V6, as distributed, had no networking at all. There
are two V6 systems with
networking in TUHS:
The first is an 'NCP' Unix (unless unless you have an ARPANet); the second is
a fairly early TCP/IP from BBN (ditto, out of the box; although one could write
an Ethernet driver for it).
I’ve also done a port of the BBN VAX stack to V6 (running on a TI990 clone), using a
PPP interface to connect. Experimental, but may have the OP's interest:
There's also a fairly nice Internet-capable V6
(well, PWB1, actually) from MIT
which I keep meaning to upload; it includes SMTP, FTP, etc, etc. I also have
visions of porting an ARP I wrote to it, and bringing up an Ethernet driver
for the DEQNA/DELQA, but I've yet to get to any of that.
I’d love to have a look at that and compare and contrast the approaches.
I’m finding that BBN’s original design, with a separate kernel thread for the network
is elegant but difficult to tune: too much priority and it crowds out user processes, too
and the slow PPP line is not kept busy.
I think I’m beginning to understand why CSRG (and later also BBN) moved to
the interrupt driven structure of 4.2BSD: perhaps it was also difficult to tune for a
VAX with ethernet.