3 * <grzes_at_gnu.univ.gda.pl> added arpreply and among modules
4 * <tommy_at_home.tig-grr.com> added limit match
6 * added (automatic) Sparc64 support, thanks to Michael Bellion and
7 Thomas Heinz from hipac.org for providing a test-box.
9 * added stp frames match type
11 * added support for deleting all user-defined chains (-X option
12 without specified chain)
15 * <csv_at_bluetail.com> Chris Vitale: basic 802.3/802.2 filtering
16 (experimental, kernel files are in the CVS)
19 * added negative rule counter support
20 * bugfix: bcnt was not updated correctly
21 * <blancher_at_cartel-securite.fr> Cedric Blancher: add ARP MAC
25 * fixed check bug in ebt_ip.c (report from
26 joe_judge_at_guardium.com).
28 * fixed problem when removing a chain (report from
29 ykphuah_at_greenpacket.com).
30 * Added --help list_extensions which, well, lists the extensions
32 * changed the way to use the atomic operations. It's now possible
33 to use the EBTABLES_ATOMIC_FILE environment variable, so it's no
34 longer necessary to explicitly state the file name. See the man.
36 * changed the way of compiling. New releases will now contain their
37 own set of kernel includes. No more copying of kernel includes to
39 * added getethertype.c (Nick) and use it. Removed name_to_number()
42 * added possibility to specify a rule number interval when deleting
45 * added ! - option possibility, which is equivalent to - ! option
47 * since last entry: added byte counters and udp/tcp port matching
49 * updated the kernel files for 2.4.20-pre5 and 2.5.32
50 * last big cleanup of kernel and userspace code just finished
56 * other things done before 2.0-rc1 that I can think of,
58 * cache align counters for better smp performance
60 * check for --xxxx-target RETURN on base chain
65 * bugfix for --atomic-commit
67 * added mark target+match
69 * added --atomic options
71 * some unlogged changes (due to lazyness)
72 * added --Lc, --Ln, --Lx
74 * user defined chains support: added -N, -X, -E options.
76 * some unlogged changes (due to lazyness)
77 * change the output for -L to make it look like it would look when
78 the user inputs the command.
79 * try to autoload modules
81 * add user defined chains support (without new commands yet,
83 * comparing rules didn't take the logical devices into account
85 * update help for -s and -d
86 * add VLAN in ethertypes
87 * add SYMLINK option for compiling
89 * allow -i and --logical-in in BROUTING
90 * update the manual page
91 * rename /etc/etherproto into /etc/ethertypes (seems to be a more
93 * add MAC mask for -s and -d, also added Unicast, Multicast and
94 Broadcast specification for specifying a (family of) MAC
98 * added redirect target.
99 * added --redirect-target, --snat-target and --dnat-target options.
100 * added logical_out and logical_in
101 * snat bugfix (->size)
103 * fixed some things in the manual.
106 * -j standard no longer works, is this cryptic? good :)
107 * lots of beautification.
108 - made some code smaller
109 - made everything fit within 80 columns
110 * fix problems with -i and -o option
111 * print_memory now prints useful info
112 * trying to see the tables when ebtables is not loaded in kernel
113 no longer makes this be seen as a bug.
115 ebtables v2.0 released, changes:
116 * A complete rewrite, made everything modular.
117 * Fixed a one year old bug in br_db.c. A similar bug was present
118 in ebtables.c. It was visible when the number of rules got
120 * Removed the option to allow/disallow counters. Frames passing
121 by are always counted now.
122 * Didn't really add any new functionality. However, it will be
123 _alot_ easier and prettier to do so now. Feel free to add an
125 * There are 4 types of extensions:
127 - Matches: like iptables has.
128 - Watchers: these only watch frames that passed all the matches
129 of the rule. They don't change the frame, nor give a verdict.
130 The log extension is a watcher.
132 * user32/kernel64 architectures like the Sparc64 are unsupported.
133 If you want me to change this, give me access to such a box,
134 and don't pressure me.