man/man4/bpf.4

.\" -*- nroff -*-
.\"
.\"	$NetBSD: bpf.4,v 1.61 2018/06/26 06:47:57 msaitoh Exp $
.\"
.\" Copyright (c) 1990, 1991, 1992, 1993, 1994
.\"	The Regents of the University of California.  All rights reserved.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that: (1) source code distributions
.\" retain the above copyright notice and this paragraph in its entirety, (2)
.\" distributions including binary code include the above copyright notice and
.\" this paragraph in its entirety in the documentation or other materials
.\" provided with the distribution, and (3) all advertising materials mentioning
.\" features or use of this software display the following acknowledgement:
.\" ``This product includes software developed by the University of California,
.\" Lawrence Berkeley Laboratory and its contributors.'' Neither the name of
.\" the University nor the names of its contributors may be used to endorse
.\" or promote products derived from this software without specific prior
.\" written permission.
.\" THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR IMPLIED
.\" WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF
.\" MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
.\"
.\" This document is derived in part from the enet man page (enet.4)
.\" distributed with 4.3BSD Unix.
.\"
.Dd June 22, 2018
.Dt BPF 4
.Os
.Sh NAME
.Nm bpf
.Nd Berkeley Packet Filter raw network interface
.Sh SYNOPSIS
.Cd "pseudo-device bpfilter"
.Sh DESCRIPTION
The Berkeley Packet Filter
provides a raw interface to data link layers in a protocol
independent fashion.
All packets on the network, even those destined for other hosts,
are accessible through this mechanism.
.Pp
The packet filter appears as a character special device,
.Pa /dev/bpf .
After opening the device, the file descriptor must be bound to a
specific network interface with the
.Dv BIOCSETIF
ioctl.
A given interface can be shared by multiple listeners, and the filter
underlying each descriptor will see an identical packet stream.
.Pp
Associated with each open instance of a
.Nm
file is a user-settable packet filter.
Whenever a packet is received by an interface,
all file descriptors listening on that interface apply their filter.
Each descriptor that accepts the packet receives its own copy.
.Pp
Reads from these files return the next group of packets
that have matched the filter.
To improve performance, the buffer passed to read must be
the same size as the buffers used internally by
.Nm .
This size is returned by the
.Dv BIOCGBLEN
ioctl (see below), and can be set with
.Dv BIOCSBLEN .
Note that an individual packet larger than this size is necessarily
truncated.
.Pp
Since packet data is in network byte order, applications should use the
.Xr byteorder 3
macros to extract multi-byte values.
.Pp
A packet can be sent out on the network by writing to a
.Nm
file descriptor.
The writes are unbuffered, meaning only one packet can be processed per write.
Currently, only writes to Ethernets and SLIP links are supported.
.Sh IOCTLS
The
.Xr ioctl 2
command codes below are defined in
.In net/bpf.h .
All commands require these includes:
.Bd -literal -offset indent
#include <sys/types.h>
#include <sys/time.h>
#include <sys/ioctl.h>
#include <net/bpf.h>
.Ed
.Pp
Additionally,
.Dv BIOCGETIF
and
.Dv BIOCSETIF
require
.Pa <net/if.h> .
.Pp
The (third) argument to the
.Xr ioctl 2
should be a pointer to the type indicated.
.Bl -tag -width indent -offset indent
.It Dv BIOCGBLEN ( u_int )
Returns the required buffer length for reads on
.Nm
files.
.It Dv BIOCSBLEN ( u_int )
Sets the buffer length for reads on
.Nm
files.
The buffer must be set before the file is attached to an interface with
.Dv BIOCSETIF .
If the requested buffer size cannot be accommodated, the closest
allowable size will be set and returned in the argument.
A read call will result in
.Er EINVAL
if it is passed a buffer that is not this size.
.It Dv BIOCGDLT ( u_int )
Returns the type of the data link layer underlying the attached interface.
.Er EINVAL
is returned if no interface has been specified.
The device types, prefixed with
.Dq DLT_ ,
are defined in
.In net/bpf.h .
