<?xml version="1.0" encoding="latin1" ?>
<!DOCTYPE erlref SYSTEM "erlref.dtd">
<erlref>
<header>
<copyright>
<year>2009</year><year>2009</year>
<holder>Ericsson AB. All Rights Reserved.</holder>
</copyright>
<legalnotice>
The contents of this file are subject to the Erlang Public License,
Version 1.1, (the "License"); you may not use this file except in
compliance with the License. You should have received a copy of the
Erlang Public License along with this software. If not, it can be
retrieved online at http://www.erlang.org/.
Software distributed under the License is distributed on an "AS IS"
basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See
the License for the specific language governing rights and limitations
under the License.
</legalnotice>
<title>inet_res</title>
<prepared>[email protected]</prepared>
<docno></docno>
<date>2009-09-11</date>
<rev>A</rev>
</header>
<module>inet_res</module>
<modulesummary>A Rudimentary DNS Client</modulesummary>
<description>
<p>Performs DNS name resolving towards recursive name servers</p>
<p>See also
<seealso marker="erts:inet_cfg">
ERTS User's Guide: Inet configuration
</seealso> for more
information on how to configure an Erlang runtime system for IP
communication and how to enable this DNS client by defining
<c><![CDATA['dns']]></c> as a lookup method. It then acts
as a backend for the resolving functions in
<seealso marker="kernel:inet">inet</seealso>.</p>
<p>This DNS client can resolve DNS records even if it
is not used for normal name resolving in the node.</p>
<p>This is not a full-fledged resolver. It is just a
DNS client that relies on asking trusted recursive nameservers.</p>
</description>
<section>
<title>Name Resolving</title>
<p>UDP queries are used unless resolver option
<c>usevc</c> is <c>true</c>, which forces TCP queries.
If the query is to large for UDP, TCP is used instead.
For regular DNS queries 512 bytes is the size limit.
When EDNS is enabled (resolver option
<c>edns</c> is set to the EDNS version i.e <c>0</c>
instead of <c>false</c>), resolver option
<c>udp_payload_size</c> sets the limit. If a nameserver
replies with the TC bit set (truncation), indicating
the answer is incomplete, the query is retried
to that nameserver using TCP. The resolver option
<c>udp_payload_size</c> also sets the advertised
size for the max allowed reply size, if EDNS is
enabled, otherwise the nameserver uses the limit
512 byte. If the reply is larger it gets truncated,
forcing a TCP re-query.</p>
<p>For UDP queries, the resolver options <c>timeout</c>
and <c>retry</c> control retransmission.
Each nameserver in the <c>nameservers</c> list is
tried with a timeout of <c>timeout</c> / <c>retry</c>.
Then all nameservers are tried again doubling the
timeout, for a total of <c>retry</c> times.</p>
<p>For queries that not use the <c>search</c> list,
if the query to all <c>nameservers</c> results in
<c>{error,nxdomain}</c>or an empty answer, the same
query is tried for the <c>alt_nameservers</c>.</p>
</section>
<section>
<title>DATA TYPES</title>
<p>As defined in the module
<seealso marker="kernel:inet">inet</seealso>:</p>
<code type="none">
hostent() = #hostent{}
posix() = some atom()s
ip_address() = tuple of integers of arity 4 or 8</code>
<p>Resolver types:</p>
<code type="none">These correspond to resolver options:
res_option() =
[ {alt_nameservers, [ nameserver() ]}
| {edns, 0 | false} % Use EDNS
| {inet6, bool()} % Return IPv6 addresses
| {nameservers, [ nameserver() ]} % List of nameservers
| {recurse, bool()} % Request server recursion
| {retry, integer()} % UDP retries
| {timeout, integer()} % UDP query timeout
| {udp_payload_size, integer()} % EDNS payload size
| {usevc, bool()} ] % Use TCP (Virtual Circuit)
nameserver() = {ip_address(),Port}
Port = integer(1..65535)
res_error() =
formerr |
qfmterror |
servfail |
nxdomain |
notimp |
refused |
badvers |
timeout
</code>
<p>DNS types:</p>
<marker id="dns_types"/>
<code type="none">dns_name() = string() with no adjacent dots
rr_type() = a | aaaa | cname | gid | hinfo | ns | mb | md | mg | mf
| minfo | mx | naptr | null | ptr | soa | spf | srv | txt
| uid | uinfo | unspec | wks
query_type() = axfr | mailb | maila | any | rr_type()
dns_class() = in | chaos | hs | any
dns_msg() = DnsMsg
This is the start of a hiearchy of opaque data structures
that can be examined with access functions in inet_dns
that return lists of {Field,Value} tuples. The arity 2
functions just return the value for a given field.
