NetBSD/sys/netinet6/IMPLEMENTATION

$NetBSD: IMPLEMENTATION,v 1.2 1999/07/03 21:30:17 thorpej Exp $

# NOTE: this is from original KAME distribution.
# Some portion of this document is not applicable to the code merged into
# NetBSD-current (especially tcp part).  Check sys/netinet6/TODO as well.

			Implementation Note

			KAME Project
			http://www.kame.net/
			$Date: 1999/07/03 21:30:17 $

1. IPv6

1.1 Conformance

The KAME kit conforms, or tries to conform, to the latest set of IPv6
specifications.  For future reference we list some of the relevant documents
below (NOTE: this is not a complete list - this is too hard to maintain...).
For details please refer to specific chapter in the document, RFCs, manpages
come with KAME, or comments in the source code.

Conformance tests have been performed on the previous KAME STABLE kit at
TAHI project.
Results can be viewed at http://www.tahi.org/report/KAME/freebsd-stable-9903/.
We also attended Univ. of New Hampshire IOL tests (http://www.iol.unh.edu/)
in the past, with our past snapshots.

RFC1639: FTP Operation Over Big Address Records (FOOBAR)
    * RFC2428 is preferred over RFC1639.  ftp clients will first try RFC2428,
      then RFC1639 if failed.
RFC1933: Transition Mechanisms for IPv6 Hosts and Routers
    * IPv4 compatible address is not supported.
    * automatic tunnelling (4.3) is not supported.
    * "gif" interface implements IPv[46]-over-IPv[46] tunnel in a generic way,
      and it covers "configured tunnel" described in the spec.
      See 1.5 in this document for details.
RFC1981: Path MTU Discovery for IPv6
RFC2080: RIPng for IPv6
    * KAME-supplied route6d, bgpd and hroute6d support this.
RFC2283: Multiprotocol Extensions for BGP-4
    * so-called "BGP4+".
    * KAME-supplied bgpd supports this.
RFC2292: Advanced Sockets API for IPv6
    * For supported library functions/kernel APIs, see sys/netinet6/ADVAPI
      (KAME/FreeBSD228 only)
RFC2362: Protocol Independent Multicast-Sparse Mode (PIM-SM)
    * RFC2362 defines packet formats for PIM-SM.  draft-ietf-pim-ipv6-01.txt
      is written based on this.
RFC2373: IPv6 Addressing Architecture
    * KAME supports node required addresses, and conforms to the scope
      requirement.
RFC2374: An IPv6 Aggregatable Global Unicast Address Format
    * KAME supports 64-bit length of Interface ID.
RFC2375: IPv6 Multicast Address Assignments
    * Userland applications use the well-known addresses assigned in the RFC.
RFC2428: FTP Extensions for IPv6 and NATs
    * RFC2428 is preferred over RFC1639.  ftp clients will first try RFC2428,
      then RFC1639 if failed.
RFC2460: IPv6 specification
RFC2461: Neighbor discovery for IPv6
    * See 1.2 in this document for details.
RFC2462: IPv6 Stateless Address Autoconfiguration
    * See 1.4 in this document for details.
RFC2463: ICMPv6 for IPv6 specification
    * See 1.8 in this document for details.
RFC2464: Transmission of IPv6 Packets over Ethernet Networks
RFC2467: Transmission of IPv6 Packets over FDDI Networks
RFC2472: IPv6 over PPP
RFC2492: IPv6 over ATM Networks
    * only PVC is supported.
RFC2545: Use of BGP-4 Multiprotocol Extensions for IPv6 Inter-Domain Routing
RFC2553: Basic Socket Interface Extensions for IPv6
    * IPv4 mapped address (3.7) and special behavior of IPv6 wildcard bind
      socket (3.8) are,
	- supported on KAME/FreeBSD31 and it seems to be working well.
	- also supported on KAME/FreeBSD228, but it is not tested well
	  so disabled by default.
	- not supported on KAME/NetBSD and KAME/BSDI.
      see 1.12 in this document for details.
draft-ietf-ipngwg-ipv6router-alert-04: IPv6 router alert option
draft-ietf-ipngwg-router-renum-08: Router renumbering for IPv6
draft-ietf-ipngwg-icmp-namelookups-02: IPv6 Name Lookups Through ICMP
draft-ietf-ipngwg-icmp-name-lookups-03: IPv6 Name Lookups Through ICMP
draft-ietf-ipngwg-aaaa-03: DNS Extensions to support IPv6
draft-ietf-ipngwg-mld-01: Multicast Listener Discovery for IPv6
draft-ietf-ipngwg-jumbo-00: The IPv6 Jumbo Payload Option
    * See 1.7 in this document for details.
draft-ietf-ipngwg-jumbograms-00: IPv6 Jumbograms
    * See 1.7 in this document for details.
draft-ietf-pim-ipv6-01.txt: PIM for IPv6
    * no sparse mode support.  pim6dd implements dense mode only.
draft-itojun-ipv6-tcp-to-anycast-00:
	Disconnecting TCP connection toward IPv6 anycast address
draft-yamamoto-wideipv6-comm-model-00
    * See 1.6 in this document for details.

1.2 Neighbor Discovery

Neighbor Discovery is fairly stable.  Currently Address Resolution,
Duplicated Address Detection, and Neighbor Unreachability Detection
are supported.  In the near future we will be adding Proxy Neighbor
Advertisement support in the kernel and Unsolicited Neighbor Advertisement
transmission command as admin tool.

If DAD fails, the address will be marked "duplicated" and message will be
generated to syslog (and usually to console).  The "duplicated" mark
can be cheked with ifconfig.  It is administrators' responsibility to check
for and recover from DAD failures.
The behavior should be improved in the near future.

