DHCP Conditional Evaluation

Specify conditional behavior for DHCP servers and clients

Description:

The Internet Systems Consortium DHCP client and server both provide the ability to perform conditional behavior depending on the contents of packets they receive. This chapter describes the syntax for specifying this conditional behavior.

Conditional behavior

Conditional behavior is specified using the if statement and the else or elsif statements. A conditional statement can appear anywhere that a regular statement (e.g., an option statement) can appear, and can enclose one or more such statements. A typical conditional statement in a server might be:

if option dhcp-user-class = "accounting" {
  max-lease-time 17600;
  option domain-name "accounting.example.org";
  option domain-name-servers ns1.accounting.example.org,
                    ns2.accounting.example.org;
} elsif option dhcp-user-class = "sales" {
  max-lease-time 17600;
  option domain-name "sales.example.org";
  option domain-name-servers ns1.sales.example.org,
                    ns2.sales.example.org;
} elsif option dhcp-user-class = "engineering" {
  max-lease-time 17600;
  option domain-name "engineering.example.org";
  option domain-name-servers ns1.engineering.example.org,
                    ns2.engineering.example.org;
} else {
  max-lease-time 600;
  option domain-name "misc.example.org";
  option domain-name-servers ns1.misc.example.org,
                    ns2.misc.example.org;
}

On the client side, an example of conditional evaluation might be:

# example.org filters DNS at its firewall, so we have to use their DNS
# servers when we connect to their network.   If we are not at
# example.org, prefer our own DNS server.
if not option domain-name = "example.org" {
  prepend domain-name-servers 127.0.0.1;
}

The if statement and the elsif continuation statement both take boolean expressions as their arguments. That is, they take expressions that, when evaluated, produce a boolean result. If the expression evaluates to true, then the statements enclosed in braces following the if statement are executed, and all subsequent elsif and else clauses are skipped. Otherwise, each subsequent elsif clause's expression is checked, until an elsif clause is encountered whose test evaluates to true. If such a clause is found, the statements in braces following it are executed, and then any subsequent elsif and else clauses are skipped. If all the if and elsif clauses are checked, but none of their expressions evaluate true, then if there's an else clause, the statements enclosed in braces following the else are evaluated. Boolean expressions that evaluate to null are treated as false in conditionals.

Boolean expressions

The DHCP distribution supports the following boolean expressions:

data-expression-1 = data-expression-2
The = operator compares the values of two data expressions, returning true if they are the same, false if they are not. If either the left-hand side or the right-hand side are null, the result is also null.
data-expression-1 ~= data-expression-2 data-expression-1 ~~ data-expression-2
The ~= and ~~ operators (not available on all systems) perform extended regex matching of the values of two data expressions, returning true if data-expression-1 matches against the regular expression evaluated by data-expression-2, or false if it does not match or encounters some error. If either the left-hand side or the right-hand side is null, the result is also false. The ~~ operator differs from the ~= operator in that it is case-insensitive.
boolean-expression-1 and boolean-expression-2
The and operator evaluates to true if the boolean expression on the left-hand side and the boolean expression on the right-hand side both evaluate to true. Otherwise, it evaluates to false. If either the expression on the left-hand side or the expression on the right-hand side is null, the result is null.
boolean-expression-1 or boolean-expression-2
The or operator evaluates to true if either the boolean expression on the left-hand side or the boolean expression on the right-hand side evaluate to true. Otherwise, it evaluates to false. If either the expression on the left-hand side or the expression on the right-hand side is null, the result is null.
not boolean-expression
The not operator evaluates to true if boolean-expression evaluates to false, and returns false if boolean-expression evaluates to true. If boolean-expression evaluates to null, the result is also null.
exists option-name
The exists expression returns true if the specified option exists in the incoming DHCP packet being processed.
known
The known expression returns true if the client whose request is currently being processed is known — that is, if there's a host declaration for it.
static
The static expression returns true if the lease assigned to the client whose request is currently being processed is derived from a static address assignment.

Data expressions

Several of the boolean expressions above depend on the results of evaluating data expressions. A list of these expressions is provided here.