.It Dv BIOCGDLTLIST ( struct bpf_dltlist )
Returns an array of the available types of the data link layer
underlying the attached interface:
.Bd -literal -offset indent
struct bpf_dltlist {
	u_int bfl_len;
	u_int *bfl_list;
};
.Ed
.Pp
The available types are returned in the array pointed to by the
.Va bfl_list
field while their length in u_int is supplied to the
.Va bfl_len
field.
.Er ENOMEM
is returned if there is not enough buffer space and
.Er EFAULT
is returned if a bad address is encountered.
The
.Va bfl_len
field is modified on return to indicate the actual length in u_int
of the array returned.
If
.Va bfl_list
is
.Dv NULL ,
the
.Va bfl_len
field is set to indicate the required length of an array in u_int.
.It Dv BIOCSDLT ( u_int )
Changes the type of the data link layer underlying the attached interface.
.Er EINVAL
is returned if no interface has been specified or the specified
type is not available for the interface.
.It Dv BIOCPROMISC
Forces the interface into promiscuous mode.
All packets, not just those destined for the local host, are processed.
Since more than one file can be listening on a given interface,
a listener that opened its interface non-promiscuously may receive
packets promiscuously.
This problem can be remedied with an appropriate filter.
.Pp
The interface remains in promiscuous mode until all files listening
promiscuously are closed.
.It Dv BIOCFLUSH
Flushes the buffer of incoming packets,
and resets the statistics that are returned by
.Dv BIOCGSTATS .
.It Dv BIOCGETIF ( struct ifreq )
Returns the name of the hardware interface that the file is listening on.
The name is returned in the ifr_name field of
.Fa ifr .
All other fields are undefined.
.It Dv BIOCSETIF ( struct ifreq )
Sets the hardware interface associated with the file.
This command must be performed before any packets can be read.
The device is indicated by name using the
.Dv ifr_name
field of the
.Fa ifreq .
Additionally, performs the actions of
.Dv BIOCFLUSH .
.It Dv BIOCSRTIMEOUT , BIOCGRTIMEOUT ( struct timeval )
Sets or gets the read timeout parameter.
The
.Fa timeval
specifies the length of time to wait before timing
out on a read request.
This parameter is initialized to zero by
.Xr open 2 ,
indicating no timeout.
.It Dv BIOCGSTATS ( struct bpf_stat )
Returns the following structure of packet statistics:
.Bd -literal -offset indent
struct bpf_stat {
	uint64_t bs_recv;
	uint64_t bs_drop;
	uint64_t bs_capt;
	uint64_t bs_padding[13];
};
.Ed
.Pp
The fields are:
.Bl -tag -width bs_recv -offset indent
.It Va bs_recv
the number of packets received by the descriptor since opened or reset
(including any buffered since the last read call);
.It Va bs_drop
the number of packets which were accepted by the filter but dropped by the
kernel because of buffer overflows
(i.e., the application's reads aren't keeping up with the packet
traffic); and
.It Va bs_capt
the number of packets accepted by the filter.
.El
.It Dv BIOCIMMEDIATE ( u_int )
Enables or disables
.Dq immediate mode ,
based on the truth value of the argument.
When immediate mode is enabled, reads return immediately upon packet
reception.
Otherwise, a read will block until either the kernel buffer
becomes full or a timeout occurs.
This is useful for programs like
.Xr rarpd 8 ,
which must respond to messages in real time.
The default for a new file is off.
.It Dv BIOCSETF ( struct bpf_program )
Sets the filter program used by the kernel to discard uninteresting
packets.
An array of instructions and its length are passed in using the following structure:
.Bd -literal -offset indent
struct bpf_program {
	u_int bf_len;
	struct bpf_insn *bf_insns;
};
.Ed
.Pp
The filter program is pointed to by the
.Va bf_insns
field while its length in units of
.Sq struct bpf_insn
is given by the
.Va bf_len
field.
Also, the actions of
.Dv BIOCFLUSH
are performed.
.Pp
See section
.Sy FILTER MACHINE
for an explanation of the filter language.
.It Dv BIOCVERSION ( struct bpf_version )
Returns the major and minor version numbers of the filter language currently
recognized by the kernel.
Before installing a filter, applications must check
that the current version is compatible with the running kernel.
Version numbers are compatible if the major numbers match and the
application minor is less than or equal to the kernel minor.