inet_dns:msg(DnsMsg) ->
[ {header, dns_header()}
| {qdlist, dns_query()}
| {anlist, dns_rr()}
| {nslist, dns_rr()}
| {arlist, dns_rr()} ]
inet_dns:msg(DnsMsg, header) -> dns_header() % for example
inet_dns:msg(DnsMsg, Field) -> Value
dhs_header() = DnsHeader
inet_dns:header(DnsHeader) ->
[ {id, integer()}
| {qr, bool()}
| {opcode, 'query' | iquery | status | integer()}
| {aa, bool()}
| {tc, bool()}
| {rd, bool()}
| {ra, bool()}
| {pr, bool()}
| {rcode, integer(0..16)} ]
inet_dns:header(DnsHeader, Field) -> Value
dns_query() = DnsQuery
inet_dns:dns_query(DnsQuery) ->
[ {domain, dns_name()}
| {type, query_type()}
| {class, dns_class()} ]
inet_dns:dns_query(DnsQuery, Field) -> Value
dns_rr() = DnsRr
inet_dns:rr(DnsRr) -> DnsRrFields | DnsRrOptFields
DnsRrFields = [ {domain, dns_name()}
| {type, rr_type()}
| {class, dns_class()}
| {ttl, integer()}
| {data, dns_data()} ]
DnsRrOptFields = [ {domain, dns_name()}
| {type, opt}
| {udp_payload_size, integer()}
| {ext_rcode, integer()}
| {version, integer()}
| {z, integer()}
| {data, dns_data()} ]
inet_dns:rr(DnsRr, Field) -> Value
dns_data() = % for dns_type()
[ dns_name() % ns, md, mf, cname, mb, mg, mr, ptr
| ip_address(v4) % a
| ip_address(v6) % aaaa
| {MName,RName,Serial,Refresh,Retry,Expiry,Minimum} % soa
| {ip_address(v4),Proto,BitMap} % wks
| {CpuString,OsString} % hinfo
| {RM,EM} % minfo
| {Prio,dns_name()} % mx
| {Prio,Weight,Port,dns_name()} % srv
| {Order,Preference,Flags,Services,Regexp,dns_name()} % naptr
| [ string() ] % txt, spf
| binary() ] % null, integer()
MName, RName = dns_name()
Serial, Refresh, Retry, Expiry, Minimum = integer(),
Proto = integer()
BitMap = binary()
CpuString, OsString = string()
RM = EM = dns_name()
Prio, Weight, Port = integer()
Order, Preference = integer()
Flags, Services = string(),
Regexp = string(utf8)
There is an info function for the types above:
inet_dns:record_type(dns_msg()) -> msg;
inet_dns:record_type(dns_header()) -> header;
inet_dns:record_type(dns_query()) -> dns_query;
inet_dns:record_type(dns_rr()) -> rr;
inet_dns:record_type(_) -> undefined.
So; inet_dns:(inet_dns:record_type(X))(X) will convert
any of these data structures into a {Field,Value} list.</code>
</section>
<funcs>
<func>
<name>getbyname(Name, Type) -> {ok,hostent()} | {error,Reason}</name>
<name>getbyname(Name, Type, Timeout) ->
{ok,hostent()} | {error,Reason}
</name>
<fsummary>Resolve a DNS record of the given type for the given host
</fsummary>
<type>
<v>Name = dns_name()</v>
<v>Type = rr_type()</v>
<v>Timeout = integer() >= 0 | infinity</v>
<v>Reason = posix() | res_error()</v>
</type>
<desc>
<p>Resolve a DNS record of the given type for the given host,
of class <c>in</c>. On success returns a <c>hostent()</c> record with
<c>dns_data()</c> elements in the address list field.
