>I intended to explicitly suggest making the destination address a
>parameter to your ip_newid() function.

It needs both source and destination, if I passed anything.

To be honest, I'd prefer a scheme with a (large) hashtable on src/dst
IP address; each hashtable bucket would be a doubly-linked list, with
the src/dst IP address, the ip_id, a refcount so that pcbs can store a
pointer to the right list entry, plus a TTL, so we can recover
long-idle entries (finding the right pcbs is another matter).

I'd anticiapte thousands to tens of thousands of active entries, for
the applications I used to run 3-4 years ago.

Getting all the book-keeping and pcb rundowns right is more than I
wanted to buy into before we start the 2.0 release cycle.


>That allows for various types of ID generation - including the
>single incremented counter (where the arg address is simply ignored).