The kernel version number is returned in the following structure:
.Bd -literal -offset indent
struct bpf_version {
	u_short bv_major;
	u_short bv_minor;
};
.Ed
.Pp
The current version numbers are given by
.Dv BPF_MAJOR_VERSION
and
.Dv BPF_MINOR_VERSION
from
.In net/bpf.h .
An incompatible filter
may result in undefined behavior (most likely, an error returned by
.Xr ioctl 2
or haphazard packet matching).
.It Dv BIOCSRSIG , BIOCGRSIG ( u_int )
Sets or gets the receive signal.
This signal will be sent to the process or process group specified by
.Dv FIOSETOWN .
It defaults to
.Dv SIGIO .
.It Dv BIOCGHDRCMPLT , BIOCSHDRCMPLT ( u_int )
Sets or gets the status of the
.Dq header complete
flag.
Set to zero if the link level source address should be filled in
automatically by the interface output routine.
Set to one if the link level source address will be written,
as provided, to the wire.
This flag is initialized to zero by default.
.It Dv BIOCGSEESENT , BIOCSSEESENT ( u_int )
These commands are obsolete but left for compatibility.
Use
.Dv BIOCSDIRECTION
and
.Dv BIOCGDIRECTION
instead.
Set or get the flag determining whether locally generated packets on the
interface should be returned by BPF.
Set to zero to see only incoming packets on the interface.
Set to one to see packets originating locally and remotely on the interface.
This flag is initialized to one by default.
.It Dv BIOCSDIRECTION
.It Dv BIOCGDIRECTION
.Pq Li u_int
Set or get the setting determining whether incoming, outgoing, or all packets
on the interface should be returned by BPF.
Set to
.Dv BPF_D_IN
to see only incoming packets on the interface.
Set to
.Dv BPF_D_INOUT
to see packets originating locally and remotely on the interface.
Set to
.Dv BPF_D_OUT
to see only outgoing packets on the interface.
This setting is initialized to
.Dv BPF_D_INOUT
by default.
.It Dv BIOCFEEDBACK , BIOCSFEEDBACK , BIOCGFEEDBACK ( u_int )
Set (or get)
.Dq packet feedback mode .
This allows injected packets to be fed back as input to the interface when
output via the interface is successful.
The first name is meant for
.Fx
compatibility, the two others follow the Get/Set convention.
.\"When
.\".Dv BPF_D_INOUT
.\"direction is set, injected
Injected
outgoing packets are not returned by BPF to avoid
duplication.
This flag is initialized to zero by default.
.El
.Sh STANDARD IOCTLS
.Nm
now supports several standard
.Xr ioctl 2 Ns 's
which allow the user to do async and/or non-blocking I/O to an open
.Nm bpf
file descriptor.
.Bl -tag -width indent -offset indent
.It Dv FIONREAD ( int )
Returns the number of bytes that are immediately available for reading.
.It Dv FIONBIO ( int )
Set or clear non-blocking I/O.
If arg is non-zero, then doing a
.Xr read 2
when no data is available will return -1 and
.Va errno
will be set to
.Er EAGAIN .
If arg is zero, non-blocking I/O is disabled.
Note: setting this
overrides the timeout set by
.Dv BIOCSRTIMEOUT .
.It Dv FIOASYNC ( int )
Enable or disable async I/O.
When enabled (arg is non-zero), the process or process group specified by
.Dv FIOSETOWN
will start receiving SIGIO's when packets
arrive.
Note that you must do an
.Dv FIOSETOWN
in order for this to take effect, as
the system will not default this for you.
The signal may be changed via
.Dv BIOCSRSIG .
.It Dv FIOSETOWN , FIOGETOWN ( int )
Set or get the process or process group (if negative) that should receive SIGIO
when packets are available.
The signal may be changed using
.Dv BIOCSRSIG
(see above).
.El
.Sh BPF HEADER
The following structure is prepended to each packet returned by
.Xr read 2 :
.Bd -literal -offset indent
struct bpf_hdr {
	struct bpf_timeval bh_tstamp;
	uint32_t bh_caplen;
	uint32_t bh_datalen;
	uint16_t bh_hdrlen;
};
.Ed
.Pp
The fields, whose values are stored in host order, are:
.Bl -tag -width bh_datalen -offset indent
.It Va bh_tstamp
The time at which the packet was processed by the packet filter.
