SIP IPv6 ABNF fix: Essential Correction for IPv6 ABNF and URI Comparison in RFC 3261
A fix for a faulty IPv6 grammar rule in SIP, plus a rule that URIs compare by binary address, not by spelling.
Gurbani, V., Ed., Carpenter, B., Ed., and B. Tate, Ed., "Essential Correction for IPv6 ABNF and URI Comparison in RFC 3261", RFC 5954, DOI 10.17487/RFC5954, August 2010, <https://www.rfc-editor.org/info/rfc5954>.
@misc{rfc5954,
series = {Request for Comments},
number = 5954,
howpublished = {RFC 5954},
publisher = {RFC Editor},
doi = {10.17487/RFC5954},
url = {https://www.rfc-editor.org/info/rfc5954},
author = {V. Gurbani and B. Carpenter and B. Tate},
title = {{Essential Correction for IPv6 ABNF and URI Comparison in RFC 3261}},
pagetotal = 7,
year = 2010,
month = aug,
} RFC text
Contents
Internet Engineering Task Force (IETF) V. Gurbani, Ed.
Request for Comments: 5954 Bell Laboratories, Alcatel-Lucent
Updates: 3261 B. Carpenter, Ed.
Category: Standards Track Univ. of Auckland
ISSN: 2070-1721 B. Tate, Ed.
BroadSoft
August 2010
Essential Correction for IPv6 ABNF and URI Comparison in RFC 3261
Abstract
This document corrects the Augmented Backus-Naur Form (ABNF)
production rule associated with generating IPv6 literals in RFC 3261.
It also clarifies the rule for Uniform Resource Identifier (URI)
comparison when the URIs contain textual representation of IP
addresses.
Status of This Memo
This is an Internet Standards Track document.
This document is a product of the Internet Engineering Task Force
(IETF). It represents the consensus of the IETF community. It has
received public review and has been approved for publication by the
Internet Engineering Steering Group (IESG). Further information on
Internet Standards is available in Section 2 of RFC 5741.
Information about the current status of this document, any errata,
and how to provide feedback on it may be obtained at
http://www.rfc-editor.org/info/rfc5954.
Copyright Notice
Copyright (c) 2010 IETF Trust and the persons identified as the
document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents
(http://trustee.ietf.org/license-info) in effect on the date of
publication of this document. Please review these documents
carefully, as they describe your rights and restrictions with respect
to this document. Code Components extracted from this document must
include Simplified BSD License text as described in Section 4.e of
the Trust Legal Provisions and are provided without warranty as
described in the Simplified BSD License.
Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 2
2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . 2
3. Problem Statement . . . . . . . . . . . . . . . . . . . . . . . 2
3.1. Extra Colon in IPv4-Mapped IPv6 Address . . . . . . . . . . 2
3.2. Comparing URIs with Textual Representation of IP
Addresses . . . . . . . . . . . . . . . . . . . . . . . . . 3
4. Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . 4
4.1. Resolution for Extra Colon in IPv4-Mapped IPv6 Address . . 4
4.2. Clarification for Comparison of URIs with Textual
Representation of IP Addresses . . . . . . . . . . . . . . 5
5. Generating a Canonical IPv6 Textual Representation . . . . . . 5
6. Security Considerations . . . . . . . . . . . . . . . . . . . . 6
7. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . 6
8. References . . . . . . . . . . . . . . . . . . . . . . . . . . 6
8.1. Normative References . . . . . . . . . . . . . . . . . . . 6
8.2. Informative References . . . . . . . . . . . . . . . . . . 6
1. Introduction
This document corrects the Augmented Backus-Naur Form (ABNF)
production rule associated with generating IPv6 literals in RFC 3261
[1]. It also clarifies the rule for Uniform Resource Identifier
(URI) comparison when the URIs contain textual representation of IP
addresses.
2. Terminology
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in RFC 2119 [2].
3. Problem Statement
3.1. Extra Colon in IPv4-Mapped IPv6 Address
The ABNF [4] for generating IPv6 literals in RFC 3261 [1] is
incorrect. When generating IPv4-mapped IPv6 addresses, the
production rule may actually generate the following construct:
[2001:db8:::192.0.2.1] - Note the extra colon before the IPv4
address.
