server/peer.c: Always send `BYE' packets on peer death.

[tripe] / server / tripe-admin.5.in

diff --git a/server/tripe-admin.5.in b/server/tripe-admin.5.in
index cc97298a0bcaf94283a55d8b7363a92b6fdb4ffe..50933a8a8efc1248c39a060f0b171390ffd6f732 100644 (file)
--- a/server/tripe-admin.5.in
+++ b/server/tripe-admin.5.in
@@ -363,18 +363,18 @@ sends us something before responding.
 .TP
 .B "\-ephemeral"
 The association with the peer is not intended to persist indefinitely.
-If a peer marked as ephemeral is killed, or the
+When a peer is killed, or the
 .BR tripe (8)
-daemon is shut down, send a
+daemon is shut down, a
 .B bye
-packet to the peer so that it forgets about us; if a peer marked as
-ephemeral sends us a
+packet is to the peer(s).  If a peer marked as ephemeral sends us a
 .B bye
 packet then it is killed (but in this case no further
 .B bye
-packet is sent).  Peers not marked as ephemeral exhibit neither of these
-behaviours; each peer must have the other marked as ephemeral for the
-association to be fully torn down if either end kills the other.
+packet is sent).  A
+.B bye
+packet from a peer which isn't marked as ephemeral leaves the peer alone
+in the hope that the connection can be reestablished.
 .TP
 .BI "\-keepalive " time
 Send a no-op packet if we've not sent a packet to the peer in the last
@@ -414,7 +414,7 @@ emits a
 .B KNOCK
 notification stating the peer's (claimed) name and address.  The server
 will already have verified that the sender is using the peer's private
-key by this point.  This option implies
+key by this point.  Prior to version 1.6.0, this option used to imply
 .BR \-ephemeral .
 .TP
 .B "\-mobile"
@@ -425,7 +425,7 @@ peers, however, it will attempt to decrypt the packet using their keys,
 and if one succeeds, the server will update its idea of the peer's
 address and emit an
 .B NEWADDR
-notification.  This option implies
+notification.  Prior to version 1.6.0, this option used to imply
 .BR \-ephemeral .
 .TP
 .BI "\-priv " tag
@@ -1542,6 +1542,51 @@ The key specifies the use of an unknown serialization format
 for hashing group elements.  Maybe the key was generated wrongly, or
 maybe the version of Catacomb installed is too old.
 .SP
+.BI "KEYMGMT " which "-keyring " file " key " tag " unsuitable-aead-cipher " cipher "no-aad"
+The key specifies the use of an authenticated encryption scheme
+.I cipher
+which does not support the processing of additional authenticated data.
+The most prominent examples of such schemes are the
+.IB cipher -naclbox
+collection, where
+.I cipher
+is
+.BR salsa20 ,
+.BR salsa20/12 ,
+.BR salsa20/8 ,
+.BR chacha20 ,
+.BR chacha12 ,
+or
+.BR chacha8 ;
+use the
+.B naclbox
+bulk transform rather than
+.B aead
+for these
+(or switch to the IETF
+.IB cipher -poly1305
+schemes instead).
+.SP
+.BI "KEYMGMT " which "-keyring " file " key " tag " unsuitable-aead-cipher " cipher "nonce-too-small"
+The key specifies the use of an authenticated encryption scheme
+.I cipher
+which doesn't even allow a 5-byte (40-bit) nonce.  Catacomb doesn't
+implement any such limited AE schemes: you must be doing something
+strange.
+.SP
+.BI "KEYMGMT " which "-keyring " file " key " tag " unsuitable-aead-cipher " cipher "nonce-too-large"
+The key specifies the use of an authenticated encryption scheme
+.I cipher
+which doesn't support any nonce size smaller than 64 bytes (512 bits).
+Catacomb doesn't implement any such extravagant AE schemes: you must be
+doing something strange.
+.SP
+.BI "KEYMGMT " which "-keyring " file " key " tag " unsuitable-aead-cipher " cipher "nonempty-ciphertext-for-empty-message"
+The key specifies the use of an authenticated encryption scheme
+.I cipher
+which produces ciphertext output even when given a completely empty
+message.  Catacomb doesn't implement any such unhelpful AE schemes: you
+must be doing something strange.
 .SP
 .BI "KEYMGMT " which "-keyring " file " key " tag " " alg " " name " no-key-size " hashsz
 The