1 * Sat Sep 3 2005 Brendan O'Dea <bod@optus.net> 2.1.5
2 - Avoid Code-Reject loop.
3 - Increase size of PPP buffers to MAXETHER.
4 - Bug fixes for CLI ringbuffer and tunnel HELLO from Yuri.
5 - Restart rather than halt BGP on receipt of CEASE (Dominique Rousseau).
6 - Add cluster_mcast_ttl option to allow a cluster to span multiple
7 subnets (suggested by Tim Devries).
9 * Mon Aug 29 2005 Brendan O'Dea <bod@optus.net> 2.1.4
10 - Drop level of "Unexpected CHAP message" log.
11 - Fix parsing of ProtocolRej (allow 1 or two byte protocols).
12 - Handle rejection of MRU negotiation by peer.
13 - Use local hostname for tunnel in SCCRP (Alex Kiernan).
15 * Wed Aug 17 2005 Brendan O'Dea <bod@optus.net> 2.1.3
16 - Fail IPCP negotiation only on ConfigRej of IP-Address.
18 * Wed Aug 10 2005 Brendan O'Dea <bod@optus.net> 2.1.2
19 - Clear cluster_master on election so that slaves will accept a new master.
20 - Provide more comments/defaults in etc/startup-config.default.
21 - Add DAE support (PoD/CoA) from Vladislav Bjelic.
22 - Clean up new warnings from gcc 4.0.
23 - Replace flags used for LCP/IPCP with state machine.
24 - Include Acct-Session-Time in interim records.
25 - Fix DAE vector, generateload (Alex Kiernan).
26 - Replace RSA MD5 with public domain version.
28 * Tue Jun 14 2005 Brendan O'Dea <bod@optusnet.com.au> 2.1.1
29 - Add missing newline to backtrace macro.
30 - Don't send CDN for each session when shutting down tunnels (this is
32 - Move tunnel shutdown from SIGQUIT signal handler to be run once from
33 still_busy(). Reject new tunnels/sessions while in the process of
35 - Clarify usage of shutdown signals in documentation.
36 - Always initialise PRNG.
37 - Sanity check length of random_vector.
38 - Fix segv in unhide_value.
39 - Ping new master when we get C_MASTER and delay next election to allow
40 the unicast limp-along code to kick in if required.
42 * Sun Jun 5 2005 Brendan O'Dea <bod@optusnet.com.au> 2.1.0
43 - Add IPv6 support from Jonathan McDowell.
44 - Add CHAP support from Jordan Hrycaj.
45 - Add interim accounting support from Vladislav Bjelic.
46 - Negotiate MRU, default 1458 to avoid fragmentation.
47 - Sanity check that cluster_send_session is not called from a child
49 - Use bounds-checking lookup functions for string constants.
50 - Add enum for RADIUS codes.
51 - Make "call_" prefix implict in CSTAT() macro.
52 - Fix some format string problems.
53 - Remove "save_state" option. Not maintained anymore; use clustering
54 to retain state across restarts.
55 - Simplify AVP unhiding code.
56 - Add optional "username" parameter to ungarden control, allowing the
57 username to be reset before going online.
58 - Add result/error codes and message to StopCCN when shutting down tunnels.
59 - Add result/error codes to CDN when shutting down sessions. Sends 2/7
60 (general error, try another LNS) when out of IP addresses, and 3
61 (adminstrative) for everything else (suggestion from Chris Gates).
62 - Use cli_error() for error messages and help.
63 - Don't use LOG() macro in initdata() until the config struct has been
64 allocated (uses config->debug).
65 - Initialise log_stream to stderr to catch errors before the config file
67 - Make "show running-config" a privileged command (contains clear text
69 - Add sessionctl plugin to provide drop/kill via nsctl.
70 - New config option: allow_duplicate_users which determines whether
71 or not to kill older sessions with the same username.
72 - Fix byte counters in accounting records.