substring (data-expr, offset, length)
The substring operator evaluates the data expression and returns the substring of the result of that evaluation that starts offset bytes from the beginning, continuing for length bytes. The offset and length are both numeric expressions. If data-expr, offset or length evaluates to null, then the result is also null. If offset is greater than or equal to the length of the evaluated data, then a zero-length data string is returned. If length is greater than the remaining length of the evaluated data after offset, then a data string containing all data from offset to the end of the evaluated data is returned.
suffix (data-expr, length)
The suffix operator evaluates data-expr and returns the last length bytes of the result of that evaluation. The length is a numeric expression. If data-expr or length evaluates to null, then the result is also null. If suffix evaluates to a number greater than the length of the evaluated data, then the evaluated data is returned.
lcase (data-expr)
The lcase function returns the result of evaluating data-expr converted to lower case. If data-expr evaluates to null, then the result is also null.
ucase (data-expr)
The ucase function returns the result of evaluating data-expr converted to upper case. If data-expr evaluates to null, then the result is also null.
option option-name
The option operator returns the contents of the specified option in the packet to which the server is responding.
config-option option-name
The config-option operator returns the value for the specified option that the DHCP client or server has been configured to send.
hardware
The hardware operator returns a data string whose first element is the type of network interface indicated in packet being considered, and whose subsequent elements are client's link-layer address. If there is no packet, or if the RFC2131 hlen field is invalid, then the result is null. Hardware types include ethernet (1), token-ring (6), and fddi (8). Hardware types are specified by the IETF, and details on how the type numbers are defined can be found in RFC2131 (in the ISC DHCP distribution, this is included in the doc/ subdirectory).
packet (offset, length)
The packet operator returns the specified portion of the packet being considered, or null in contexts where no packet is being considered. The offset and length are applied to the contents of the packet as in the substring operator.
string
A string, enclosed in quotes, may be specified as a data expression, and returns the text between the quotes, encoded in ASCII. The backslash (\) character is treated specially, as in C programming: \t means TAB, \r means carriage return, \n means newline, and \b means bell. Any octal value can be specified with \ nnn, where nnn is any positive octal number less than 0400. Any hexadecimal value can be specified with \x nn, where nn is any positive hexadecimal number less than or equal to 0xff.
colon-separated hexadecimal list
A list of hexadecimal octet values, separated by colons, may be specified as a data expression.
concat (data-expr1, ,,,, data-exprN)
The expressions are evaluated, and the results of each evaluation are concatenated in the sequence that the subexpressions are listed. If any subexpression evaluates to null, the result of the concatenation is null.
reverse (numeric-expr1, data-expr2)
The two expressions are evaluated, and then the result of evaluating the data expression is reversed in place, using hunks of the size specified in the numeric expression. For example, if the numeric expression evaluates to four, and the data expression evaluates to twelve bytes of data, then the reverse expression will evaluate to twelve bytes of data, consisting of the last four bytes of the the input data, followed by the middle four bytes, followed by the first four bytes.
leased-address
In any context where the client whose request is being processed has been assigned an IP address, this data expression returns that IP address. In any context where the client whose request is being processed has not been assigned an IP address, if this data expression is found in executable statements executed on that client's behalf, a log message indicating "there is no lease associated with this client" is syslogged to the debug level (this is considered dhcpd.conf debugging information).
binary-to-ascii (numeric-expr1, numeric-expr2, data-expr, data-expr2)
Converts the result of evaluating data-expr2 into a text string containing one number for each element of the result of evaluating data-expr2. Each number is separated from the other by the result of evaluating data-expr1. The result of evaluating numeric-expr1 specifies the base (2 through 16) into which the numbers should be converted. The result of evaluating numeric-expr2 specifies the width in bits of each number, which may be 8, 16, or 32.

As an example of the preceding three types of expressions, to produce the name of a PTR record for the IP address being assigned to a client, you could write the following expression:

concat (binary-to-ascii (10, 8, ".",
                         reverse (1, leased-address)),
                         ".in-addr.arpa.");
  
encode-int (numeric-expr, width)
The numeric-expr is evaluated and encoded as a data string of the specified width, in network byte order (most significant byte first). If the numeric expression evaluates to the null value, the result is also null.
pick-first-value (data-expr1 [ ... exprn ] )
The pick-first-value function takes any number of data expressions as its arguments. Each expression is evaluated, starting with the first in the list, until an expression is found that doesn't evaluate to a null value. That expression is returned, and none of the subsequent expressions are evaluated. If all expressions evaluate to a null value, the null value is returned.
host-decl-name
The host-decl-name function returns the name of the host declaration that matched the client whose request is currently being processed, if any. If no host declaration matched, the result is the null value.

Numeric expressions

Numeric expressions are expressions that evaluate to an integer. In general, the maximum size of such an integer should not be assumed to be representable in fewer than 32 bits, but the precision of such integers may be more than 32 bits.