The correct construct, of course, would only include two colons
before the IPv4 address.
Historically, the ABNF pertaining to IPv6 references in RFC 3261
was derived from Appendix B of RFC 2373 [7], which was flawed to
begin with (see errata for RFC 2373 [8]). RFC 2373 has been
subsequently obsoleted by RFC 4291 [6].
The ABNF for IPv6reference is reproduced from RFC 3261 below:
IPv6reference = "[" IPv6address "]"
IPv6address = hexpart [ ":" IPv4address ]
IPv4address = 1*3DIGIT "." 1*3DIGIT "." 1*3DIGIT "." 1*3DIGIT
hexpart = hexseq / hexseq "::" [ hexseq ] / "::" [ hexseq ]
hexseq = hex4 *( ":" hex4)
hex4 = 1*4HEXDIG
Note that the ambiguity occurs in the <IPv6address> production rule
where the <IPv4address> non-terminal is prefixed by the ":" token.
Because the <hexpart> production rule is defined such that two of its
alternatives already include the "::" token, this may yield to the
faulty construction of an IPv6-mapped IPv4 address with an extra
colon when expanding those alternatives.
3.2. Comparing URIs with Textual Representation of IP Addresses
In SIP, URIs are compared for a variety of reasons. Registrars
compare URIs when they receive a binding update request, for
instance. Section 19.1.4 of RFC 3261 [1] provides the rules for
comparing URIs. Among other rules, it states that:
For two URIs to be equal, the user, password, host, and port
components must match.
Does the above rule then imply that the following URIs are equal:
sip:bob@[::ffff:192.0.2.128] = sip:bob@[::ffff:c000:280]?
sip:bob@[2001:db8::9:1] = sip:bob@[2001:db8::9:01]?
sip:bob@[0:0:0:0:0:FFFF:129.144.52.38] = sip:bob@
[::FFFF:129.144.52.38]?
In all of the above examples, the textual representation of the IPv6
address is different, but these addresses are binary equivalents
(implementers are also urged to consult Section 5 of this document
for recommendations on IPv6 address text representations). Section
19.1.4 of RFC 3261 does not provide any rule for URIs containing
different textual representations of IPv6 addresses that all
correspond to the same binary equivalent.
Note that the same ambiguity occurs for IPv4 addresses, i.e., is
192.0.2.128 = 192.00.02.128? However, IPv6, with its compressed
notation and the need to represent hybrid addresses (like IPv4-
mapped IPv6 addresses) makes the representation issue more acute.
The resolution discussed in Section 4.2 applies to textual
representations of both IPv6 and IPv4 addresses.
4. Resolution
4.1. Resolution for Extra Colon in IPv4-Mapped IPv6 Address
The resolution to this ambiguity is simply to use the correct ABNF
for the <IPv6address> production rule from Appendix A of RFC 3986
[3]. For the sake of completeness, it is reproduced below:
IPv6address = 6( h16 ":" ) ls32
/ "::" 5( h16 ":" ) ls32
/ [ h16 ] "::" 4( h16 ":" ) ls32
/ [ *1( h16 ":" ) h16 ] "::" 3( h16 ":" ) ls32
/ [ *2( h16 ":" ) h16 ] "::" 2( h16 ":" ) ls32
/ [ *3( h16 ":" ) h16 ] "::" h16 ":" ls32
/ [ *4( h16 ":" ) h16 ] "::" ls32
/ [ *5( h16 ":" ) h16 ] "::" h16
/ [ *6( h16 ":" ) h16 ] "::"
h16 = 1*4HEXDIG
ls32 = ( h16 ":" h16 ) / IPv4address
IPv4address = dec-octet "." dec-octet "." dec-octet "." dec-octet
dec-octet = DIGIT ; 0-9
/ %x31-39 DIGIT ; 10-99
/ "1" 2DIGIT ; 100-199
/ "2" %x30-34 DIGIT ; 200-249
/ "25" %x30-35 ; 250-255
Accordingly, this document updates RFC 3261 as follows: the
<IPv6address> and <IPv4address> production rules from RFC 3261 MUST
NOT be used and instead, the production rules of the same name in RFC
3986 (and reproduced above) MUST be used. This will render
<hexpart>, <hexseq>, and <hex4> production rules in RFC 3261
obsolete; as such, these three production rules -- namely, <hexpart>,
<hexseq>, and <hex4> -- from RFC 3261 MUST NOT be used.
