chiark / gitweb /
cleanup: Various simple whitespace changes.
[tripe] / doc / tripe-admin.5.in
... / ...
CommitLineData
1.\" -*-nroff-*-
2.\"
3.ie t \{\
4. if \n(.g \{\
5. fam P
6. \}
7.\}
8.
9.de SP
10.TP
11..
12.TH tripe-admin 5 "18 February 2001" "Straylight/Edgeware" "TrIPE: Trivial IP Encryption"
13.SH NAME
14tripe-admin \- administrator commands for TrIPE
15.SH DESCRIPTION
16This manual page describes the administration interface provided by the
17.BR tripe (8)
18daemon.
19.PP
20The
21.BR tripectl (8)
22program can be used either interactively or in scripts to communicate
23with the server using this interface. Alternatively, simple custom
24clients can be written in scripting languages such as Perl, Python or
25Tcl, or more advanced clients such as GUI monitors can be written in C
26with little difficulty.
27.PP
28By default, the server listens for admin connections on the Unix-domain
29socket
30.BR /var/lib/tripe/tripesock .
31Administration commands use a textual protocol. Each client command or
32server response consists of a line of ASCII text terminated by a single
33linefeed character. No command may be longer than 255 characters.
34.SS "General structure"
35Each command or response line consists of a sequence of
36whitespace-separated words. The number and nature of whitespace
37characters separating two words in a client command is not significant;
38the server always uses a single space character. The first word in a
39line is a
40.I keyword
41identifying the type of command or response contained. Keywords in
42client commands are not case-sensitive; the server always uses uppercase
43for its keywords.
44.SS "Simple commands"
45For simple client command, the server responds with zero or more
46.B INFO
47lines, followed by either an
48.B OK
49line or a
50.B FAIL
51line. Each
52.B INFO
53provides information requested in the command. An
54.B OK
55response contains no further data. A
56.B FAIL
57code is followed by a machine-readable explanation of why the command
58failed.
59.PP
60Simple command processing is strictly synchronous: the server reads a
61command, processes it, and responds, before reading the next command.
62All commands can be run as simple commands. Long-running commands
63(e.g.,
64.B ADD
65and
66.BR PING )
67block the client until they finish, but the rest of the server continues
68running.
69.SS "Asynchronous messages"
70There are three types of asynchronous messages which
71aren't associated with any particular command.
72.PP
73The
74.B WARN
75message contains a machine-readable message warning of an error
76encountered while processing a command, unexpected or unusual behaviour
77by a peer, or a possible attack by an adversary. Under normal
78conditions, the server shouldn't emit any warnings.
79.PP
80The
81.B TRACE
82message contains a human-readable tracing message containing diagnostic
83information. Trace messages are controlled using the
84.B \-T
85command-line option to the server, or the
86.B TRACE
87administration command (see below). Support for tracing can be disabled
88when the package is being configured, and may not be available in your
89version.
90.PP
91Finally, the
92.B NOTE
93message is a machine-readable notification about some routine but
94interesting event such as creation or destruction of peers.
95.PP
96The presence of asynchronous messages can be controlled using the
97.B WATCH
98command.
99.SS "Background commands"
100Some commands (e.g.,
101.B ADD
102and
103.BR PING )
104take a long time to complete. To prevent these long-running commands
105from tying up a server connection, they can be run in the background.
106Not all commands can be run like this: the ones that can provide a
107.B \-background
108option, which must be supplied with a
109.IR tag .
110.PP
111A command may fail before it starts running in the background. In this
112case, the server emits a
113.B FAIL
114response, as usual. To indicate that a command has started running in
115the background, the server emits a response of the form
116.BI "BGDETACH " tag \fR,
117where
118.I tag
119is the value passed to the
120.B \-background
121option. From this point on, the server is ready to process more
122commands and reply to them.
123.PP
124Responses to background commands are indicated by a line beginning with
125one of the tokens
126.BR BGOK ,
127.BR BGFAIL ,
128or
129.BR BGINFO ,
130followed by the command tag. These correspond to the
131.BR OK ,
132.BR FAIL ,
133and
134.B INFO
135responses for simple commands:
136.B BGINFO
137indicates information from a background command which has not completed
138yet; and
139.B BGOK
140and
141.B BGFAIL
142indicates that a background command succeeded or failed, respectively.