73 - Add Acct-Output-Gigawords, Acct-Input-Gigawords attributes to RADIUS
75 - Fix icmp host unreachable to use router address.
76 - Include endpoint address in accounting dump files.
77 - Convert mainloop to use epoll rather than select.
78 - Add note about fragmentation in Docs/manual.html, and a sample
79 iptables rule for MSS clamping.
81 + Show session open time in "show session"/"show user" detailed output.
82 + Have slaves with BGP configured drop BGP on receipt of a shutdown
83 signal, but hang about for an additional 5s to process any remaining
85 + Run regular_cleanups after processing the results of the select,
86 looking at a sufficient slice of each table to ensure that all
87 entries are examined at least once per second.
89 + Cluster changes from Michael, intended to prevent a stray master
90 from trashing a cluster:
91 = Ignore heartbeats from peers claiming to be the master before the
92 timeout on the old master has expired.
93 = A master receiving a stray heartbeat sends a unicast HB back, which
94 should cause the rogue to die due to the tie-breaker code.
95 = Keep probing the master for late heartbeats.
96 = Drop BGP as soon as we become master with the minumum required peers.
97 = Any PING seen from a master forces an election (rather than just
98 where basetime is zero).
99 = A slave which receives a LASTSEEN message (presumably a restarted
100 master) sends back new message type, C_MASTER which indicates the
101 address of the current master.
102 + New config option: cluster_master_min_adv which determines the minimum
103 number of up to date slaves required before the master will drop
106 + Add handling of "throttle=N" RADIUS attributes.
107 + Fix RADIUS indexing (should have 16K entries with 64 sockets).
109 + Fix leak in session freelist when initial RADIUS session allocation
112 + Add a Cisco-Avpair with intercept details to RADIUS Start/Stop
115 + Only send RADIUS stop record in sessionshutdown when there's an ip address.
116 + Reset .die on master takeover (so that dying sessions don't have to
117 hang around until the new master has the same uptime as the old one).
118 + Update .last_packet in cluster_handle_bytes only when there have
119 been bytes received from the modem (dead sessions were having the
120 idle timeout reset by stray packets).
122 + Ensure that sessionkill is not called on an unopened session (borks
124 + Bump MAXSESSION to 60K.
125 + Fix off-by-one errors in session/tunnel initialisation and
126 sessiont <-> sessionidt functions.
127 + Use session[s].opened consistently when checking for in-use sessions
128 (rather than session[s].tunnel).
129 + Use <= cluster_highest_sessionid rather than < MAXSESSION in a
131 + Don't kill a whole tunnel if we're out of sessions.
132 + Change session[s].ip to 0 if set from RADIUS to 255.255.255.254;
133 avoids the possibility that it will be interpreted as a valid IP
135 + Avoid a possible buffer overflow in processpap.
136 + Kill session if authentication was rejected.
138 + More DoS prevention: add packet_limit option to apply a hard limit
139 to downstream packets per session.
140 + Fix "clear counters".
141 + Log "Accepted connection to CLI" at 4 when connection is from localhost
142 to reduce noise in logs.
143 + Show time since last counter reset in "show counters".
145 + Throttle outgoing LASTSEEN packets to at most one per second for a
148 * Fri Dec 17 2004 Brendan O'Dea <bod@optusnet.com.au> 2.0.13
149 - Better cluster master collision resolution: keep a counter of state
150 changes, propagated in the heartbeats; the master with the highest #
151 of changes (that has kept in contact with the LAC through the
153 - Skip newlines in ringbuffer messages to CLI.
154 - Drop "Session N is closing" message level to 4; don't process PPPIP
155 packets in this state.
156 - Use gzip --best for man pages, include pid_file in sample
157 startup-config (from Jonathan's Debian package patches).
158 - Read multiple packets off cluster_sockfd as well as udpfd, tunfd in an
159 attempt to avoid losing the cluster in high load (DoS) conditions.
160 - Add counters for select_called, multi_read_used and multi_read_exceeded.