The use of the <IPv4address> production rule from RFC 3986 no longer
allows syntactically valid -- though semantically invalid -- SIP URIs
of the form "sip:bob@444.555.666.777".
4.2. Clarification for Comparison of URIs with Textual Representation
of IP Addresses
The resolution to this ambiguity is a simple clarification
acknowledging that the textual representation of an IP address
varies, but it is the binary equivalence of the IP address that must
be taken into consideration when comparing two URIs that contain
varying textual representations of an IP address.
Accordingly, the existing rule from the bulleted list in Section
19.1.4 of RFC 3261 MUST be modified as follows:
OLD:
o For two URIs to be equal, the user, password, host, and port
components must match.
NEW:
o For two URIs to be equal, the user, password, host, and port
components must match. If the host component contains a textual
representation of IP addresses, then the representation of those
IP addresses may vary. If so, the host components are considered
to match if the different textual representations yield the same
binary IP address.
In addition, the text in the following paragraph MUST be added to the
existing list of examples in Section 19.1.4 of RFC 3261 in order to
demonstrate the intent of the modified rule:
The following URIs are equivalent because the underlying binary
representation of the IP addresses are the same although their
textual representations vary:
sip:bob@[::ffff:192.0.2.128]
sip:bob@[::ffff:c000:280]
sip:bob@[2001:db8::9:1]
sip:bob@[2001:db8::9:01]
sip:bob@[0:0:0:0:0:FFFF:129.144.52.38]
sip:bob@[::FFFF:129.144.52.38]
5. Generating a Canonical IPv6 Textual Representation
Implementers SHOULD generate IPv6 text representation as defined in
RFC 5952 [5].
6. Security Considerations
This document does not introduce any new security considerations
beyond those described in RFC 3261 [1].
7. Acknowledgments
The ABNF for IPv6 was developed by Roy T. Fielding and Andrew Main
and published in RFC 3986.
Jeroen van Bemmel, Peter Blatherwick, Gonzalo Camarillo, Paul
Kyzivat, Jonathan Rosenberg, Michael Thomas, and Dale Worley provided
invaluable discussion points on the SIP WG mailing list on the URI
equivalency problem. Alfred Hoenes urged the use of angle brackets
(as specified in Section 2.1 of RFC 5234 [4]) to denote productions.
8. References
8.1. Normative References
[1] Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, A.,
Peterson, J., Sparks, R., Handley, M., and E. Schooler, "SIP:
Session Initiation Protocol", RFC 3261, June 2002.
[2] Bradner, S., "Key words for use in RFCs to Indicate Requirement
Levels", BCP 14, RFC 2119, March 1997.
[3] Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform
Resource Identifier (URI): Generic Syntax", STD 66, RFC 3986,
January 2005.
[4] Crocker, D. and P. Overell, "Augmented BNF for Syntax
Specifications: ABNF", STD 68, RFC 5234, January 2008.
[5] Kawamura, S. and M. Kawashima, "A Recommendation for IPv6
Address Text Representation", RFC 5952, August 2010.
8.2. Informative References
[6] Hinden, R. and S. Deering, "IP Version 6 Addressing
Architecture", RFC 4291, February 2006.
[7] Hinden, R. and S. Deering, "IP Version 6 Addressing
Architecture", RFC 2373, July 1998.
[8] "RFC Editor Errata", <http://www.rfc-editor.org/errata.php>.
Authors' Addresses
Vijay K. Gurbani (editor)
Bell Laboratories, Alcatel-Lucent
1960 Lucent Lane
Room 9C-533
Naperville, IL 60563
USA
Phone: +1 630 224-0216
EMail: vkg@bell-labs.com
Brian E. Carpenter (editor)
Department of Computer Science
University of Auckland
PB 92019
Auckland, 1142
New Zealand
EMail: brian.e.carpenter@gmail.com
Brett Tate (editor)
BroadSoft
EMail: brett@broadsoft.com
AI ELI5
On this page
§1What This Correction Is
This document corrects the Augmented Backus-Naur Form (ABNF) production rule that RFC 3261 (SIP) uses for generating IPv6 literals. It also clarifies how Uniform Resource Identifiers (URIs) are compared when they contain textual representations of IP addresses.