143.PP
144A background command will never issue an
145.B OK
146or
147.B BGINFO
148response: it will always detach and then issue any
149.B BGINFO
150lines followed by
151.B BGOK
152response.
153.SS "Network addresses"
154A network address is a sequence of words. The first is a token
155identifying the network address family. The length of an address and
156the meanings of the subsequent words depend on the address family.
157Address family tokens are not case-sensitive on input; on output, they
158are always in upper-case.
159.PP
160At present, only one address family is understood.
161.TP
162.BI "INET " address " " port
163An Internet socket, naming an IPv4 address and UDP port. On output, the
164address is always in numeric dotted-quad form, and the port is given as
165a plain number. On input, DNS hostnames and symbolic port names are
166permitted. Name resolution does not block the main server, but will
167block the requesting client. This hopefully makes life simpler for
168stupid clients. Complex clients which don't wish to be held up can open
169extra connections or do the resolution themselves.)
170.PP
171If, on input, no recognised address family token is found, the following
172words are assumed to represent an
173.B INET
174address.
175.SS "Key-value output"
176Some commands (e.g.,
177.B STATS
178and
179.BR SERVINFO )
180produce output in the form of
181.IB key = value
182pairs, one per word. Neither the
183.I key
184nor the
185.I value
186contain spaces.
187.SS "Trace lists"
188Commands which enable or disable kinds of output (e.g.,
189.B TRACE
190and
191.BR WATCH )
192work in similar ways. They take a single optional argument, which
193consists of a string of letters selecting message types, optionally
194interspersed with
195.RB ` + '
196to enable, or
197.RB ` \- '
198to disable, the subsequently listed types.
199.PP
200If the argument is omitted, the available message types are displayed,
201one to an
202.B INFO
203line, in a fixed-column format. Column zero contains the key letter for
204selecting that message type; column one contains either a space or a
205.RB ` + '
206sign, if the message type is disabled or enabled respectively; and a
207textual description of the message type begins at column 3 and continues
208to the end of the line.
209.PP
210Lowercase key letters control individual message types. Uppercase key
211letters control collections of message types.
212.SH "COMMAND REFERENCE"
213.\"* 10 Commands
214The commands provided are:
215.SP
216.BI "ADD " peer " \fR[" options "\fR] " address "\fR..."
217Adds a new peer. The peer is given the name
218.IR peer ;
219the peer's public key is assumed to be in the file
220.B keyring.pub
221(or whatever alternative file was specified in the
222.B \-K
223option on the command line). The
224.I address
225is the network address (see above for the format) at which the peer can
226be contacted. The following options are recognised.
227.RS
228.\"+opts
229.TP
230.BI "\-background " tag
231Run the command in the background, using the given
232.IR tag .
233.TP
234.BI "\-keepalive " time
235Send a no-op packet if we've not sent a packet to the peer in the last
236.I time
237interval. This is useful for persuading port-translating firewalls to
238believe that the `connection' is still active. The
239.I time
240is expressed as a nonnegative integer followed optionally by
241.BR d ,
242.BR h ,
243.BR m ,
244or
245.BR s
246for days, hours, minutes, or seconds respectively; if no suffix is
247given, seconds are assumed.
248.TP
249.BI "\-tunnel " tunnel
250Use the named tunnel driver, rather than the default.
251.\"-opts
252.RE
253.SP
254.BI "ADDR " peer
255Emits an
256.B INFO
257line reporting the IP address and port number stored for
258.IR peer .
259.SP
260.BI "CHECKCHAL " challenge
261Verifies a challenge as being one earlier issued by
262.B GETCHAL
263and not previously either passed to
264.B CHECKCHAL
265or in a greeting message.
266.SP
267.B "DAEMON"
268Causes the server to disassociate itself from its terminal and become a
269background task. This only works once. A warning is issued.
270.TP
271.BI "EPING \fR[" options "\fR] " peer
272Sends an encrypted ping to the peer, and expects an encrypted response.
273This checks that the peer is running (and not being impersonated), and
274that it can encrypt and decrypt packets correctly. Options and
275responses are the same as for the
276.B PING
277command.
278.SP
279.BI "FORCEKX " peer
280Requests the server to begin a new key exchange with
281.I peer
282immediately.
283.SP
284.B "GETCHAL"
285Requests a challenge. The challenge is returned in an
286.B INFO
287line, as a base64-encoded string. See
288.BR CHECKCHAL .
