Re: trpt(8)

[Greg Troxel <gdt%ir.bbn.com@localhost>]

Joerg Sonnenberger <joerg%britannica.bec.de@localhost> writes:

> is there any reason to preserve trpt(8)? It feels next-to-useless when
> compared to tcpdump...

I'm not sure, but something like that can/could be way more useful.  In
particular, it could log the reason retransmission, cwnd changes, etc.
I have an implementation of this (called PROTOTRACE) that lets the TCP
code put notifications of interesting events into a bpf tap, and has
tcpdump print it.  Then, when one run runs tcpdump2xplot you can see
cwnd and other interesting stuff.  I have permission to release this but
haven't had the cycles.   So given that, trpt may well be outdated and
pointless.  And it may not work usefully enough to be worth keeping now
- I just wanted to point out that protocol state chagnes are in theory
richer than packet captures.