extract-int (data-expr, width)
The extract-int operator extracts an integer value in network byte order from the result of evaluating the specified data expression. Width is the width in bits of the integer to extract. Currently, the only supported widths are 8, 16, and 32. If the evaluation of the data expression doesn't provide sufficient bits to extract an integer of the specified size, the null value is returned.
lease-time
The duration of the current lease; that is, the difference between the current time and the time that the lease expires.
number
Any number between zero and the maximum representable size may be specified as a numeric expression.
client-state
The current state of the client instance being processed. This is only useful in DHCP client configuration files. Possible values are:
  • Booting — the DHCP client is in the INIT state, and doesn't yet have an IP address. The next message transmitted will be a DHCPDISCOVER, which will be broadcast.
  • Reboot — the DHCP client is in the INIT-REBOOT state. It has an IP address, but is not yet using it. The next message to be transmitted will be a DHCPREQUEST, which will be broadcast. If no response is heard, the client will bind to its address and move to the BOUND state.
  • Select — the DHCP client is in the SELECTING state; it has received at least one DHCPOFFER message, but is waiting to see if it may receive other DHCPOFFER messages from other servers. No messages are sent in the SELECTING state.
  • Request — the DHCP client is in the REQUESTING state; it has received at least one DHCPOFFER message, and has chosen which one it will request. The next message to be sent will be a DHCPREQUEST message, which will be broadcast.
  • Bound — the DHCP client is in the BOUND state; it has an IP address. No messages are transmitted in this state.
  • Renew — the DHCP client is in the RENEWING state; it has an IP address, and is trying to contact the server to renew it. The next message to be sent will be a DHCPREQUEST message, which will be unicast directly to the server.
  • Rebind — the DHCP client is in the REBINDING state; it has an IP address, and is trying to contact any server to renew it. The next message to be sent will be a DHCPREQUEST, which will be broadcast.

Action expressions

log (priority, data-expr)
Logging statements may be used to send information to the standard logging channels. A logging statement includes an optional priority (fatal, error, info, or debug), and a data expression.

Logging statements take only a single data expression argument, so if you want to output multiple data values, you will need to use the concat operator to concatenate them.

execute (command-path [, data-expr1, ... data-exprN]);
The execute statement runs an external command. The first argument is a string literal containing the name or path of the command to run. The other arguments, if present, are either string literals or data-expressions which evaluate to text strings, to be passed as command-line arguments to the command. The execute statement is synchronous; the program will block until the external command being run has finished.

Lengthy program execution (for example, in an "on commit" in dhcpd.conf) may result in bad performance and timeouts. Only external applications with very short execution times are suitable for use.

Passing user-supplied data to an external application might be dangerous. Make sure the external application checks input buffers for validity. Non-printable ASCII characters will be converted into dhcpd.conf language octal escapes (777), so make sure your external command handles them as such.

It's possible to use the execute statement in any context, not only on events. If you put it in a regular scope in the configuration file, you will execute that command every time a scope is evaluated.

Dynamic DNS updates

The DHCP client and server have the ability to dynamically update the Domain Name System. Within the configuration files, you can define how you want the Domain Name System to be updated. These updates are RFC2136-compliant, so any DNS server supporting RFC2136 should be able to accept updates from the DHCP server.

Security

Support for TSIG and DNSSEC is not yet available. When you set your DNS server up to allow updates from the DHCP server or client, you may be exposing it to unauthorized updates. To avoid this, the best you can do right now is to use IP address-based packet filtering to prevent unauthorized hosts from submitting update requests. Obviously, there is currently no way to provide security for client updates - this will require TSIG or DNSSEC, neither of which is yet available in the DHCP distribution.

Dynamic DNS (DDNS) updates are performed by using the dns-update expression. The dns-update expression is a boolean expression that takes four parameters. If the update succeeds, the result is true. If it fails, the result is false. The four parameters are:

  • the resource record type (RR)
  • the left hand side of the RR
  • the right hand side of the RR
  • the ttl that should be applied to the record

The simplest example of the use of the function can be found in the reference section of the dhcpd.conf file, where events are described. In this example, several statements are being used to make the arguments to the dns-update.

In the example, the first argument to the first dns-update expression is a data expression that evaluates to the A RR type. The second argument is constructed by concatenating the DHCP host-name option with a text string containing the local domain, in this case ssd.example.net. The third argument is constructed by converting the address the client has been assigned from a 32-bit number into an ASCII string with each byte separated by a period (.). The fourth argument, the TTL, specifies the amount of time remaining in the lease (note that this isn't really correct, since the DNS server will pass this TTL out whenever a request comes in, even if that is only a few seconds before the lease expires).

If the first dns-update statement succeeds, it's followed up with a second update to install a PTR RR. The installation of a PTR record is similar to installing an A RR except that the left hand side of the record is the leased address, reversed, with .in-addr.arpa concatenated. The right hand side is the fully qualified domain name of the client to which the address is being leased.

Contributing author

The Internet Systems Consortium DHCP Distribution was written by Ted Lemon under a contract with Vixie Labs. Funding for this project was provided through Internet Systems Consortium. Information about Internet Systems Consortium can be found at http://www.isc.org .

See also:

RFC2132, RFC2131

Last modified: 2013-12-21

comments powered by Disqus