</p><p>
This function uses the resolver option <c>search</c> that
is a list of domain names. If the name to resolve contains
no dots, it is prepended to each domain name in the
search list, and they are tried in order. If the name
contains dots, it is first tried as an absolute name
and if that fails the search list is used. If the name
has a trailing dot it is simply supposed to be
an absolute name and the search list is not used.
</p>
</desc>
</func>
<func>
<name>gethostbyaddr(Address) -> {ok,hostent()} | {error,Reason}</name>
<name>gethostbyaddr(Address, Timeout) ->
{ok,hostent()} | {error,Reason}
</name>
<fsummary>Return a hostent record for the host with the given address
</fsummary>
<type>
<v>Address = ip_address()</v>
<v>Timeout = integer() >= 0 | infinity</v>
<v>Reason = posix() | res_error()</v>
</type>
<desc>
<p>Backend functions used by
<seealso marker="kernel:inet#gethostbyaddr/1">
inet:gethostbyaddr/1
</seealso>.
</p>
</desc>
</func>
<func>
<name>gethostbyname(Name) -> {ok,hostent()} | Reason}</name>
<name>gethostbyname(Name, Family) ->
{ok,hostent()} | {error,Reason}}
</name>
<name>gethostbyname(Name, Family, Timeout) ->
{ok,hostent()} | {error,Reason}
</name>
<fsummary>Return a hostent record for the host with the given name
</fsummary>
<type>
<v>Name = dns_name()</v>
<v>Timeout = integer() >= 0 | infinity</v>
<v>Reason = posix() | res_error()</v>
</type>
<desc>
<p>Backend functions used by
<seealso marker="kernel:inet#gethostbyname/1">
inet:gethostbyname/1,2
</seealso>.
</p><p>
This function uses the resolver option <c>search</c> just like
<seealso marker="#getbyname/2">getbyname/2,3</seealso>.
</p><p>
If the resolver option <c>inet6</c> is <c>true</c>,
an IPv6 address is looked up, and if that fails
the IPv4 address is looked up and returned on
IPv6 mapped IPv4 format.
</p>
</desc>
</func>
<func>
<name>lookup(Name, Class, Type) -> [ dns_data() ]
</name>
<name>lookup(Name, Class, Type, Opts) -> [ dns_data() ]
</name>
<name>lookup(Name, Class, Type, Opts, Timeout) -> [ dns_data() ]
</name>
<fsummary>Resolve the DNS data for the record of the given type and class
for the given name
</fsummary>
<type>
<v>Name = dns_name() | ip_address()</v>
<v>Type = rr_type()</v>
<v>Opts = res_option() | verbose</v>
<v>Timeout = integer() >= 0 | infinity</v>
<v>Reason = posix() | res_error()</v>
</type>
<desc>
<p>Resolve the DNS data for the record of the given type and class
for the given name. On success filters out the answer records
with the correct <c>Class</c> and <c>Type</c> and returns
a list of their data fields. So a lookup for type <c>any</c>
will give an empty answer since the answer records have
specific types that are not <c>any</c>. An empty answer
as well as a failed lookup returns an empty list.
</p><p>
Calls <seealso marker="#resolve/3">resolve/2..4</seealso>
with the same arguments and filters the result, so
<c>Opts</c> is explained there.
</p>
</desc>
</func>
<func>
<name>resolve(Name, Class, Type) -> {ok,dns_msg()} | Error
</name>
<name>resolve(Name, Class, Type, Opts) -> {ok,dns_msg()} | Error
</name>
<name>resolve(Name, Class, Type, Opts, Timeout) -> {ok,dns_msg()} | Error
</name>
<fsummary>Resolve a DNS record of the given type and class
for the given name
</fsummary>
<type>
<v>Name = dns_name() | ip_address()</v>
<v>Type = rr_type()</v>
<v>Opts = res_option() | verbose | atom()</v>
<v>Timeout = integer() >= 0 | infinity</v>
<v>Error = {error,Reason} | {error,{Reason,dns_msg()}}</v>
<v>Reason = posix() | res_error()</v>
</type>
<desc>
<p>Resolve a DNS record of the given type and class for the given name.