Some of the network driver loops multicast packets back to itself,
even if instructed not to do so (especially in promiscuous mode).
In such cases DAD may fail, because DAD engine sees inbound NS packet
(actually from the node itself) and considers it as a sign of duplicate.

Neighbor Discovery specification (RFC2461) does not talk about neighbor
cache handling in the following cases:
(1) there was no neighbor cache entry, and unsolicited RS/NS/NA/redirect
    packet, without link-layer address was received
(2) neighbor cache handling on medium without link-layer address
    (we need a neighbor cache entry for IsRouter bit)
For (1), we implemented workaround based on discussions on IETF ipngwg mailing
list.  For more details, see the comments in the source code and email
thread started from (IPng 7155), dated Feb 6 1999.

IPv6 adderss autoconfiguration assumes that a host has only single network
interface (single non-loopback interface).  This is because the spec assumes
that there's no routing table on IPv6 host (an IPv6 host has default router
list and prefix list only).
Because of this, if you accept RAs on a node with multiple network interfaces,
the node will not behave as you might expect.  RFC2461 Appendix A talks
little bit about this issue.

IPv6 on-link determination rule (RFC2461) is quite different from assumptions
in BSD network code.  At this moment, KAME does not implement on-link
determination rule when default router list is empty (RFC2461, section 5.2,
last sentence in 2nd paragraph - note that the spec misuse the word "host"
and "node" in several places in the section).

To avoid possible DoS attacks and infinite loops, KAME stack will accept
only 10 options on ND packet.  Therefore, if you have 20 prefix options
attached to RA, only the first 10 prefixes will be recognized.
If this troubles you, please contact KAME team and/or modify
sys/netinet6/nd6.c:nd6_options().

1.3 Scope Index

IPv6 uses scoped addresses.  Therefore, it is very important to
specify scope index (interface index for link-local address, or
site index for site-local address) with an IPv6 address.  Without
scope index, scoped IPv6 address is ambiguous to the kernel, and
kernel will not be able to determine the outbound interface for a
packet.

Ordinary userland applications should use advanced API (RFC2292) to
specify scope index, or interface index.  For similar purpose,
sin6_scope_id member in sockaddr_in6 structure is defined in RFC2553.
However, the semantics for sin6_scope_id is rather vague.  If you
care about portability of your application, we suggest you to use
advanced API rather than sin6_scope_id.

In the kernel, an interface index for link-local scoped address is
embedded into in6_addr.s6_addr16[1].  For example, you may see
something like:
	fe80:1::200:f8ff:fe01:6317
in the routing table and interface address structure (struct
in6_ifaddr).  The address above is a link-local unicast address
which belongs to a network interface whose interface identifier is
1.  The embedded index enables us to identify IPv6 link local
addresses over multiple interfaces effectively and with only a
little code change.
Routing daemons and configuration programs, like route6d and
ifconfig, will need to manipulate the "embedded" scope index.
These programs use routing sockets and ioctls (like SIOCGIFADDR_IN6)
and the kernel API will return IPv6 addresses with in6_addr.s6_addr16[1]
filled in.  The APIs are for manipulating kernel internal structure.
Programs that use these APIs have to be prepared about differences
in kernels anyway.

When you specify scoped address to the command line, NEVER write the
embedded form (such as ff02:1::1 or fe80:2::fedc).  This is not supposed
to work.  Always use standard form, like ff02::1 or fe80::fedc, with
command line option for specifying interface (like "ping6 -I ne0 ff02::1).
In general, if a command does not have command line option to specify
outgoing interface, that command is not ready to accept scoped address.
This may seem to be opposite from IPv6's premise to support "dentist office"
situation.  We believe that specifications need some improvements for this.

1.4 Plug and Play

The KAME kit implements most of the IPv6 stateless address
autoconfiguration in the kernel.
Neighbor Discovery functions are implemented in the kernel as a whole.
Router Advertisement (RA) input for hosts is implemented in the
kernel.  Router Solicitation (RS) output for endhosts, RS input
for routers, and RA output for routers are implemented in the
userland.

When the kernel boots up, an IPv6 link-local address is assigned to each
interface.  Also, direct route for the link-local address is added to
routing table. Here is an output of netstat command:

Internet6:
Destination                   Gateway                   Flags      Netif Expire
fe80:1::/64                   link#1                    UC           ed0
fe80:2::/64                   link#2                    UC           ep0

Each interface joins the solicited multicast address and the
link-local all-nodes multicast addresses (e.g.  fe80:1::1:ff01:6317
and ff02:1::1, respectively).  In addition to a link-local address,
the loopback address (::1) is assigned to the loopback interface.  It
adds ::1/128 and ff01::/32 to routing table and joins ff01::1.

When a host hears Router Advertisement from the router, default route
and network address prefix (usually global address prefix) are added.
To generate a Router Solicitation packet at any time, use the "rtsol" command.
Also, "rtsold" daemon is available.  It generates Router Solicitation
whenever necessary, and it works great for nomadic usage (notebooks/laptops).
If one wishes to ignore Router Advertisements, use sysctl to set
net.inet6.ip6.accept_rtadv to 0.

To generate Router Advertisement from a router, use the "rtadvd" daemon.

RFC2462 has validation rule against incoming RA prefix information option,
in 5.5.3 (e).  This is to protect hosts from malicious (or misconfigured)
routers that advertise very short prefix lifetime.
There was an update from Jim Bound to ipngwg mailing list (look
for "(ipng 6712)" in the archive) and KAME implements Jim's update.

See 1.2 in the document for relationship between DAD and autoconfiguration.

DHCPv6 server/client is not implemented yet.  "Managed" and "Other" bits in
RA have no special effect to stateful autoconfiguration procedure
("Managed" bit actually prevents stateless autoconfiguration, but no special
action will be taken for DHCPv6 client).

