diff options
Diffstat (limited to 'lib/diameter/doc/src')
| -rw-r--r-- | lib/diameter/doc/src/notes.xml | 183 |
1 files changed, 183 insertions, 0 deletions
diff --git a/lib/diameter/doc/src/notes.xml b/lib/diameter/doc/src/notes.xml index e6ac332c10..479fab21b2 100644 --- a/lib/diameter/doc/src/notes.xml +++ b/lib/diameter/doc/src/notes.xml @@ -42,6 +42,189 @@ first.</p> <!-- ===================================================================== --> +<section><title>diameter 1.9</title> + + <section><title>Fixed Bugs and Malfunctions</title> + <list> + <item> + <p> + Don't discard outgoing answers unnecessarily.</p> + <p> + Answers missing a Result-Code AVP or setting an E-bit + inappropriately were discarded even if encode was + successful.</p> + <p> + Own Id: OTP-11492</p> + </item> + <item> + <p> + Increase supervision timeouts.</p> + <p> + At diameter application shutdown, DPR could be omitted on + open peer connections because of short supervision + timeouts.</p> + <p> + Own Id: OTP-12412</p> + </item> + <item> + <p> + Fix retransmission of messages sent as header/avps list.</p> + <p> + Extracting End-to-End and Hop-by-Hop Identifiers resulted + in a function clause error, resulting in a handle_error + callback.</p> + <p> + Own Id: OTP-12415</p> + </item> + <item> + <p> + Fix diameter_avp decode of Grouped AVPs having decode + errors.</p> + <p> + Components of such an AVP were not extracted, causing it + to be represented by a single diameter_avp record instead + of the intended list.</p> + <p> + Dictionary files must be recompiled for the fix to have + effect.</p> + <p> + Own Id: OTP-12475</p> + </item> + <item> + <p> + Fix ordering of AVPs in relayed messages.</p> + <p> + The order was reversed relative to the received order, + with a Route-Record AVP prepended.</p> + <p> + Thanks to Andrzej Trawiński.</p> + <p> + Own Id: OTP-12551</p> + </item> + <item> + <p> + Fix issues with DiameterURI encode/decode.</p> + <p> + RFC 6773 changed the default port and transport, but the + RFC 3588 defaults were used even if the RFC 6733 common + dictionary was in use. The RFC 3588 defaults are now only + used when the common dictionary is + diameter_gen_base_rfc3588.</p> + <p> + Both RFC 3588 and 6733 disallow + transport=udp;protocol=diameter. Encode of the + combination now fails.</p> + <p> + Decode of ports numbers outside the range 0-65535 and + fully qualified domain names longer than 255 octets now + fails.</p> + <p> + Note that RFC 3588 is obsolete, and that there is a + diameter_gen_base_rfc6733. The change in defaults is a + potential interoperability problem when moving to RFC + 6733 with peers that do not send all URI components. The + fact that 6733 allows 5xxx result codes in answer + messages setting the E-bit, which RFC 3588 doesn't, is + another.</p> + <p> + Own Id: OTP-12589</p> + </item> + </list> + </section> + + + <section><title>Improvements and New Features</title> + <list> + <item> + <p> + Add service_opt() string_decode.</p> + <p> + To disable the decode of potentially large binaries to + string. This prevents large strings from being copied + when incoming Diameter messages are passed between + processes, a vulnerability that can lead to memory being + exhausted given sufficiently malicious peers.</p> + <p> + The value is a boolean(), true being the default for + backwards compatibility. Setting false causes both + diameter_caps records and decoded messages to contain + binary() in relevant places that previously had string(): + diameter_app(3) callbacks need to be prepared for the + change.</p> + <p> + The Diameter types affected are OctetString and the + derived types UTF8String, DiameterIdentity, DiameterURI, + IPFilterRule, and QoSFilterRule. Time and Address are + unaffected.</p> + <p> + Own Id: OTP-11952</p> + </item> + <item> + <p> + Add transport_opt() pool_size.</p> + <p> + To allow for pools of accepting transport processes, + which can better service multiple simultaneous peer + connections. The option can also be used with connecting + transports, to establish multiple connections to the same + peer without having to configure multiple transports.</p> + <p> + Own Id: OTP-12428</p> + </item> + <item> + <p> + Allow DPR to be sent with diameter:call/4.</p> + <p> + It has been possible to send, but the answer was regarded + as unsolicited and discarded. DPA now causes the + transport process in question to be terminated, as for + DPR that diameter itself sends.</p> + <p> + Own Id: OTP-12542</p> + </item> + <item> + <p> + Discard requests after DPR.</p> + <p> + RFC 6733 is imprecise, but the tone is that messages + received after DPR are an exception to be dealt with only + because of the possibility of unordered delivery over + SCTP. As a consequence, and because a request following + DPR is unlikely to be answered due to the impending loss + of the peer connection, discard outgoing requests + following an outgoing or incoming DPR. Incoming requests + are also discarded, with the exception of DPR itself. + Answers are sent and received as usual.</p> + <p> + Own Id: OTP-12543</p> + </item> + <item> + <p> + Add transport_opt() dpr_timeout.</p> + <p> + To cause a peer connection to be closed following an + outgoing DPA when the peer fails to do so. It is the + recipient of DPA that should close the connection + according to RFC 6733.</p> + <p> + Own Id: OTP-12609</p> + </item> + <item> + <p> + Add service_opt() incoming_maxlen.</p> + <p> + To bound the expected size of incoming Diameter messages. + Messages larger than the specified number of bytes are + discarded, to prevent a malicious peer from generating + excessive load.</p> + <p> + Own Id: OTP-12628</p> + </item> + </list> + </section> + +</section> + <section><title>diameter 1.8</title> <section><title>Fixed Bugs and Malfunctions</title> |