289.SP
290.BI "GREET " peer " " challenge
291Sends a greeting packet containing the
292.I challenge
293(base-64 encoded) to the named
294.IR peer .
295The expectation is that this will cause the peer to recognize us and
296begin a key-exchange.
297.SP
298.B "HELP"
299Causes the server to emit an
300.B INFO
301line for each command it supports. Each line lists the command name,
302followed by the names of the arguments. This may be helpful as a memory
303aid for interactive use, or for program clients probing for features.
304.SP
305.BI "IFNAME " peer
306Emits an
307.B INFO
308line containing the name of the network interface used to collect IP
309packets which are to be encrypted and sent to
310.IR peer .
311Used by configuration scripts so that they can set up routing tables
312appropriately after adding new peers.
313.SP
314.BI "KILL " peer
315Causes the server to forget all about
316.IR peer .
317All keys are destroyed, and no more packets are sent. No notification
318is sent to the peer: if it's important that the peer be notified, you
319must think of a way to do that yourself.
320.SP
321.B "LIST"
322For each currently-known peer, an
323.B INFO
324line is written containing the peer's name, as given to
325.BR ADD .
326.SP
327.BI "NOTIFY " tokens\fR...
328Issues a
329.B USER
330notification to all interested administration clients.
331.SP
332.BI "PEERINFO " peer
333Returns information about a peer, in key-value form. The following keys
334are returned.
335.RS
336.TP
337.B tunnel
338The tunnel driver used for this peer.
339.TP
340.B keepalive
341The keepalive interval, in seconds, or zero if no keepalives are to be
342sent.
343.RE
344.SP
345.BI "PING \fR[" options "\fR] " peer
346Send a transport-level ping to the peer. The ping and its response are
347not encrypted or authenticated. This command, possibly in conjunction
348with tracing, is useful for ensuring that UDP packets are actually
349flowing in both directions. See also the
350.B EPING
351command.
352.IP
353An
354.B INFO
355line is printed describing the outcome:
356.RS
357.TP
358.BI "ping-ok " millis
359A response was received
360.I millis
361after the ping was sent.
362.TP
363.BI "ping-timeout"
364No response was received within the time allowed.
365.TP
366.BI "ping-peer-died"
367The peer was killed (probably by another admin connection) before a
368response was received.
369.RE
370.IP
371Options recognized for this command are:
372.RS
373.\"+opts
374.TP
375.BI "\-background " tag
376Run the command in the background, using the given
377.IR tag .
378.TP
379.BI "\-timeout " time
380Wait for
381.I time
382seconds before giving up on a response. The default is 5 seconds. (The
383time format is the same as for the
384.B "ADD \-keepalive"
385option.)
386.\"-opts
387.RE
388.SP
389.B "PORT"
390Emits an
391.B INFO
392line containing just the number of the UDP port used by the
393.B tripe
394server. If you've allowed your server to allocate a port dynamically,
395this is how to find out which one it chose.
396.SP
397.B "RELOAD"
398Instructs the server to recheck its keyring files. The server checks
399these periodically anyway but it may be necessary to force a recheck,
400for example after adding a new peer key.
401.SP
402.B "QUIT"
403Instructs the server to exit immediately. A warning is sent.
404.SP
405.B "SERVINFO"
406Returns information about the server, in the form of key-value pairs.
407The following keys are used.
408.RS
409.TP
410.B implementation
411A keyword naming the implementation of the
412.BR tripe (8)
413server. The current implementation is called
414.BR edgeware-tripe .
415.TP
416.B version
417The server's version number, as reported by
418.BR VERSION .
419.TP
420.B daemon
421Either
422.B t
423or
424.BR nil ,
425if the server has or hasn't (respectively) become a daemon.
426.RE
427.SP
428.BI "SETIFNAME " peer " " new-name
429Informs the server that the
430.IR peer 's
431tunnel-interface name has been changed to
432.IR new-name .
433This is useful if firewalling decisions are made based on interface
434names: a setup script for a particular peer can change the name, and
435then update the server's records so that they're accurate.
436.SP
437.BI "STATS " peer
438Emits a number of
439.B INFO
440lines, each containing one or more statistics in the form
441.IB name = value \fR.
442The statistics-gathering is experimental and subject to change.