1.5 Generic tunnel interface

GIF (Generic InterFace) is a pseudo interface for configured tunnel.
Details are described in gif(4) manpage.
Currently
	v6 in v6
	v6 in v4
	v4 in v6
	v4 in v4
are available.  Use "gifconfig" to assign physical (outer) source
and destination address to gif interfaces.
Configuration that uses same address family for inner and outer IP
header (v4 in v4, or v6 in v6) is dangerous.  It is very easy to
configure interfaces and routing tables to perform infinite level
of tunneling.  Please be warned.

gif can be configured to be ECN-friendly.  See 4.5 for ECN-friendliness
of tunnels, and gif(4) manpage for how to configure.

1.6 Source Address Selection

Source selection of KAME is scope oriented (there are some exceptions -
see below).  For a given destination, a source IPv6 address is selected
by the following rule:
    1. If the source address is explicitly specified by the user
       (e.g. via the advanced API), the specified address is used.
    2. If there is an address assigned to the outgoing interface
       (which is usually determined by looking up the routing table)
       that has the same scope as the destination address, the address
       is used.
       This is the most typical case.
    3. If there is no address that satisfies the above condition,
       choose a global address assigned to one of the interfaces
       on the sending node.
    4. If there is no address that satisfies the above condition and
       there is no global address on the sending node, choose the
       address associated with the routing table entry for the destination.
       This is the last resort, which may cause scope violation.

For instance, ::1 is selected for ff01::1, fe80:1::200:f8ff:fe01:6317
for fe80:1::2a0:24ff:feab:839b.  If the outgoing interface has
multiple address for the scope, a source is selected longest match
basis (rule 3).  Suppose 3ffe:501:808:1:200:f8ff:fe01:6317 and
3ffe:2001:9:124:200:f8ff:fe01:6317 are given to the outgoing
interface.  3ffe:501:808:1:200:f8ff:fe01:6317 is chosen as the source
for the destination 3ffe:501:800::1.

Note that the above rule is not documented in the IPv6 spec.  It is
considered "up to implementation" item.
There are some cases where we do not use the above rule.  One
example is connected TCP session, and we use the address kept in tcb
as the source.
Another example is source address for Neighbor Advertisement.
Under the spec (RFC2461 7.2.2) NA's source should be the target
address of the corresponding NS's target.  In this case we follow
the spec rather than the above longest-match rule.

1.7 Jumbo Payload

KAME supports the Jumbo Payload hop-by-hop option used to send IPv6
packets with payloads longer than 65,535 octets.  But since currently
KAME does not support any physical interface whose MTU is more than
65,535, such payloads can be seen only on the loopback interface(i.e.
lo0).

If you want to try jumbo payloads, you first have to reconfigure the
kernel so that the MTU of the loopback interface is more than 65,535
bytes; add the following to the kernel configuration file:
	options		"LARGE_LOMTU"		#To test jumbo payload
and recompile the new kernel.

Then you can test jumbo payloads by the ping6 command with -b and -s
options.  The -b option must be specified to enlarge the size of the
socket buffer and the -s option specifies the length of the packet,
which should be more than 65,535.  For example, type as follows; 
	% ping6 -b 70000 -s 68000 ::1

The IPv6 specification requires that the Jumbo Payload option must not
be used in a packet that carries a fragment header.  If this condition
is broken, an ICMPv6 Parameter Problem message must be sent to the
sender.  KAME kernel follows the specification, but you cannot usually
see an ICMPv6 error caused by this requirement.

If KAME kernel receives an IPv6 packet, it checks the frame length of
the packet and compares it to the length specified in the payload
length field of the IPv6 header or in the value of the Jumbo Payload
option, if any.  If the former is shorter than the latter, KAME kernel
discards the packet and increments the statistics. You can see the
statistics as output of netstat command with `-s -p ip6' option:
	% netstat -s -p ip6
	ip6:
		(snip)
		1 with data size < data length

So, KAME kernel does not send an ICMPv6 error unless the erroneous
packet is an actual Jumbo Payload, that is, its packet size is more
than 65,535 bytes.  As described above, KAME kernel currently does not
support physical interface with such a huge MTU, so it rarely returns an
ICMPv6 error.

TCP/UDP over jumbogram is not supported at this moment.  This is because
we have no medium (other than loopback) to test this.  Contact us if you
need this.

IPsec does not work on jumbograms.  This is due to some specification twists
in supporting AH with jumbograms.

1.8 Loop prevention in header processing

IPv6 specification allows arbitrary number of extension headers to
be placed onto packets.  If we implement IPv6 packet processing
code in the way BSD IPv4 code is implemented, kernel stack may
overflow due to deep function call chain.  KAME sys/netinet6 code
is carefully designed to avoid kernel stack overflow.  Because of
this, KAME sys/netinet6 code defines its own protocol switch
structure, as "struct ip6protosw" (see netinet6/ip6protosw.h).
IPv4 part (sys/netinet) remains untouched for compatibility.
Because of this, if you receive IPsec-over-IPv4 packet with massive
number of IPsec headers, kernel stack may blow up.  IPsec-over-IPv6 is okay.

1.9 ICMPv6

After RFC2463 was published, IETF ipngwg has decided to disallow ICMPv6 error
packet against ICMPv6 redirect, to prevent ICMPv6 storm on a network medium.
KAME already implements this into the kernel.

1.10 Applications

For userland programming, we support IPv6 socket API as specified in
RFC2553, RFC2292 and upcoming internet drafts.

TCP/UDP over IPv6 is available and quite stable.  You can enjoy "telnet",
"ftp", "rlogin", "rsh", "ssh", etc.  These applications are protocol
independent.  That is, they automatically chooses IPv4 or IPv6
according to DNS.

