Internet Engineering Task Force (IETF)                           F. Gont
Request for Comments: 7112                           Huawei Technologies
Updates: 2460                                                  V. Manral
Category: Standards Track                                 Ionos Networks
ISSN: 2070-1721                                                R. Bonica
                                                       Juniper Networks
                                                           January 2014


             Implications of Oversized IPv6 Header Chains

Abstract

  The IPv6 specification allows IPv6 Header Chains of an arbitrary
  size.  The specification also allows options that can, in turn,
  extend each of the headers.  In those scenarios in which the IPv6
  Header Chain or options are unusually long and packets are
  fragmented, or scenarios in which the fragment size is very small,
  the First Fragment of a packet may fail to include the entire IPv6
  Header Chain.  This document discusses the interoperability and
  security problems of such traffic, and updates RFC 2460 such that the
  First Fragment of a packet is required to contain the entire IPv6
  Header Chain.

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/rfc7112.














Gont, et al.                 Standards Track                    [Page 1]

RFC 7112         Implications of Oversized Header Chains    January 2014


Copyright Notice

  Copyright (c) 2014 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. Requirements Language ...........................................3
  3. Terminology .....................................................3
  4. Motivation ......................................................4
  5. Updates to RFC 2460 .............................................5
  6. IANA Considerations .............................................5
  7. Security Considerations .........................................6
  8. Acknowledgements ................................................6
  9. References ......................................................7
     9.1. Normative References .......................................7
     9.2. Informative References .....................................7

1.  Introduction

  With IPv6, optional internet-layer information is carried in one or
  more IPv6 Extension Headers [RFC2460].  Extension Headers are placed
  between the IPv6 header and the Upper-Layer Header in a packet.  The
  term "Header Chain" refers collectively to the IPv6 header, Extension
  Headers, and Upper-Layer Header occurring in a packet.  In those
  scenarios in which the IPv6 Header Chain is unusually long and
  packets are fragmented, or scenarios in which the fragment size is
  very small, the Header Chain may span multiple fragments.

  While IPv4 had a fixed maximum length for the set of all IPv4 options
  present in a single IPv4 packet, IPv6 does not have any equivalent
  maximum limit at present.  This document updates the set of IPv6
  specifications to create an overall limit on the size of the
  combination of IPv6 options and IPv6 Extension Headers that is
  allowed in a single IPv6 packet.  Namely, it updates RFC 2460 such
  that the First Fragment of a fragmented datagram is required to
  contain the entire IPv6 Header Chain.



Gont, et al.                 Standards Track                    [Page 2]

RFC 7112         Implications of Oversized Header Chains    January 2014


  It should be noted that this requirement does not preclude the use of
  large payloads but, instead, merely requires that all headers,
  starting from the IPv6 base header and continuing up to the Upper-
  Layer Header (e.g., TCP or the like) be present in the First
  Fragment.

2.  Requirements Language

  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 [RFC2119].

3.  Terminology

  For the purposes of this document, the terms Extension Header, IPv6
  Header Chain, First Fragment, and Upper-Layer Header are used as
  follows:

  Extension Header:

     Extension Headers are defined in Section 4 of [RFC2460].  As a
     result of [RFC7045], [IANA-PROTO] provides a list of assigned
     Internet Protocol Numbers and designates which of those protocol
     numbers also represent Extension Headers.

  First Fragment:

     An IPv6 fragment with Fragment Offset equal to 0.

  IPv6 Header Chain:

     The IPv6 Header Chain contains an initial IPv6 header, zero or
     more IPv6 Extension Headers, and optionally, a single Upper-Layer
     Header.  If an Upper-Layer Header is present, it terminates the
     header chain; otherwise, the "No Next Header" value (Next Header =
     59) terminates it.

     The first member of the IPv6 Header Chain is always an IPv6
     header.  For a subsequent header to qualify as a member of the
     header chain, it must be referenced by the "Next Header" field of
     the previous member of the header chain.  However, if a second
     IPv6 header appears in the header chain, as is the case when IPv6
     is tunneled over IPv6, the second IPv6 header is considered to be
     an Upper-Layer Header and terminates the header chain.  Likewise,
     if an Encapsulating Security Payload (ESP) header appears in the
     header chain, it is considered to be an Upper-Layer Header, and it
     terminates the header chain.




