Permalink
Browse files

First major import from Bruce Simpson (bms)

Includes:  Options Support, TLV updates, IGMPv2, IGMPv3 and other rototilling
  • Loading branch information...
gvnn3 committed Apr 23, 2008
1 parent 795c53c commit 054173f3824db84fb7a3d7879574b013ca309f46
@@ -7,8 +7,14 @@ syntax: glob
*.ps
*.dvi
*.pdf
*.swp
build/*
docs/Web/*
docs/latex/*
docs/man/*
make.out
pcs/pcap/config.h
pcs/pcap/config.pkl
pcs/pcap/pcap.c
tests/etherdump.out
tests/pcaptest.dump
@@ -1,3 +1,5 @@
All contributors are listed in chronological order
John-Mark Gurney (jmg at freebsd.org)
Li-Wen Hsu (lwhsu at lwhsu.ckefgisc.org)
@@ -9,3 +11,6 @@ Kip Macy (kmacy at freebsd dot org)
Tom McLaughlin (tmclaugh at sdf.lonestar.org)
Chuck Tuffli (ctuffli at users.sourceforge.net)
Bruce M Simpson (bms at freebsd.org)
34 TODO
@@ -1,3 +1,37 @@
TODO for PCS
*) Eliminate the assumption that a TLV or LV always encodes its length
in bytes. This is inconsistent with the definition of Field.
*) Eliminate the assumption that a TLV always contains itself or
indeed counts itself in its length, see below.
*) In IGMP, add a pointer to the type numbers:
http://www.iana.org/assignments/igmp-type-numbers
Think carefully about how to deal with the fact that at least
6 different protocols use "igmp" as their encapsulation.
Probably an IGMP "map" is best, as the header fields are often
never referred to outside of their own context.
*) Fix the pretty-printing for TCP options.
*) Fix the lack of variable-width bounds in the PCS meta-language.
At the moment, LengthValueFields and TypeLengthValueFields have
no notion of a variable width, so they always encode and decode
as fixed length fields. This makes it very difficult to capture
most packet structures correctly in PCS.
*) Fix the IP4Connector so that it will work and can be reliably used
instead of the PcapConnector. Without an ARP module, going straight
to Ethernet is tedious, and doesn't even work for other link layers.
*) Reimplement the IP-layer diagnostic tools to use the IP4Connector.
*) Come up with a fix for BPF loopback injection on FreeBSD and push the
fix upstream.
*) Add many many more packets
*) Teach BPF and pcap how to request allmulti.
Oops, something went wrong.

0 comments on commit 054173f

Please sign in to comment.