1.11 Kernel Internals

 (*) TCP stack is merged between netinet and netinet6 on KAME/FreeBSD31,
     so if your platform is FreeBSD31, then there is no difference
     between tcp6 and tcp in following description.

The current KAME has escaped from the IPv4 netinet logic.  While
ip_forward() calls ip_output(), ip6_forward() directly calls
if_output() since routers must not divide IPv6 packets into fragments.

ICMPv6 should contain the original packet as long as possible up to
1280.  UDP6/IP6 port unreach, for instance, should contain all
extension headers and the *unchanged* UDP6 and IP6 headers.
So, all IP6 functions except TCP6 never convert network byte
order into host byte order, to save the original packet.

tcp6_input(), udp6_input() and icmp6_input() can't assume that IP6
header is preceding the transport headers due to extension
headers.  So, in6_cksum() was implemented to handle packets whose IP6
header and transport header is not continuous.  TCP/IP6 nor UDP/IP6
header structure don't exist for checksum calculation.

To process IP6 header, extension headers and transport headers easily,
KAME requires network drivers to store packets in one internal mbuf or
one or more external mbufs.  A typical old driver prepares two
internal mbufs for 100 - 208 bytes data, however, KAME's reference
implementation stores it in one external mbuf.

"netstat -s -p ip6" tells you whether or not your driver conforms
KAME's requirement.  In the following example, "cce0" violates the
requirement. (For more information, refer to Section 2.)

        Mbuf statistics:
                317 one mbuf
                two or more mbuf::
                        lo0 = 8
			cce0 = 10
                3282 one ext mbuf
                0 two or more ext mbuf

Each input function calls IP6_EXTHDR_CHECK in the beginning to check
if the region between IP6 and its header is
continuous.  IP6_EXTHDR_CHECK calls m_pullup() only if the mbuf has
M_LOOP flag, that is, the packet comes from the loopback
interface.  m_pullup() is never called for packets coming from physical
network interfaces.

TCP6 reassembly makes use of IP6 header to store reassemble
information.  IP6 is not supposed to be just before TCP6, so
ip6tcpreass structure has a pointer to TCP6 header.  Of course, it has
also a pointer back to mbuf to avoid m_pullup().

Like TCP6, both IP and IP6 reassemble functions never call m_pullup().

xxx_ctlinput() calls in_mrejoin() on PRC_IFNEWADDR.  We think this is
one of 4.4BSD implementation flaws.  Since 4.4BSD keeps ia_multiaddrs
in in_ifaddr{}, it can't use multicast feature if the interface has no
unicast address.  So, if an application joins to an interface and then
all unicast addresses are removed from the interface, the application
can't send/receive any multicast packets.  Moreover, if a new unicast
address is assigned to the interface, in_mrejoin() must be called.
KAME's interfaces, however, have ALWAYS one link-local unicast
address.  These extensions have thus not been implemented in KAME.

1.12 IPv4 mapped address and IPv6 wildcard socket

RFC2553 describes IPv4 mapped address (3.7) and special behavior
of IPv6 wildcard bind socket (3.8).  The spec allows you to:
- Transmit IPv4 packet over AF_INET6 socket by using special form of
  the address like ::ffff:10.1.1.1.
- Accept IPv4 connections by AF_INET6 wildcard bind socket.
but it may look complicated spec.

We KAME team have 4 OS platforms right now, and behavior is silghtly
different between them.  To summarize:
- All KAME implementations treat tcp/udp port number space separately
  between IPv4 and IPv6.
- KAME/NetBSD and KAME/BSDI does not support IPv4 mapped address, nor
  special wildcard bind on AF_INET6.
- KAME/FreeBSD31 supports IPv4 mapped address, and special wildcard bind on
  AF_INET6.  They seem to be working well and enabled by default.
  You can disable those two by runtime and kernel compile configuration.
  (you can't enable only one of them: they comes together)
- KAME/FreeBSD228 supports both, but it is not tested well so disabled by
  default.
  You can enable those two by runtime and kernel compile configuration.
  (you can't enable only one of them: they comes together)
- KAME/FreeBSD228 and KAME/FreeBSD31 implements slightly different behavior
  for special wildcard bind on AF_INET6.

The following sections will give you the details, and how you can
configure the behavior on KAME/FreeBSD{228,31}.

Advise to application implementers: to implement a portable IPv6 application
(which works on multiple IPv6 kernels), we believe that the following
is the key to the success:
- NEVER hardcode AF_INET or AF_INET6.
- Use getaddrinfo() and getnameinfo() throughout the system.
  Never use gethostby*(), getaddrby*(), inet_*() or getipnodeby*().
- If you would like to listen to connections, use getaddrinfo() (maybe
  with AI_PASSIVE), and make sockets for all the "struct addrinfo" returned.
- If you would like to connect to destination, use getaddrinfo() and try
  all the destination returned, like telnet does.
- Some of the IPv6 stack is shipped with buggy getaddrinfo().  Ship a minimal
  working version with your application and use that as last resort.
It looks that RFC2553 talks too little on wildcard bind issue,
especially on the port space issue, failure mode and relationship
between AF_INET/INET6 wildcard bind.  There can be several separate
interpretation for this RFC (see 1.12.2 - we have two different
implementation for this, and RFC2553 seems to fit to both of them).
So, to implement portable application you should assume nothing
about the behavior in the kernel.  Using getaddrinfo() is the safest way.
Port number space and wildcard bind issues were discussed in detail
on ipv6imp mailing list, in mid March 1999 and it looks that there's
no concrete consensus (means, up to implementers).  You may want to
check the mailing list archives.

1.12.1 KAME/NetBSD and KAME/BSDI

The platforms do not support IPv4 mapped address.
The IPv4 mapped address support needs tweaked implementation in
DNS support libraries, as documented in RFC2553 6.1.  However, since
the platforms do not support this, you do not need to worry about
RFC2553 6.1 and story goes much simpler.
(KAME library actually implements the tweaks, but it is safe to ignore that)

Port number space is totally separate between AF_INET and
AF_INET6 sockets.  You can always perform wildcard bind on both of
the adderss families, on the same port.