162 - Retain counters of shutdown sessions to dump once per minute.
163 - Use standard uintN_t types for portability.
165 * Wed Dec 1 2004 Brendan O'Dea <bod@optusnet.com.au> 2.0.12
166 - The "This time, for sure!" release.
167 - Fix throttlectl plugin argument parsing.
169 * Wed Dec 1 2004 Brendan O'Dea <bod@optusnet.com.au> 2.0.11
170 - Don't send a RADIUS start record when ungardening on shutdown.
172 * Wed Dec 1 2004 Brendan O'Dea <bod@optusnet.com.au> 2.0.10
173 - Fix byte ordering of LCP header length (thanks Yuri).
174 - Increase ip_conntrack_max due to dropped packets.
176 * Tue Nov 30 2004 Brendan O'Dea <bod@optusnet.com.au> 2.0.9
177 - Revise CCP, send ConfigReq once only.
178 - Don't copy the old buffer into Config{Nak,Rej} LCP responses (oops);
179 add length checks when appending.
180 - Do copy the identifier from the request and update length.
181 - Have makeppp print a backtrace on overflow.
182 - Check control serial before clearing window, prevents looping tunnel
183 setup in some instances.
184 - Implement named access-lists which may be applied to a session
185 either via Filter-Id RADIUS responses, or using the CLI.
186 - Drop ip address from LOG.
187 - autothrottle: revise parsing; ignore lcp:interface-config avpairs
188 which don't start with serv[ice-policy].
191 * Sat Nov 20 2004 Brendan O'Dea <bod@optusnet.com.au> 2.0.8
192 - Ignore gateway address in Framed-Route (from Jonathan McDowell).
193 - Don't route Framed-IP-Address if contained in a Framed-Route.
194 - Call sessionshutdown() when a tunnel is dropped rather than
195 sessionkill() to ensure that RADIUS stop records are sent.
196 - Cleanup: make a bunch of global functions/variables static.
197 - Remove reference to old -a command line argument.
198 - Add l2tpns(8) and nsctl(8) manpages from Jonathan McDowell.
199 - Add startup-config(5) manpage.
200 - Revise nsctl to allow arbitrary strings/args to be passed to plugins.
201 - Add snoopctl, throttlectl plugins.
202 - Fix deletion from linked list.
203 - Allow LCP re-negotiation after connection completes (thanks Yuri).
205 * Mon Nov 15 2004 Brendan O'Dea <bod@optusnet.com.au> 2.0.7
206 - Fix socket creation in host_unreachable() (thanks to Bjørn Augestad)
207 - Don't assume BGP peer sends back negotiated hold time, pick smallest
209 * Thu Nov 11 2004 Brendan O'Dea <bod@optusnet.com.au> 2.0.6
210 - Make BGP keepalive/hold time configurable
211 - Revise BGP config to use "router bgp AS" syntax (requires libcli >= 1.8.2)
213 * Tue Nov 9 2004 Brendan O'Dea <bod@optusnet.com.au> 2.0.5
214 - Handle routing properly in lone-master case
215 - Fix intercepts: don't double-snoop throttled customers, ensure
216 byte/packet counts are only updated once
217 - Add a callback to allow plugins to fetch values from the running config
219 * Mon Nov 8 2004 Brendan O'Dea <bod@optusnet.com.au> 2.0.4
220 - Added setrxspeed plugin
221 - Added peer_address config option
222 - Rename wrapper macros to LOG()/LOG_HEX(), use p->log() in plugins
223 - Replace some PPP{PAP,CHAP} magic numebrs with constants
224 - Nak asyncmap (unless == 0)
225 - Bundle ConfigRej options
226 - Clean up initlcp handling
228 * Wed Nov 3 2004 Brendan O'Dea <bod@optusnet.com.au> 2.0.3
229 - Added support for hidden AVPs by Robert Clark
230 - l2tpns-chap-response.patch from Robert Clark
231 - Merge l2tpns-config-hostname.patch from Robert Clark
232 - l2tpns-dont-timeshift-unidirectional-traffic.patch from Robert Clark - Dump accounting data if cin OR cout is non-zero
233 - Don't write accounting files if no accounting dir is set - Yuri
234 - Fix checking for mmap success
235 - Renegotiate MRU - Yuri
236 - Take LCP ConfigReq length from the packet length field - Yuri
237 - Hostname set via command line not config
238 - Make number of throttle buckets configurable
239 - Shared_malloc returns NULL on failure
241 - Unfsck 4->8 indenting change
242 - Use 2 seperate u16 values for throttle rate in/out
243 - Defer adding radius fds to the select loop until become_master
245 * Thu Sep 02 2004 David Parrish <david@dparrish.com> 2.0.2
246 - Combined LCP patches from Iain and Yuri. This should allow Windows 2k/XP
247 clients to connect, as well Linksys DSL modems.