443.SP
444.BR "TRACE " [\fIoptions\fP]
445Selects trace outputs: see
446.B "Trace lists"
447above. Message types provided are:
448.RS
449.PP
450Currently, the following tracing options are supported:
451.TP
452.B t
453Tunnel events: reception of packets to be encrypted, and injection of
454successfully-decrypted packets.
455.TP
456.B r
457Peer management events: creation and destruction of peer attachments,
458and arrival of messages.
459.TP
460.B a
461Administration interface: acceptance of new connections, and handling of
462the backgroud name-resolution required by the
463.B ADD
464command.
465.TP
466.B s
467Handling of symmetric keysets: creation and expiry of keysets, and
468encryption and decryption of messages.
469.TP
470.B x
471Key exchange: reception, parsing and emission of key exchange messages.
472.TP
473.B m
474Key management: loading keys and checking for file modifications.
475.TP
476.B l
477Display information about challenge issuing and verification.
478.TP
479.B p
480Display contents of packets sent and received by the tunnel and/or peer
481modules.
482.TP
483.B c
484Display inputs, outputs and intermediate results of cryptographic
485operations. This includes plaintext and key material. Use with
486caution.
487.TP
488.B A
489All of the above.
490.PP
491Note that the
492.B p
493(packet contents)
494and
495.B c
496(crypto details)
497outputs provide extra detail for other outputs. Specifying
498.B p
499without
500.BR r
501or
502.B t
503isn't useful; neither is specifying
504.B c
505without one of
506.BR s ,
507.BR l ,
508.B x
509or
510.BR m .
511.RE
512.SP
513.B "TUNNELS"
514For each available tunnel driver, an
515.B INFO
516line is printed giving its name.
517.SP
518.B "VERSION"
519Causes the server to emit an
520.B INFO
521line stating its software version, as two words: the server name, and
522its version string. The server name
523.B tripe
524is reserved to the Straylight/Edgeware implementation.
525.SP
526.BR "WATCH " [\fIoptions\fP]
527Enables or disables asynchronous messages
528.IR "for the current connection only" .
529See
530.B "Trace lists"
531above. The default watch state for the connection the server opens
532automatically on stdin/stdout is to show warnings and trace messages;
533other connections show no asynchronous messages. (This is done in order
534to guarantee that a program reading the server's stdout does not miss
535any warnings.)
536.RS
537.PP
538Message types provided are:
539.TP
540.B t
541.B TRACE
542messages.
543.TP
544.B n
545.B NOTE
546messages.
547.TP
548.B w
549.B WARN
550messages.
551.TP
552.B A
553All of the above.
554.RE
555.SP
556.BI "WARN " tokens\fR...
557Issues a
558.B USER
559warning to all interested administration clients.
560.SH "ERROR MESSAGES"
561.\"* 20 Error messages (FAIL codes)
562The following
563.B FAIL
564(or
565.BR BGFAIL )
566messages are sent to clients as a result of errors during command
567processing.
568.SP
569.BI "already-daemon"
570(For
571.BR DAEMON .)
572The
573.B tripe
574server is already running as a daemon.
575.SP
576.BI "bad-addr-syntax " message
577(For commands accepting socket addresses.) The address couldn't be
578understood.
579.SP
580.BI "bad-syntax " cmd " " message
581(For any command.) The command couldn't be understood: e.g., the number
582of arguments was wrong.
583.SP
584.BI "bad-time-spec " word
585The
586.I word
587is not a valid time interval specification. Acceptable time
588specifications are nonnegative integers followed optionally by
589.BR d ,
590.BR h ,
591.BR m ,
592or
593.BR s ,
594for days, hours, minutes, or seconds, respectively.
595.SP
596.BI "bad-trace-option " char
597(For
598.BR TRACE .)
599An unknown trace option was requested.
600.SP
601.BI "bad-watch-option " char
602(For
603.BR WATCH .)
604An unknown watch option was requested.
605.SP
606.BI "daemon-error " ecode " " message
607(For
608.BR DAEMON .)
609An error occurred during the attempt to become a daemon, as reported by
610.IR message .
611.SP
612.BI "invalid-port " number
613(For
614.BR ADD .)
615The given port number is out of range.
616.SP
617.BI "peer-create-fail " peer
618(For
619.BR ADD .)
620Adding
621.I peer
622failed for some reason. A warning should have been emitted explaining
623why.