If a server application would like to accept IPv4 and IPv6 connections,
it should use AF_INET and AF_INET6 socket (you'll need two sockets).
Applicsations should use proper socket for connections.  IPv4 connections
must be made on AF_INET socket, and IPv6 connections must be made
on AF_INET6 socket.  getaddrinfo() library helps you in writing
AF-independent application, and managing sockets with different AFs.
(some of the implementers think that we should totaly get rid of gethostby*
family of the functions and migrate to get{addr,name}info, since
it is very clean and helps you support new AFs in the future)

1.12.2 KAME/FreeBSD31

The platform can be configued to support IPv4 mapped address/special AF_INET6
wildcard bind (enabled by default).  If you disable it, it behaves as described
in 1.12.1.

The IPv4 mapped address support needs tweaked implementation in
DNS support libraries.  This is documented in RFC2553 6.1.
KAME libraries (namely libinet6.a) actually support that.

RFC2553 does not talk about how port number space should be designed
(i.e. should they be separate between AF_INET and AF_INET6, or
should they be common)
In KAME with the behavior enabled, port number space is separate
between AF_INET and AF_INET6 sockets in most cases.  The only
exception is wildcard bind socket, where the special behavior appears.

If a server application would like to accept IPv4 and IPv6 connections,
it can use IPv6 socket with wildcard bind, or use two sockets (for
AF_INET6 and AF_INET).  You can handle IPv4 and IPv6 connections
by using AF_INET6 socket.  Porting of an application can be simpler in
this case, like:
- change AF_INET into AF_INET6
- use gethostbyname2(hostname, AF_INET6) or getipnodebyname(), instead of
  gethostbyname(hostname)
- use struct sockaddr_in6 instead of sockaddr_in

To provide services to both IPv4 and IPv6 clients, you will run a single
server which binds to single AF_INET6 wildcard socket.  This server will
accept both IPv4 and IPv6 connections to the tcp/udp port.
If you run two daemons, which binds to AF_INET6 wildcard socket and
AF_INET socket (say, sendmail4 and sendmail6), story start to look
a bit complicated.  The next sections have the detail.

(* the following paragraphs are different between KAME/FreeBSD31 and 228)

Wildcard bind on AF_INET6 behaves like "wildcard bind between two
address families".  It will grab IPv4 connection if and only if
there is no socket that binds to more specific destination.
Here, wildcard bind on AF_INET is regarded as "more specific bind"
than wildcard bind on AF_INET6.
In other words, wildcard bind on AF_INET6 is the only thing that
has special behavior.  It will not affect wildcard bind on AF_INET.
If the folllowing events happen, IPv4 connection will be routed
to application B, and IPv6 connection will be routed to application A.
- application A perform wildcard bind on AF_INET6, port X
- application B perform wildcard bind on AF_INET, port X
- IPv4 connection arrives
- IPv6 connection arrives
If the following events happen, the behavior is the same.  IPv4
connection will be routed to application B, and IPv6 connection
will be routed to application A.
- application B perform wildcard bind on AF_INET, port X
- application A perform wildcard bind on AF_INET6, port X
- IPv4 connection arrives
- IPv6 connection arrives

If the following events happen, KAME/FreeBSD31 will behave like this:
- sendmail4 is running.  It is doing wildcard bind on AF_INET.
- Invoke sendmail6.  It will do a wildcard bind on AF_INET6.
- Stop sendmail4.  Here, on KAME/FreeBSD31, IPv4 and IPv6 conections will
  be routed to sendmail6.  This is different from KAME/FreeBSD228.

1.12.3 KAME/FreeBSD228

The platform can be configued to support IPv4 mapped address/special AF_INET6
wildcard bind (disabled by default).  If you disable it, it behaves as
described in 1.12.1.

KAME/FreeBSD228 works mostly the same as KAME/FreeBSD31, except the paragraphs
under (*) mark in 1.12.2.  Please refer to 1.12.2 for other parts.

	Wildcard bind on AF_INET6 affects the behavior of wildcard bind on
	AF_INET.
	If events happen in the following order, application B fails to
	bind.  Therefore, both IPv4 and IPv6 connections will be routed to
	application A.
	- application A perform wildcard bind on AF_INET6, port X
	- application B perform wildcard bind on AF_INET, port X -> fail
	- IPv4 connection arrives
	- IPv6 connection arrives
	If the following events happen, IPv4 connection will be routed to
	application B, and IPv6 connection will be routed to application
	A.  This behavior is the same as described in 1.12.1.
	- application B perform wildcard bind on AF_INET, port X
	- application A perform wildcard bind on AF_INET6, port X
	- IPv4 connection arrives
	- IPv6 connection arrives

	If the following events happen, KAME/FreeBSD228 will behave like this:
	- sendmail4 is running.  It is doing wildcard bind on AF_INET.
	- Invoke sendmail6.  It will do a wildcard bind on AF_INET6.
	- Stop sendmail4.  Here, on KAME/FreeBSD228, sendmail6 will be able to
	  get IPv6 connections only.  This is different from KAME/FreeBSD31.

1.12.4 configuration and implementation

On KAME/FreeBSD31 and KAME/FreeBSD228, the behavior is 
configurable by following procedure.  To enable it:
- Add the "MAPPED_ADDR_ENABLED" kernel config option into your
  kernel config file (see "sys/i386/conf/GENERIC.v6" sample file)
  and build your kernel, and
- set sysctl variable appropriately, like:
	# sysctl -w net.inet6.ip6.mapped_addr=1
Note that, to enable the behavior you'll need to do the both of the above.
If you do not do the both, the behavior is disabled.