248 - Apply patch to fix -v option from Juergen Kammer.
249 - Makefile fix from Juergen Kammer to not overwrite existing config files on
251 - Configurable radius port patch from Juergen Kammer.
252 - Send my_address if no bind_address when doing IPCP
253 - Write pid file if filename is set
254 - Add startup script and monitor script from Yuri
255 - Some logging correctness fixes from Iain Wade
256 - Add support for LCP Ident and CallBack (rejection only) from Yuri
257 - Initiate LCP if not attempted by the client, or in renegotiation - Yuri
258 - Indentation and style cleanups
259 - Per-user upload and download throttle rates - Yuri
260 - Make autothrottle.so understand cisco lcp:interface-config - Yuri
261 - Show filter stats in show session - Yuri
262 - Cleanup from Michael to change sid to unique_id
263 - Add plugin to remove domain name from auth requests
264 - Add .spec file for RPM generation
266 * Tue Jul 13 2004 Brendan O'Dea <bod@optusnet.com.au> 2.0.1
267 - Update INSTALL, Docs/manual.html documentation.
268 - Add INTERNALS documentation.
269 - Add lock_pages option.
270 - TerminateAck fix from Yuri
271 - Adject cli_loop args for libcli 1.8.0
272 - Allow for backward compatabity in C_PING packets
273 - Don't send RADIUS stop messages from sessionshutdown when called from
276 - Fix for LASTSEEN breakage: don't do anything in the CLI other than
277 flag changes to be made by the parent.
278 - Split out master parts from cluster_check_master() into cluster_check_slaves().
279 - Set hostname in CLI prompt.
280 - Make cluster_hb_interval work; include interval/timeout in heartbeats
281 so that a change on the master is propagated immediately to the slaves.
282 - Use fast heartbeats when there are slaves not up to date.
283 - Ensure basetime of shut down master is set to zero (prevent delayed election).
284 - Fix radius session leak on IPCP timeout.
285 - Fix some off-by-one errors in tunnel/session loops.
286 - Add "limp along" fix for when a slave drops temporarily from the mcast group.
287 - Rename l2tpns.cfg as startup-config to match CONFIGFILE.
288 - Update cli callbacks to work with libcli 1.6.
289 This supports privileged and unprivileged commands, as well as a configuration
291 - Add help for all cli commands.
292 - Add "show version" command.
293 - Fix uptime counter display.
294 - Fix nasty bug where cluster basetime can be set to 0 when sending initial
296 - Don't rmmod ip_conntrack, as this can take a lot of time.
297 - Re-order logging in routeset such that the action is given before any error.
298 - Use the correct gateway address when deleting routes.
299 - Remove any routes when address changes.
300 - Require authentication if telnet from remote ip.
301 - Require enable password always.
302 - Return error if show pool done on slave.
303 - We MUST immediately exit if we're the wrong master!
