From 502189ba42469d3332bc0658caa2bd0de1e3fcb9 Mon Sep 17 00:00:00 2001
From: Anders Svensson <anders@erlang.org>
Date: Mon, 24 Aug 2015 16:14:49 +0200
Subject: Add service_opt() strict_mbit

There are differing opinions on whether or not reception of an arbitrary
AVP setting the M-bit is an error. 1.3.4 of RFC 6733 says this about
how an existing Diameter application may be modified:

   o  The M-bit allows the sender to indicate to the receiver whether or
      not understanding the semantics of an AVP and its content is
      mandatory.  If the M-bit is set by the sender and the receiver
      does not understand the AVP or the values carried within that AVP,
      then a failure is generated (see Section 7).

   It is the decision of the protocol designer when to develop a new
   Diameter application rather than extending Diameter in other ways.
   However, a new Diameter application MUST be created when one or more
   of the following criteria are met:

   M-bit Setting

      An AVP with the M-bit in the MUST column of the AVP flag table is
      added to an existing Command/Application.  An AVP with the M-bit
      in the MAY column of the AVP flag table is added to an existing
      Command/Application.

The point here is presumably interoperability: that the command grammar
should specify explicitly what mandatory AVPs much be understood, and
that anything more is an error.

On the other hand, 3.2 says thus about command grammars:

   avp-name         = avp-spec / "AVP"
                      ; The string "AVP" stands for *any* arbitrary AVP
                      ; Name, not otherwise listed in that Command Code
                      ; definition.  The inclusion of this string
                      ; is recommended for all CCFs to allow for
                      ; extensibility.

This renders 1.3.4 pointless unless "*any* AVP" is qualified by "not
setting the M-bit", since the sender can effectively violate 1.3.4
without this necessitating an error at the receiver. If clients add
arbitrary AVPs setting the M-bit then request handling becomes more
implementation-dependent.

The current interpretation in diameter is strict: if a command grammar
doesn't explicitly allow an AVP setting the M-bit then reception of such
an AVP is regarded as an error. The strict_mbit option now allows this
behaviour to be changed, false turning all responsibility for the M-bit
over to the user.
---
 lib/diameter/doc/src/diameter.xml | 43 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 43 insertions(+)

(limited to 'lib/diameter/doc')

diff --git a/lib/diameter/doc/src/diameter.xml b/lib/diameter/doc/src/diameter.xml
index 854bc5b432..b0cff32c9a 100644
--- a/lib/diameter/doc/src/diameter.xml
+++ b/lib/diameter/doc/src/diameter.xml
@@ -912,6 +912,49 @@ Options <c>monitor</c> and <c>link</c> are ignored.</p>
 Defaults to the empty list.</p>
 </item>
 
+<marker id="strict_mbit"/>
+<tag><c>{strict_mbit, boolean()}</c></tag>
+<item>
+<p>
+Whether or not to regard an AVP setting the M-bit as erroneous when
+the command grammar in question does not explicitly allow the AVP.
+If <c>true</c> then such AVPs are regarded as 5001 errors,
+DIAMETER_AVP_UNSUPPORTED.
+If <c>false</c> then the M-bit is ignored and policing
+it becomes the receiver's responsibility.</p>
+
+<p>
+Defaults to <c>true</c>.</p>
+
+<warning>
+<p>
+RFC 6733 is unclear about the semantics of the M-bit.
+One the one hand, the CCF specification in section 3.2 documents AVP
+in a command grammar as meaning <b>any</b> arbitrary AVP; on the
+other hand, 1.3.4 states that AVPs setting the M-bit cannot be added
+to an existing command: the modified command must instead be
+placed in a new Diameter application.</p>
+<p>
+The reason for the latter is presumably interoperability:
+allowing arbitrary AVPs setting the M-bit in a command makes its
+interpretation implementation-dependent, since there's no
+guarantee that all implementations will understand the same set of
+arbitrary AVPs in the context of a given command.
+However, interpreting <c>AVP</c> in a command grammar as <b>any</b>
+AVP, regardless of M-bit, renders 1.3.4 meaningless, since the receiver
+can simply ignore any AVP it thinks isn't relevant, regardless of the
+sender's intent.</p>
+<p>
+Beware of confusing mandatory in the sense of the M-bit with mandatory
+in the sense of the command grammar.
+The former is a semantic requirement: that the receiver understand the
+semantics of the AVP in the context in question.
+The latter is a syntactic requirement: whether or not the AVP must
+occur in the message in question.</p>
+</warning>
+
+</item>
+
 <marker id="string_decode"/>
 <tag><c>{string_decode, boolean()}</c></tag>
 <item>
-- 
cgit v1.2.3