The behavior is enabled by default on KAME/FreeBSD31, but disabled
by default on KAME/FreeBSD228, because those functions are not well
tested on KAME/FreeBSD228.
Implementation is quite different between KAME/FreeBSD31 and
KAME/FreeBSD228, because tcp stack code and protocol contorol block
(pcb) implementation are merged between netinet and netinet6 in
KAME/FreeBSD31 code, but they are separate in KAME/FreeBSD228 code.
This is the reason for different behavior in two OSes.

2. Network Drivers

KAME requires three items to be added into the standard drivers:

(1) mbuf clustering requirement. In this stable release, we changed
    MINCLSIZE into MHLEN+1 for all the operating systems in order to make
    all the drivers behave as we expect.  

(2) multicast.  If "ifmcstat" yields no multicast group for a
    interface, that interface has to be patched.

(3) If you are using notebook PCs, you'll need to be a one-liner for
    initializing your interface when your card is plugged in.
    (NOTE: KAME/NetBSD does not need this)

To avoid troubles, we suggest you to comment out the device drivers
for unsupported/unnecessary cards, from the configuration file.
If you accidentally enable unsupported drivers, some of the userland
tools may not work correctly (routing daemons are typical example).

In the following sections, "official support" means that KAME developers
are using that ethernet card/driver frequently.

2.1 FreeBSD 2.2.x-RELEASE

Here is a list of FreeBSD 2.2.x-RELEASE drivers and its conditions:

	driver	mbuf(1)		multicast(2)	PCMCIA		official
						patch(3)	support?
	---	---		---		---		---
	(Ethernet)
	ar	looks ok	-		-		-
	cnw	ok		ok		ok		yes (*)
	ed	ok		ok		ok		yes
	ep	ok		ok		ok		yes
	fe	ok		ok		ok		yes
	sn	looks ok	-		ok		-   (*)
	vx	looks ok	-		-		-
	wlp	ok		ok		ok		-   (*)
	xl	ok		ok		-		yes
	zp	ok		ok		-		-
	(FDDI)
	fpa	looks ok	?		-		-
	(ATM)
	en	ok		ok		-		yes
	(Serial)
	lp	?		-		-		not work
	sl	?		-		-		not work
	sr	looks ok	ok		-		-   (**)

You may want to add an invocation of "rtsol" in "/etc/pccard_ether",
if you are using notebook computers and PCMCIA ethernet card.

(*) These drivers are distributed with PAO (http://www.jp.freebsd.org/PAO/).

(**) There was some report says that, if you make sr driver up and down and
then up, the kernel may hang up.  We have disabled frame-relay support from
sr driver and after that this looks to be working fine.  If you need
frame-relay support to come back, please contact KAME deverlopers.

2.2 BSD/OS

The following lists BSD/OS device drivers and its conditions:

	driver	mbuf(1)		multicast(2)	PCMCIA		official
						patch(3)	support?
	---	---		---		---		---
	(Ethernet)
	cnw	ok		ok		ok		yes
	de	ok		ok		no need		-
	ef	ok		ok		ok		yes
	exp	ok		ok		no need		-
	mz	ok		ok		ok		yes
	ne	ok		ok		ok		yes
	we	ok		ok		ok		-
	(FDDI)
	fpa	ok		ok		-		-
	(ATM)
	en	maybe		ok		-		-
	(Serial)
	ntwo	ok		ok		-		yes
	sl	?		-		-		not work
	appp	?		-		-		not work

You may want to use "@insert" directive in /etc/pccard.conf to invoke
"rtsol" command right after dynamic insertion of PCMCIA ethernet cards.

2.3 NetBSD

The following table lists the network drivers we have tried so far.
Note that, for NetBSD, we do not need one-liner patch to PCMCIA driver.

	driver		mbuf(1)	multicast(2)	official
						support?
	---		---	---		---
	(Ethernet)
	ne pci/i386	ok	ok		yes
	ep pcmcia/i386	ok	ok		-
	le sbus/sparc	ok	ok		yes
	(ATM)
	en pci/i386	ok	ok		-

2.4 FreeBSD 3.x-RELEASE

Here is a list of FreeBSD 3.x-RELEASE drivers and its conditions:

	driver	mbuf(1)		multicast(2)	PCMCIA		official
						patch(3)	support?
	---	---		---		---
	(Ethernet)
	fe	ok		ok		?		yes
	fxp	ok		ok		-		yes
	lnc	?		ok		-		-
	cnw	ok		ok		ok(*)		-(**)
	ep	ok		ok		ok(*)		-
	sn	?		?		?		-(**)

(*)  With PAO3 patch applied. (http://www.jp.freebsd.org/PAO/).
(**) These drivers are distributed with PAO as PAO3
     (http://www.jp.freebsd.org/PAO/).

More drivers will just simply work on KAME FreeBSD 3.x-RELEASE but have not
been checked yet.

3. Translator

We categorize IPv4/IPv6 translator into 4 types.

Translator A --- It is used in the early stage of transition to make
it possible to establish a connection from an IPv6 host in an IPv6
island to an IPv4 host in the IPv4 ocean.

Translator B --- It is used in the early stage of transition to make
it possible to establish a connection from an IPv4 host in the IPv4
ocean to an IPv6 host in an IPv6 island.

Translator C --- It is used in the late stage of transition to make it
possible to establish a connection from an IPv4 host in an IPv4 island
to an IPv6 host in the IPv6 ocean.

Translator D --- It is used in the late stage of transition to make it
possible to establish a connection from an IPv6 host in the IPv6 ocean
to an IPv4 host in an IPv4 island.

KAME provides an TCP relay translator for category A.  This is called
"FAITH".  We also provide IP header translator for category A.