624.SP
625.BI "peer-exists " peer
626(For
627.BR ADD .)
628There is already a peer named
629.IR peer .
630.SP
631.B "ping-send-failed"
632The attempt to send a ping packet failed, probably due to lack of
633encryption keys.
634.SP
635.BI "resolve-error " hostname
636(For
637.BR ADD .)
638The DNS name
639.I hostname
640could not be resolved.
641.SP
642.BI "resolver-timeout " hostname
643(For
644.BR ADD .)
645The DNS name
646.I hostname
647took too long to resolve.
648.SP
649.BI "unknown-command " token
650The command
651.B token
652was not recognised.
653.SP
654.BI "unknown-peer " name
655(For
656.BR ADDR ,
657.BR IFNAME ,
658.BR KILL ,
659.BR SETIFNAME ,
660and
661.BR STATS .)
662There is no peer called
663.IR name .
664.SP
665.BI "unknown-service " service
666(For
667.BR ADD .)
668The service name
669.I service
670couldn't be found in
671.BR /etc/services .
672.SH "NOTIFICATIONS"
673.\"* 30 Notification broadcasts (NOTE codes)
674The following notifications are sent to clients who request them.
675.SP
676.BI "ADD " peer " " ifname " " address \fR...
677A new peer has been added. The peer's name is
678.IR peer ,
679its tunnel is network interface
680.IR ifname ,
681and its network address is
682.IR address .
683.SP
684.BI "DAEMON"
685The server has forked off into the sunset and become a daemon.
686.SP
687.BI "GREET " challenge " " address \fR...
688A valid greeting was received, with the given challenge (exactly as it
689was returned by
690.B GETCHAL
691earlier).
692.SP
693.BI "KILL " peer
694The peer
695.I peer
696has been killed.
697.SP
698.BI "KXDONE " peer
699Key exchange with
700.I peer
701finished successfully.
702.SP
703.BI "KXSTART " peer
704Key exchange with
705.I peer
706has begun or restarted. If key exchange keeps failing, this message
707will be repeated periodically.
708.SP
709.BI "NEWIFNAME " peer " " old-name " " new-name
710The given
711.IR peer 's
712tunnel interface name has been changed from
713.I old-name
714to
715.IR new-name ,
716as a result of a
717.B SETIFNAME
718command.
719.SP
720.BI "USER " tokens\fR...
721An administration client issued a notification using the
722.B NOTIFY
723command.
724.SH "WARNINGS"
725.\"* 40 Warning broadcasts (WARN codes)
726.\"+sep
727There are many possible warnings. They are categorized according to
728their first tokens.
729.PP
730Many of these warnings report system errors. These are reported as a
731pair of tokens, described below as
732.I ecode
733and
734.IR message .
735The
736.I ecode
737is a string of the form
738.BI E number
739giving the
740.BR errno (3)
741value of the error; the
742.I message
743is the `human-readable' form of the message, as reported by
744.BR strerror (3).
745.SS "ABORT warnings"
746These all indicate that the
747.B tripe
748server has become unable to continue. If enabled, the server will dump
749core in its configuration directory.
750.SP
751.BI "ABORT repeated-select-errors"
752The main event loop is repeatedly failing. If the server doesn't quit,
753it will probably waste all available CPU doing nothing.
754.SS "ADMIN warnings"
755These indicate a problem with the administration socket interface.
756.SP
757.BI "ADMIN accept-error " ecode " " message
758There was an error while attempting to accept a connection from a new
759client.
760.SP
761.BI "ADMIN client-write-error " ecode " " message
762There was an error sending data to a client. The connection to the
763client has been closed.
764.SS "CHAL warnings"
765These indicate errors in challenges, either in the
766.B CHECKCHAL
767command or in greeting packets.
768.SP
769.B "CHAL impossible-challenge"
770The server hasn't issued any challenges yet. Quite how anyone else
771thought he could make one up is hard to imagine.
772.SP
773.B "CHAL incorrect-tag"
774Challenge received contained the wrong authentication data. It might be
775very stale, or a forgery.
776.SP
777.B "CHAL invalid-challenge"
778Challenge received was the wrong length. We might have changed MAC
779algorithms since the challenge was issued, or it might just be rubbish.