305 * Wed Jun 23 2004 David Parrish <david@dparrish.com> 2.0.0
307 - Completely replace active/standby clustering with a new peer-to-peer
308 clustering method which allows much greater throughput and is a lot more fault
310 - Add internal tbf implementation for throttling without relying on tc and
312 - Add support for iBGP and eBGP to advertise routes
313 - Add cli commands "show cluster", "show bgp", "show ipcache", "show throttle",
314 "show tbf", "suspend bgp", "restart bgp", "show user"
315 - Interception destination must be set per-user
316 - If SMP machine, allow use of SCHED_FIFO, which should improve performance
317 - Added config option to send GARP at startup
318 - Added plugin_become_master and plugin_new_session_master plugin hooks
319 - Remove useless sessionsendarp(). This isn't needed now that we are using TUN
321 - ICMP rate limiting so not every unreachable packet is replied with an ICMP
323 - mangle table is not required on anything but the cluster master, so slaves
324 will drop the mangle table and attempt to unload the ip_conntrack module
325 - Statically assigned IP addresses (by Radius) work now
326 - Add -d command-line flag to detach and become a daemon
327 - Configuration file is now "/etc/l2tpns/startup-config"
328 - Reduced MIN_IP_SIZE to 0x19 to stop a pile of Short IP warnings
329 - Resend initial IPCP request until it's acknowleged by the client
330 - Better radius session cleanup logic
331 - Many miscellaenous bugfixes and performance enhancements
332 - Thanks to Michael O'Reilly and Brendan O'Dea for most of these new features
334 * Mon May 24 2004 David Parrish <david@dparrish.com> 1.2.0
335 - Fix SEGFAULT in garden module
336 - Use multiple radius sockets to allow more concurrent authentication requests
337 - Add username parameter to "show users" command
338 - Fix counting tunnel rx errors as tunnel tx errors
339 - Add "show throttle" command
340 - Add gcc __attribute__ to logging functions
341 - Fix warnings shown by __attribute__
342 - Make sure regular cleanup happens regularly under high load
343 - Add variable cleanup_interval for changing cleanup interval
344 - Add support for reading more than one packet per fd in each processing loop
345 - This is configurable with the multi_read_count variable
346 - Remove segv handler so core dumps can happen
347 - Use nonblocking sockets
348 - Increase tun queue length
349 - Fix minimum length of IP packets
350 - Remove per-packet plugin hooks (they are slow)
351 - Don't drop session if no free RADIUS
352 - Don't expire more than 1000 sessions per cleanup interval
353 - Remove -a and -c command-line options. They don't work anyway
354 - Don't require file: in log_filename
355 - Bump version to 1.2.0
356 - Check return code when throttling users
358 * Mon Apr 5 2004 David Parrish <david@dparrish.com> 1.1.1
359 - Don't mention configure anymore, it's not used
360 - Added the autosnoop and autothrottle modules
361 - Don't default to using a htb for the class root
363 * Fri Mar 5 2004 David Parrish <david@dparrish.com> 1.1.0
364 - Change all strcpy() calls to strncpy() to avoid buffer overflow potential
365 - Add ICMP host unreachable support
366 - Logging to syslog if log_file = "syslog:facility"
367 - Now requires libcli 1.5
368 - All configuration moves to a config structure
369 - Ability to modify and write config on the fly through command-line interface
370 - Config file support is removed, and now handled by the cli
371 - Show hostname in cli prompt
372 - Keep current state type for tunnels
373 - Add uptime command do CLI, which also shows real-time bandwidth utilisation
374 - Add goodbye command to cluster master, which forces droppping a slave
375 - Cache IP address allocation, so that reconnecting users get the same address
376 - Fix tunnel resend timeouts, so that dead tunnels will be cleaned up
377 - Allocate tunnels and radius without using a linked list which had issues
378 - Fix some off-by-one errors in tunnel and session and radius arrays
379 - Save and reload ip address pool when dieing
380 - Check version and size of reloaded data when restarting
381 - Remove plugin_config support
382 - Remove old support for TBF which didn't work anyway. HTB is required to do throttling now.
383 - Add COPYING and Changes files