3.1 FAITH TCP relay translator

FAITH system uses TCP relay daemon called "faithd" helped by the KAME kernel.
FAITH will reserve an IPv6 address prefix, and relay TCP connection
toward that prefix to IPv4 destination.

For example, if the reserved IPv6 prefix is 3ffe:0501:0200:ffff::, and
the IPv6 destination for TCP connection is 3ffe:0501:0200:ffff::163.221.202.12,
the connection will be relayed toward IPv4 destination 163.221.202.12.

	destination IPv4 node (163.221.202.12)
	  ^
	  | IPv4 tcp toward 163.221.202.12
	FAITH-relay dual stack node
	  ^
	  | IPv6 TCP toward 3ffe:0501:0200:ffff::163.221.202.12
	source IPv6 node

faithd must be invoked on FAITH-relay dual stack node.

For more details, consult kit/src/faithd/README.

3.2 IPv6-to-IPv4 header translator

(to be written)

4. IPsec

IPsec is mainly organized by three components.

(1) Policy Management
(2) Key Management
(3) AH and ESP handling

4.1 Policy Management

The policy management code is experimental, but this is almostly
conformed to RFC2401.  You can manage the SPD by both the command
for administrater and the socket operation.
By the command "setkey", there are three directive as following:

        "discard" means to discard the packet.
        "none" means not to do IPsec.
        "ipsec" means to do IPsec.

"ipsec" is followed by some IPsec requests like "protocol/level".

	protocol is to be either "ah" or "esp".
	level is to be either "default", "use" or "require".

	"use" means that if a security association is available, use it
	for outbound packet, request SA to the Key management daemon
	through PF_KEY v2, and accept any inbound packets.
	
	"require" means that if a security association does not exist
	for outbound traffic, acquire one and discard the packet untill
	SA is found, and require inbound packets to use security.

	"default" is consulted to system wide default defined "sysctl" MIBs:
		net.inet.ipsec.esp_trans_deflev
		net.inet.ipsec.esp_net_deflev
		net.inet.ipsec.ah_trans_deflev
		net.inet.ipsec.ah_net_deflev
		net.inet6.ipsec6.esp_trans_deflev
		net.inet6.ipsec6.esp_net_deflev
		net.inet6.ipsec6.ah_trans_deflev
		net.inet6.ipsec6.ah_net_deflev

		They are 1:use or 2:require.

When you need to setup IPsec tunnel mode, you use the format of
"procol/level/peer".  "peer" is IP address of the tunnel end-point.

By the socket operation, there are three directive as following:
        "ipsec" means to do IPsec, see above.
        "entrust" means to consult to SPD defined by the command.
        "bypass" means not to do IPsec. This is for priveleged socket.

If kernel doesn't find out policy entry, then system wide default
is applied.  You can specify the system wide default as discarding
packet or not to do IPsec.

	net.inet.ipsec.def_policy
	net.inet6.ipsec6.def_policy

	They are 0:discard or 1:none.

You can see these values in netinet6/ipsec.h.

The policy entry is not re-ordered with its
indexes, so the order of entry when you add is very significant.
But we think it should be fixed in the future.

4.2 Key Management

The key management code implemented in this kit (sys/netkey) is a
home-brew PFKEY v2 implementation.  This conforms to RFC2367.

The home-brew IKE daemon, "racoon" is included in the kit
(kit/src/racoon).  It can perform key exchanges in some limited
conditions, however, it may take some more time to be stabilized.

4.3 AH and ESP handling

IPsec module is implemented as "hooks" to the standard IPv4/IPv6
processing.  When sending a packet, ip{,6}_output() checks if ESP/AH
processing is required by checking if a matching SPD (Security
Policy Database) is found.  If ESP/AH is needed,
{esp,ah}{4,6}_output() will be called and mbuf will be updated
accordingly.  When a packet is received, {esp,ah}4_input() will be
called based on protocol number, i.e. (*inetsw[proto])().
{esp,ah}4_input() will decrypt/check authenticity of the packet,
and strips off daisy-chained header and padding for ESP/AH.  It is
safe to strip off the ESP/AH header on packet reception, since we
will never use the received packet in "as is" form.

By using ESP/AH, TCP4/6 MSS will be affected by extra daisy-chained
headers inserted by ESP/AH.  Our code takes care of the case.

Basic crypto functions can be found in directory "sys/crypto".  ESP/AH
transform are listed in {esp,ah}_core.c with wrapper functions.  If you
wish to add some algorithm, add wrapper function in {esp,ah}_core.c, and
add your crypto algorithm code into sys/crypto.

Tunnel mode is partially supported in this release, with the following
restrictions:
- IPsec tunnel is not combined with GIF generic tunneling interface.
  It needs a great care because we may create an infinite loop between
  ip_output() and tunnelifp->if_output().  Opinion varies if it is better
  to unify them, or not.
- MTU and Don't Fragment bit (IPv4) considerations need more checking, but
  basically works fine.
- Authentication model for AH tunnel must be revisited.  We'll need to
  improve the policy management engine, eventually.