780.SP
781.B "CHAL replay duplicated-sequence"
782Challenge received was a definite replay of an old challenge. Someone's
783up to something!
784.SP
785.B "CHAL replay old-sequence"
786Challenge received was old, but maybe not actually a replay. Try again.
787.SS "KEYMGMT warnings"
788These indicate a problem with the keyring files, or the keys stored in
789them.
790.SP
791.BI "KEYMGMT bad-private-key " message
792The private key could not be read, or failed a consistency check. If
793there was a problem with the file, usually there will have been
794.B key-file-error
795warnings before this.
796.SP
797.BI "KEYMGMT bad-public-keyring " message
798The public keyring couldn't be read. Usually, there will have been
799.B key-file-error
800warnings before this.
801.SP
802.BI "KEYMGMT key-file-error " file ":" line " " message
803Reports a specific error with the named keyring file. This probably
804indicates a bug in
805.BR key (1).
806.SP
807.BI "KEYMGMT public-key " tag " " tokens\fR...
808These messages all indicate a problem with the public key named
809.IR tag .
810.SP
811.BI "KEYMGMT public-key " tag " algorithm-mismatch"
812The algorithms specified on the public key don't match the ones for our
813private key. All the peers in a network have to use the same
814algorithms.
815.SP
816.BI "KEYMGMT public-key " tag " bad " message
817The public key couldn't be read, or is invalid.
818.SP
819.BI "KEYMGMT public-key " tag " bad-public-group-element"
820The public key is invalid. This may indicate a malicious attempt to
821introduce a bogus key.
822.SP
823.BI "KEYMGMT public-key " tag " bad-algorithm-selection"
824The algorithms listed on the public key couldn't be understood. The
825algorithm selection attributes are probably malformed and need fixing.
826.SP
827.BI "KEYMGMT public-key " tag " incorrect-group"
828The public key doesn't use the same group as our private key. All the
829peers in a network have to use the same group.
830.SP
831.BI "KEYMGMT public-key " tag " not-found"
832The public key for peer
833.I tag
834wasn't in the public keyring.
835.SP
836.BI "KEYMGMT public-key " tag " unknown-type"
837The type of the public key isn't understood. Maybe you need to upgrade
838your copy of
839.BR tripe .
840(Even if you do, you'll have to regenerate your keys.)
841.SS "KX warnings"
842These indicate problems during key-exchange. Many indicate either a bug
843in the server (either yours or the remote one), or some kind of attack
844in progress. All name a
845.I peer
846as the second token: this is the peer the packet is apparently from,
847though it may have been sent by an attacker instead.
848.PP
849In the descriptions below,
850.I msgtoken
851is one of the tokens
852.BR pre-challenge ,
853.BR cookie ,
854.BR challenge ,
855.BR reply ,
856.BR switch-rq ,
857or
858.BR switch-ok .
859.SP
860.BI "KX " peer " bad-expected-reply-log"
861The challenges
862.B tripe
863uses in its protocol contain a check value which proves that the
864challenge is honest. This message indicates that the check value
865supplied is wrong: someone is attempting to use bogus challenges to
866persuade your
867.B tripe
868server to leak private key information. No chance!
869.SP
870.BI "KX " peer " decrypt-failed reply\fR|\fBswitch-ok"
871A symmetrically-encrypted portion of a key-exchange message failed to
872decrypt.
873.SP
874.BI "KX " peer " invalid " msgtoken
875A key-exchange message was malformed. This almost certainly indicates a
876bug somewhere.
877.SP
878.BI "KX " peer " incorrect cookie\fR|\fBswitch-rq\fR|\fBswitch-ok"
879A message didn't contain the right magic data. This may be a replay of
880some old exchange, or random packets being sent in an attempt to waste
881CPU.
882.SP
883.BI "KX " peer " public-key-expired"
884The peer's public key has expired. It's maintainer should have given
885you a replacement before now.
886.SP
887.BI "KX " peer " sending-cookie"
888We've received too many bogus pre-challenge messages. Someone is trying
889to flood us with key-exchange messages and make us waste CPU on doing
890hard asymmetric crypto sums.
891.SP
892.BI "KX " peer " unexpected " msgtoken
893The message received wasn't appropriate for this stage of the key
894exchange process. This may mean that one of our previous packets got
895lost. For
896.BR pre-challenge ,
897it may simply mean that the peer has recently restarted.