The returned <c>dns_msg()</c> can be examined using
access functions in <c>inet_db</c> as described
in <seealso marker="#dns_types">DNS types</seealso>.
</p><p>
If <c>Name</c> is an <c>ip_address()</c>, the domain name
to query for is generated as the standard reverse
".IN-ADDR.ARPA." name for an IPv4 address, or the
".IP6.ARPA." name for an IPv6 address.
In this case you most probably want to use
<c>Class = in</c> and <c>Type = ptr</c> but it
is not done automatically.
</p><p>
<c>Opts</c> override the corresponding resolver options.
If the option <c>nameservers</c> is given, it is
also assumed that it is the complete list of nameserves,
so the resolver option <c>alt_nameserves</c> is ignored.
Of course, if that option is also given to this function,
it is used.
</p><p>
The <c>verbose</c> option (or rather <c>{verbose,true}</c>),
causes diagnostics printout through
<seealso marker="stdlib:io#format/3">io:format/2</seealso>
of queries, replies retransmissions, etc, similar
to from utilities like <c>dig</c>, <c>nslookup</c> et.al.
</p><p>
If <c>Opt</c> is an arbitrary atom it is interpreted
as <c>{Opt,true}</c> unless the atom string starts with
<c>"no"</c> making the interpretation <c>{Opt,false}</c>.
For example: <c>usevc</c> is an alias for <c>{usevc,true}</c>,
and <c>nousevc</c> an alias for <c>{usevc,false}</c>.
</p><p>
The <c>inet6</c> option currently has no effect on this function.
You probably want to use <c>Type = a | aaaa</c> instead.
</p>
</desc>
</func>
</funcs>
<section>
<title>Examples</title>
<p>Access functions example: how
<seealso marker="#lookup/3">lookup/3</seealso>
could have been implemented using
<seealso marker="#resolve/3">resolve/3</seealso>
from outside the module.
</p><code type="none">
example_lookup(Name, Class, Type) ->
case inet_res:resolve(Name, Class, Type) of
{ok,Msg} ->
[inet_dns:rr(RR, data)
|| RR <- inet_dns:msg(Msg, anlist),
inet_dns:rr(RR, type) =:= Type,
inet_dns:rr(RR, class) =:= Class];
{error,_} ->
[]
end.</code>
</section>
<section>
<title>Legacy Functions</title>
<p>These have been deprecated due to the annoying double
meaning of the nameservers/timeout argument, and
because they had no decent place for a resolver options list.</p>
</section>
<funcs>
<func>
<name>nslookup(Name, Class, Type) -> {ok,dns_msg()} | {error,Reason}
</name>
<name>nslookup(Name, Class, Type, Timeout) ->
{ok,dns_msg()} | {error,Reason}
</name>
<name>nslookup(Name, Class, Type, Nameservers) ->
{ok,dns_msg()} | {error,Reason}
</name>
<fsummary>Resolve a DNS record of the given type and class
for the given name
</fsummary>
<type>
<v>Name = dns_name() | ip_address()</v>
<v>Type = rr_type()</v>
<v>Nameservers = [ nameserver() ]</v>
<v>Timeout = integer() >= 0 | infinity</v>
<v>Reason = posix() | res_error()</v>
</type>
<desc>
<p>Resolve a DNS record of the given type and class for the given name.
</p>
</desc>
</func>
<func>
<name>nnslookup(Name, Class, Type, Nameservers) ->
{ok,dns_msg()} | {error,posix()}
</name>
<name>nnslookup(Name, Class, Type, Nameservers, Timeout) ->
{ok,dns_msg()} | {error,posix()}
</name>
<fsummary>Resolve a DNS record of the given type and class
for the given name
</fsummary>
<type>
<v>Name = dns_name() | ip_address()</v>
<v>Type = rr_type()</v>
<v>Nameservers = [ nameserver() ]</v>
<v>Timeout = integer() >= 0 | infinity</v>
<v>Reason = posix() | res_error()</v>
</type>
<desc>
<p>Resolve a DNS record of the given type and class for the given name.
</p>
</desc>
</func>
</funcs>
</erlref>
