot of non-CCITT PAD
documents).
Second, AX.25 could more appropriately be called A-HDLC, in that AX.25
uses only the HDLC portion of X.25 and contains none of the packet-level
protocol of X.25.
I didn't participate in the development of the AX.25 spec, but it seems
reasonable to speculate that the PID was added as a higher-level
protocol demultiplexing mechanism, perhaps without a whole lot of
analysis. I understand that some implementations assume, and probably
not unreasonably, that the PID is constant over the life of an AX.25
connection.
While it may have been appropriate to put the PID some place where it
is specified only once per connection, the AX.25 designers threw away
the most obvious place to put it, namely the X.25 [packet-level] call
request packet.
So, it seems just as reasonable to conclude that the AX.25 PID construct
was not well thought out or perhaps not well documented, as it is to
conclude that you can expect to change the value of the PID over the life
of a connection.
-tjs