898.SP
899.BI "KX " peer " unknown-challenge"
900The peer is asking for an answer to a challenge which we don't know
901about. This may mean that we've been inundated with challenges from
902some malicious source
903.I who can read our messages
904and discarded the valid one.
905.SP
906.BI "KX " peer " unknown-message 0x" nn
907An unknown key-exchange message arrived.
908.SS "PEER warnings"
909These are largely concerned with management of peers and the low-level
910details of the network protocol. The second word is usually the name of
911a peer, or
912.RB ` \- '
913if none is relevant.
914.SP
915.BI "PEER " peer " bad-packet no-type"
916An empty packet arrived. This is very strange.
917.SP
918.BI "PEER " peer " bad-packet unknown-category 0x" nn
919The message category
920.I nn
921(in hex) isn't understood. Probably a strange random packet from
922somewhere; could be an unlikely bug.
923.SP
924.BI "PEER " peer " bad-packet unknown-type 0x" nn
925The message type
926.I nn
927(in hex) isn't understood. Probably a strange random packet from
928somewhere; could be an unlikely bug.
929.SP
930.BI "PEER " peer " corrupt-encrypted-ping"
931The peer sent a ping response which matches an outstanding ping, but its
932payload is wrong. There's definitely a bug somewhere.
933.SP
934.BI "PEER " peer " corrupt-transport-ping"
935The peer (apparently) sent a ping response which matches an outstanding
936ping, but its payload is wrong. Either there's a bug, or the bad guys
937are playing tricks on you.
938.SP
939.BI "PEER " peer " decrypt-failed"
940An encrypted IP packet failed to decrypt. It may have been mangled in
941transit, or may be a very old packet from an expired previous session
942key. There is usually a considerable overlap in the validity periods of
943successive session keys, so this shouldn't occur unless the key exchange
944takes ages or fails.
945.SP
946.BI "PEER " peer " malformed-encrypted-ping"
947The peer sent a ping response which is hopelessly invalid. There's
948definitely a bug somewhere.
949.SP
950.BI "PEER " peer " malformed-transport-ping"
951The peer (apparently) sent a ping response which is hopelessly invalid.
952Either there's a bug, or the bad guys are playing tricks on you.
953.SP
954.BI "PEER " peer " packet-build-failed"
955There wasn't enough space in our buffer to put the packet we wanted to
956send. Shouldn't happen.
957.SP
958.BI "PEER \- socket-read-error " ecode " " message
959An error occurred trying to read an incoming packet.
960.SP
961.BI "PEER " peer " socket-write-error " ecode " " message
962An error occurred attempting to send a network packet. We lost that
963one.
964.SP
965.BI "PEER " peer " unexpected-encrypted-ping 0x" id
966The peer sent an encrypted ping response whose id doesn't match any
967outstanding ping. Maybe it was delayed for longer than the server was
968willing to wait, or maybe the peer has gone mad.
969.SP
970.BI "PEER \- unexpected-source " address\fR...
971A packet arrived from
972.I address
973(a network address \(en see above), but no peer is known at that
974address. This may indicate a misconfiguration, or simply be a result of
975one end of a connection being set up before the other.
976.SP
977.BI "PEER " peer " unexpected-transport-ping 0x" id
978The peer (apparently) sent a transport ping response whose id doesn't
979match any outstanding ping. Maybe it was delayed for longer than the
980server was willing to wait, or maybe the peer has gone mad; or maybe
981there are bad people trying to confuse you.
982.SS "SERVER warnings"
983These indicate problems concerning the server process as a whole.
984.SP
985.BI "SERVER ignore signal " name
986A signal arrived, but the server ignored it. Currently this happens for
987.B SIGHUP
988because that's a popular way of telling daemons to re-read their
989configuration files. Since
990.B tripe
991re-reads its keyrings automatically and has no other configuration
992files, it's not relevant, but it seemed better to ignore the signal than
993let the server die.
994.SP
995.BI "SERVER quit signal " \fR[\fInn\fR|\fIname\fR]
996A signal arrived and
997.B tripe
998is going to quit.
999.SP
1000.BI "SERVER quit admin-request"
1001A client of the administration interface issued a
1002.B QUIT
1003command.
1004.SP
1005.BI "SERVER select-error " ecode " " message
1006An error occurred in the server's main event loop. This is bad: if it
1007happens too many times, the server will abort.