Gont, et al.                 Standards Track                    [Page 3]

RFC 7112         Implications of Oversized Header Chains    January 2014


  Upper-Layer Header:

     In the general case, the Upper-Layer Header is the first member of
     the header chain that is neither an IPv6 header nor an IPv6
     Extension Header.  However, if either an ESP header, or a second
     IPv6 header occur in the header chain, they are considered to be
     Upper-Layer Headers, and they terminate the header chain.

     Neither the upper-layer payload, nor any protocol data following
     the upper-layer payload, is considered to be part of the IPv6
     Header Chain.  In a simple example, if the Upper-Layer Header is a
     TCP header, the TCP payload is not part of the IPv6 Header Chain.
     In a more complex example, if the Upper-Layer Header is an ESP
     header, neither the payload data, nor any of the fields that
     follow the payload data in the ESP header are part of the IPv6
     Header Chain.

4.  Motivation

  Many forwarding devices implement stateless firewalls.  A stateless
  firewall enforces a forwarding policy on a packet-by-packet basis.
  In order to enforce its forwarding policy, the stateless firewall may
  need to glean information from both the IPv6 and upper-layer headers.

  For example, assume that a stateless firewall discards all traffic
  received from an interface unless it is destined for a particular TCP
  port on a particular IPv6 address.  When this firewall is presented
  with a fragmented packet that is destined for a different TCP port,
  and the entire header chain is contained within the First Fragment,
  the firewall discards the First Fragment and allows subsequent
  fragments to pass.  Because the First Fragment was discarded, the
  packet cannot be reassembled at the destination.  Insomuch as the
  packet cannot be reassembled, the forwarding policy is enforced.

  However, when the firewall is presented with a fragmented packet and
  the header chain spans multiple fragments, the First Fragment does
  not contain enough information for the firewall to enforce its
  forwarding policy.  Lacking sufficient information, the stateless
  firewall either forwards or discards that fragment.  Regardless of
  the action that it takes, it may fail to enforce its forwarding
  policy.










Gont, et al.                 Standards Track                    [Page 4]

RFC 7112         Implications of Oversized Header Chains    January 2014


5.  Updates to RFC 2460

  When a host fragments an IPv6 datagram, it MUST include the entire
  IPv6 Header Chain in the First Fragment.

  A host that receives a First Fragment that does not satisfy the
  above-stated requirement SHOULD discard the packet and SHOULD send an
  ICMPv6 error message to the source address of the offending packet
  (subject to the rules for ICMPv6 errors specified in [RFC4443]).
  However, for backwards compatibility, implementations MAY include a
  configuration option that allows such fragments to be accepted.

  Likewise, an intermediate system (e.g., router or firewall) that
  receives an IPv6 First Fragment that does not satisfy the above-
  stated requirement MAY discard that packet, and it MAY send an ICMPv6
  error message to the source address of the offending packet (subject
  to the rules for ICMPv6 error messages specified in [RFC4443]).
  Intermediate systems having this capability SHOULD support
  configuration (e.g., enable/disable) of whether or not such packets
  are dropped by the intermediate system.

  If a host or intermediate system discards a First Fragment because it
  does not satisfy the above-stated requirement and sends an ICMPv6
  error message due to the discard, then the ICMPv6 error message MUST
  be Type 4 ("Parameter Problem") and MUST use Code 3 ("First Fragment
  has incomplete IPv6 Header Chain").  The Pointer field contained by
  the ICMPv6 Parameter Problem message MUST be set to zero.  The format
  for the ICMPv6 error message is the same regardless of whether a host
  or intermediate system originates it.

  As a result of the above-mentioned requirement, a packet's header
  chain length cannot exceed the Path MTU associated with its
  destination.  Hosts discover the Path MTU using procedures such as
  those defined in [RFC1981] and [RFC4821].  Hosts that do not discover
  the Path MTU MUST limit the IPv6 Header Chain length to 1280 bytes.
  Limiting the IPv6 Header Chain length to 1280 bytes ensures that the
  header chain length does not exceed the IPv6 minimum MTU [RFC2460].