It is a short, targeted patch and formally updates RFC 3261.
It is a small repair to an older rulebook. One rule for writing addresses was wrong, and another rule about when two addresses count as the same was missing.
§3.1The Extra Colon Bug
The ABNF for IPv6 literals in RFC 3261 is incorrect. When generating IPv4-mapped IPv6 addresses it can produce [2001:db8:::192.0.2.1], with an extra colon before the IPv4 part. The correct form has only two colons there.
The cause is in the IPv6address rule, which prefixes IPv4address with a : token. Two alternatives of the hexpart rule already end with ::, so adding another colon gives the faulty result. The RFC says this ABNF was derived from Appendix B of RFC 2373, which was flawed to begin with. RFC 2373 was later obsoleted by RFC 4291.
The old rule sometimes let you write an address with one colon too many, like typing a double comma in a sentence. The result looked fine to the rule but was not a real address.
§4.1The Grammar Fix
The fix is to use the IPv6address rule from Appendix A of RFC 3986. This document reproduces it, along with h16, ls32, IPv4address and dec-octet.
The update to RFC 3261 is:
- The RFC 3261
IPv6addressandIPv4addressrules MUST NOT be used. The same-named rules from RFC 3986 MUST be used. - This makes the
hexpart,hexseqandhex4rules from RFC 3261 obsolete, so they MUST NOT be used.
A side effect is that the stricter IPv4address rule no longer allows syntactically valid but semantically invalid SIP URIs such as sip:bob@444.555.666.777.
Instead of patching the broken rule, the authors swap in the correct version that another rulebook already has. Impossible addresses like 444.555.666.777 also stop being allowed.
§3.2Comparing URIs With IP Addresses
SIP compares URIs for many reasons. For example, registrars compare them when they receive a binding update. Section 19.1.4 of RFC 3261 said that for two URIs to be equal, the user, password, host and port components must match. That leaves a question when the same IPv6 address is written differently, such as sip:bob@[2001:db8::9:1] and sip:bob@[2001:db8::9:01]. They look different as text but are binary equivalents.
The same ambiguity exists for IPv4 (is 192.0.2.128 equal to 192.00.02.128?), but IPv6 compressed notation and hybrid addresses make it more acute.
Writing 7 and 07 uses different letters for the same number. The old rule did not say whether a computer should treat those as the same address or different ones.
§4.2The New Comparison Rule
The text says it is the binary equivalence of the IP address that must be considered. Section 19.1.4 of RFC 3261 is modified. The old rule (user, password, host and port components must match) gets this addition: if the host contains a textual IP address that may vary in representation, the host components match if the different textual representations yield the same binary IP address.
The RFC also requires adding examples to Section 19.1.4 of RFC 3261. The following pairs are equivalent:
sip:bob@[::ffff:192.0.2.128]andsip:bob@[::ffff:c000:280]sip:bob@[2001:db8::9:1]andsip:bob@[2001:db8::9:01]sip:bob@[0:0:0:0:0:FFFF:129.144.52.38]andsip:bob@[::FFFF:129.144.52.38]
This applies to textual forms of both IPv6 and IPv4 addresses.
The new rule says to compare where the address points, not how it is spelled. If two spellings point to the same place, they match.
OLD:
For two URIs to be equal, the user, password, host, and port
components must match.
NEW:
For two URIs to be equal, the user, password, host, and port
components must match. If the host component contains a textual
representation of IP addresses, then the representation of those
IP addresses may vary. If so, the host components are considered
to match if the different textual representations yield the same
binary IP address.§5Canonical IPv6 Text and Security
Implementers SHOULD generate IPv6 text representations as defined in RFC 5952, which keeps the output consistent. Section 3.2 of this RFC also points readers there for recommendations on IPv6 text.
On security, the RFC says it introduces no new considerations beyond those in RFC 3261.
When writing out an address, people should use the standard tidy spelling so everyone writes it the same way. This fix does not create any new security problems.