1008.SS "SYMM warnings"
1009These are concerned with the symmetric encryption and decryption
1010process.
1011.SP
1012.BI "SYMM replay old-sequence"
1013A packet was received with an old sequence number. It may just have
1014been delayed or duplicated, or it may have been an attempt at a replay
1015attack.
1016.SP
1017.BI "SYMM replay duplicated-sequence"
1018A packet was received with a sequence number we've definitely seen
1019before. It may be an accidental duplication because the 'net is like
1020that, or a deliberate attempt at a replay.
1021.SS "TUN warnings"
1022These concern the workings of the system-specific tunnel driver. The
1023second word is the name of the tunnel interface in question, or
1024.RB ` \- '
1025if none.
1026.SP
1027.BI "TUN \- bsd no-tunnel-devices"
1028The driver couldn't find an available tunnel device. Maybe if you
1029create some more
1030.BI /dev/tun nn
1031files, it will work.
1032.SP
1033.BI "TUN - " tun-name " open-error " device " " ecode " " message
1034An attempt to open the tunnel device file
1035.I device
1036failed.
1037.SP
1038.BI "TUN \- linux config-error " ecode " " message
1039Configuring the Linux TUN/TAP interface failed.
1040.SP
1041.BI "TUN " ifname " " tun-name " read-error " ecode " " message
1042Reading from the tunnel device failed.
1043.SP
1044.BI "TUN " ifname " slip bad-escape"
1045The SLIP driver encountered a escaped byte it wasn't expecting to see.
1046The erroneous packet will be ignored.
1047.SP
1048.BI "TUN " ifname " slip eof"
1049The SLIP driver encountered end-of-file on its input descriptor.
1050Pending data is discarded, and no attempt is made to read any more data
1051from that interface ever.
1052.SP
1053.BI "TUN " ifname " slip escape-end"
1054The SLIP driver encountered an escaped `end' marker. This probably
1055means that someone's been sending it junk. The erroneous packet is
1056discarded, and we hope that we've rediscovered synchronization.
1057.SP
1058.BI "TUN \- slip fork-error " ecode " " message
1059The SLIP driver encountered an error forking a child process while
1060allocating a new dynamic interface.
1061.SP
1062.BI "TUN \- slip no-slip-interfaces"
1063The driver ran out of static SLIP interfaces. Either preallocate more,
1064or use dynamic SLIP interface allocation.
1065.SP
1066.BI "TUN " ifname " slip overflow"
1067The SLIP driver gave up reading a packet because it got too large.
1068.SP
1069.BI "TUN \- slip pipe-error " ecode " " message
1070The SLIP driver encountered an error creating pipes while allocating a
1071new dynamic interface.
1072.SP
1073.BI "TUN \- slip read-ifname-failed " ecode " " message
1074The SLIP driver encountered an error reading the name of a dynamically
1075allocated interface. Maybe the allocation script is broken.
1076.SP
1077.BI "TUN \- unet config-error " ecode " " message
1078Configuring the Linux Unet interface failed. Unet is obsolete and
1079shouldn't be used any more.
1080.SP
1081.BI "TUN \- unet getinfo-error " ecode " " message
1082Reading information about the Unet interface failed. Unet is obsolete
1083and shouldn't be used any more.
1084.SP
1085.BI "TUN \- unet ifname-too-long"
1086The Unet interface's name overflowed, so we couldn't read it properly.
1087Unet is obsolete and shouldn't be used any more.
1088.SS "USER warnings"
1089These are issued by administration clients using the
1090.B WARN
1091command.
1092.SP
1093.BI "USER " tokens\fR...
1094An administration client issued a warning.
1095.\"-sep
1096.SH "SUMMARY"
1097.SS "Command responses"
1098.nf
1099.BI "BGFAIL " tag " " tokens \fR...
1100.BI "BGINFO " tag " " tokens \fR...
1101.BI "BGOK " tag
1102.BI "FAIL " tokens \fR...
1103.BI "INFO " tokens \fR...
1104.B OK
1105.fi
1106.\"= summary
1107.SH "SEE ALSO"
1108.BR tripectl (1),
1109.BR tripe (8).
1110.PP
1111.IR "The Trivial IP Encryption Protocol" .
1112.SH "AUTHOR"
1113Mark Wooding, <mdw@distorted.org.uk>