4.4 Conformance to RFCs and IDs

The IPsec code in the kernel conforms (or, tries to conform) to the
following standards:
    "old IPsec" specification documented in rfc182[5-9].txt
    "new IPsec" specification documented in rfc240[1-6].txt, rfc241[01].txt,
	rfc2451.txt and draft-mcdonald-simple-ipsec-api-01.txt (draft expired,
	but you can take from ftp://ftp.kame.net/pub/internet-drafts/).
	(NOTE: IKE specifications, rfc241[7-9].txt are implemented in userland,
	as "racoon" IKE daemon)

Currently supported algorithms are:
    old IPsec AH
	null crypto checksum (no document, just for debugging)
	keyed MD5 with 128bit crypto checksum (rfc1828.txt)
	keyed SHA1 with 128bit crypto checksum (no document)
	HMAC MD5 with 128bit crypto checksum (rfc2085.txt)
	HMAC SHA1 with 128bit crypto checksum (no document)
    old IPsec ESP
	null encryption (no document, similar to rfc2410.txt)
	DES-CBC mode (rfc1829.txt)
    new IPsec AH
	null crypto checksum (no document, just for debugging)
	keyed MD5 with 96bit crypto checksum (no document)
	keyed SHA1 with 96bit crypto checksum (no document)
	HMAC MD5 with 96bit crypto checksum (rfc2403.txt
	HMAC SHA1 with 96bit crypto checksum (rfc2404.txt)
    new IPsec ESP
	null encryption (rfc2410.txt)
	DES-CBC with derived IV
		(draft-ietf-ipsec-ciph-des-derived-01.txt, draft expired)
	DES-CBC with explicit IV (rfc2405.txt)
	3DES-CBC with explicit IV (rfc2451.txt)
	BLOWFISH CBC (rfc2451.txt)
	CAST128 CBC (rfc2451.txt)
	RC5 CBC (rfc2451.txt)
	each of the above can be combined with:
	    ESP authentication with HMAC-MD5(96bit)
	    ESP authentication with HMAC-SHA1(96bit)

The following algorithms are NOT supported:
    old IPsec AH
	HMAC MD5 with 128bit crypto checksum + 64bit replay prevention
		(rfc2085.txt)
	keyed SHA1 with 160bit crypto checksum + 32bit padding (rfc1852.txt)

4.5 ECN consideration on IPsec tunnels

KAME IPsec implements ECN-friendly IPsec tunnel, described in
draft-ipsec-ecn-00.txt.
Normal IPsec tunnel is described in RFC2401.  On encapsulation,
IPv4 TOS field (or, IPv6 traffic class field) will be copied from inner
IP header to outer IP header.  On decapsulation outer IP header
will be simply dropped.  The decapsulation rule is not compatible
with ECN, since ECN bit on the outer IP TOS/traffic class field will be
lost.
To make IPsec tunnel ECN-friendly, we should modify encapsulation
and decapsulation procedure.  This is described in
http://www.aciri.org/floyd/papers/draft-ipsec-ecn-00.txt, chapter 3.

KAME IPsec tunnel implementation can give you three behaviors, by setting
net.inet.ipsec.ecn (or net.inet6.ipsec6.ecn) to some value:
- RFC2401: no consideration for ECN (sysctl value -1)
- ECN forbidden (sysctl value 0)
- ECN allowed (sysctl value 1)
Note that the behavior is configurable in per-node manner, not per-SA manner
(draft-ipsec-ecn-00 wants per-SA configuration, but it looks too much for me).

The behavior is summarized as follows (see source code for more detail):

		encapsulate			decapsulate
		---				---
RFC2401		copy all TOS bits		drop TOS bits on outer
		from inner to outer.		(use inner TOS bits as is)

ECN forbidden	copy TOS bits except for ECN	drop TOS bits on outer
		(masked with 0xfc) from inner	(use inner TOS bits as is)
		to outer.  set ECN bits to 0.

ECN allowed	copy TOS bits except for ECN	use inner TOS bits with some
		CE (masked with 0xfe) from	change.  if outer ECN CE bit
		inner to outer.			is 1, enable ECN CE bit on
		set ECN CE bit to 0.		the inner.

General strategy for configuration is as follows:
- if both IPsec tunnel endpoint are capable of ECN-friendly behavior,
  you'd better configure both end to "ECN allowed" (sysctl value 1).
- if the other end is very strict about TOS bit, use "RFC2401"
  (sysctl value -1).
- in other cases, use "ECN forbidden" (sysctl value 0).
The default behavior is "ECN forbidden" (sysctl value 0).

For more information, please refer to:
	http://www.aciri.org/floyd/papers/draft-ipsec-ecn-00.txt
	RFC2481 (Explicit Congestion Notification)
	KAME sys/netinet6/{ah,esp}_input.c

(Thanks goes to Kenjiro Cho <kjc@csl.sony.co.jp> for detailed analysis)

5. IPComp

IPComp stands for IP payload compression protocol.  This is aimed for
payload compression, not the header compression like PPP VJ compression.
This may be useful when you are using slow serial link (say, cell phone)
with powerful CPU (well, recent notebook PCs are really powerful...).
The protocol design of IPComp is very similar to IPsec.

KAME implements the following specifications:
- RFC2393: IP Payload Compression Protocol (IPComp)
- RFC2394: IP Payload Compression Using DEFLATE

Here are some points to be noted:
- IPComp is treated as part of IPsec protocol suite, and SPI and
  CPI space is unified.  Spec says that there's no relationship
  between two so they are assumed to be separate.
- IPComp association (IPCA) is kept in SAD.
- It is possible to use well-known CPI (CPI=2 for DEFLATE for example),
  for outbound/inbound packet, but for indexing purposes one element from
  SPI/CPI space will be occupied anyway.
- pfkey is modified to support IPComp.  However, there's no official
  SA type number assignment yet.  Portability with other IPComp
  stack is questionable (anyway, who else implement IPComp on UN*X?).
- Spec says that IPComp output must be performed before IPsec output
  processing.  However, with manual SPD setting, you can violate this
  ordering requirement (KAME code is too generic, maybe).
- Though MTU can be significantly decreased by using IPComp, no special
  consideration is made about path MTU (spec talks nothing about MTU
  consideration).  IPComp is designed for serial links, not ethernet-like
  medium, it seems.
- You can change compression ratio on outbound packet, by changing
  deflate_policy in sys/netinet6/ipcomp_core.c (should it be sysctl accessible?
  or per-SAD configurable?)

						<end of IMPLEMENTATION>