6.  IANA Considerations

  IANA has added the following "Type 4 - Parameter Problem" message to
  the "Internet Control Message Protocol version 6 (ICMPv6) Parameters"
  registry:

     CODE     NAME/DESCRIPTION
      3       IPv6 First Fragment has incomplete IPv6 Header Chain





Gont, et al.                 Standards Track                    [Page 5]

RFC 7112         Implications of Oversized Header Chains    January 2014


7.  Security Considerations

  No new security exposures or issues are raised by this document.
  This document describes how undesirably fragmented packets can be
  leveraged to evade stateless packet filtering.  Having made that
  observation, this document updates [RFC2460] so that undesirably
  fragmented packets are forbidden.  Therefore, a security
  vulnerability is removed.

  This specification allows nodes that drop the aforementioned packets
  to signal such packet drops with ICMPv6 "Parameter Problem, IPv6
  First Fragment has incomplete IPv6 header chain" (Type 4, Code 3)
  error messages.

  As with all ICMPv6 error/diagnostic messages, deploying Source
  Address Forgery Prevention filters helps reduce the chances of an
  attacker successfully performing a reflection attack by sending
  forged illegal packets with the victim's/target's IPv6 address as the
  IPv6 source address of the illegal packet [RFC2827] [RFC3704].

  A firewall that performs stateless deep packet inspection (i.e.,
  examines application payload content) might still be unable to
  correctly process fragmented packets, even if the IPv6 Header Chain
  is not fragmented.

8.  Acknowledgements

  The authors would like to thank Ran Atkinson for contributing text
  and ideas that were incorporated into this document.

  The authors would like to thank (in alphabetical order) Ran Atkinson,
  Fred Baker, Stewart Bryant, Brian Carpenter, Benoit Claise, Dominik
  Elsbroek, Wes George, Mike Heard, Bill Jouris, Suresh Krishnan, Dave
  Thaler, Ole Troan, Eric Vyncke, and Peter Yee, for providing valuable
  comments on earlier versions of this document.
















Gont, et al.                 Standards Track                    [Page 6]

RFC 7112         Implications of Oversized Header Chains    January 2014


9.  References

9.1.  Normative References

  [RFC1981]  McCann, J., Deering, S., and J. Mogul, "Path MTU Discovery
             for IP version 6", RFC 1981, August 1996.

  [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
             Requirement Levels", BCP 14, RFC 2119, March 1997.

  [RFC2460]  Deering, S. and R. Hinden, "Internet Protocol, Version 6
             (IPv6) Specification", RFC 2460, December 1998.

  [RFC4443]  Conta, A., Deering, S., and M. Gupta, "Internet Control
             Message Protocol (ICMPv6) for the Internet Protocol
             Version 6 (IPv6) Specification", RFC 4443, March 2006.

  [RFC4821]  Mathis, M. and J. Heffner, "Packetization Layer Path MTU
             Discovery", RFC 4821, March 2007.

  [RFC7045]  Carpenter, B. and S. Jiang, "Transmission and Processing
             of IPv6 Extension Headers", RFC 7045, December 2013.

9.2.  Informative References

  [RFC2827]  Ferguson, P. and D. Senie, "Network Ingress Filtering:
             Defeating Denial of Service Attacks which employ IP Source
             Address Spoofing", BCP 38, RFC 2827, May 2000.

  [RFC3704]  Baker, F. and P. Savola, "Ingress Filtering for Multihomed
             Networks", BCP 84, RFC 3704, March 2004.

  [IANA-PROTO]
             Internet Assigned Numbers Authority, "Protocol Numbers",
             <http://www.iana.org/assignments/protocol-numbers>.
















Gont, et al.                 Standards Track                    [Page 7]

RFC 7112         Implications of Oversized Header Chains    January 2014


Authors' Addresses

  Fernando Gont
  Huawei Technologies
  Evaristo Carriego 2644
  Haedo, Provincia de Buenos Aires  1706
  Argentina

  Phone: +54 11 4650 8472
  EMail: [email protected]


  Vishwas Manral
  Ionos Networks
  Sunnyvale, CA  94089
  US

  Phone: 408-447-1497
  EMail: [email protected]


  Ronald P. Bonica
  Juniper Networks
  2251 Corporate Park Drive
  Herndon, VA  20171
  US

  Phone: 571 250 5819
  EMail: [email protected]






















Gont, et al.                 Standards Track                    [Page 8]