This structure differs from the standard
.Vt struct timeval
in that both members are of type
.Vt long .
.It Va bh_caplen
The length of the captured portion of the packet.
This is the minimum of
the truncation amount specified by the filter and the length of the packet.
.It Va bh_datalen
The length of the packet off the wire.
This value is independent of the truncation amount specified by the filter.
.It Va bh_hdrlen
The length of the BPF header, which may not be equal to
.Em sizeof(struct bpf_hdr) .
.El
.Pp
The
.Va bh_hdrlen
field exists to account for
padding between the header and the link level protocol.
The purpose here is to guarantee proper alignment of the packet
data structures, which is required on alignment sensitive
architectures and improves performance on many other architectures.
The packet filter ensures that the
.Va bpf_hdr
and the
.Em network layer
header will be word aligned.
Suitable precautions must be taken when accessing the link layer
protocol fields on alignment restricted machines.
(This isn't a problem on an Ethernet, since
the type field is a short falling on an even offset,
and the addresses are probably accessed in a bytewise fashion).
.Pp
Additionally, individual packets are padded so that each starts
on a word boundary.
This requires that an application
has some knowledge of how to get from packet to packet.
The macro
.Dv BPF_WORDALIGN
is defined in
.In net/bpf.h
to facilitate this process.
It rounds up its argument
to the nearest word aligned value (where a word is
.Dv BPF_ALIGNMENT
bytes wide).
.Pp
For example, if
.Sq Va p
points to the start of a packet, this expression
will advance it to the next packet:
.Pp
.Dl p = (char *)p + BPF_WORDALIGN(p->bh_hdrlen + p->bh_caplen)
.Pp
For the alignment mechanisms to work properly, the
buffer passed to
.Xr read 2
must itself be word aligned.
.Xr malloc 3
will always return an aligned buffer.
.Sh FILTER MACHINE
A filter program is an array of instructions, with all branches forwardly
directed, terminated by a
.Sy return
instruction.
Each instruction performs some action on the pseudo-machine state,
which consists of an accumulator, index register, scratch memory store,
and implicit program counter.
.Pp
The following structure defines the instruction format:
.Bd -literal -offset indent
struct bpf_insn {
	uint16_t code;
	u_char 	jt;
	u_char 	jf;
	uint32_t k;
};
.Ed
.Pp
The
.Va k
field is used in different ways by different instructions,
and the
.Va jt
and
.Va jf
fields are used as offsets
by the branch instructions.
The opcodes are encoded in a semi-hierarchical fashion.
There are eight classes of instructions: BPF_LD, BPF_LDX, BPF_ST, BPF_STX,
BPF_ALU, BPF_JMP, BPF_RET, and BPF_MISC.
Various other mode and
operator bits are or'd into the class to give the actual instructions.
The classes and modes are defined in
.In net/bpf.h .
.Pp
Below are the semantics for each defined BPF instruction.
We use the convention that A is the accumulator, X is the index register,
P[] packet data, and M[] scratch memory store.
P[i:n] gives the data at byte offset
.Dq i
in the packet,
interpreted as a word (n=4),
unsigned halfword (n=2), or unsigned byte (n=1).
M[i] gives the i'th word in the scratch memory store, which is only
addressed in word units.
The memory store is indexed from 0 to BPF_MEMWORDS-1.
.Va k ,
.Va jt ,
and
.Va jf
are the corresponding fields in the
instruction definition.
.Dq len
refers to the length of the packet.
.Bl -tag -width indent -offset indent
.It Sy BPF_LD
These instructions copy a value into the accumulator.
The type of the source operand is specified by an
.Dq addressing mode
and can be a constant
.Sy ( BPF_IMM ) ,
packet data at a fixed offset
.Sy ( BPF_ABS ) ,
packet data at a variable offset
.Sy ( BPF_IND ) ,
the packet length
.Sy ( BPF_LEN ) ,
or a word in the scratch memory store
.Sy ( BPF_MEM ) .
For
.Sy BPF_IND
and
.Sy BPF_ABS ,
the data size must be specified as a word
.Sy ( BPF_W ) ,
halfword
.Sy ( BPF_H ) ,
or byte
.Sy ( BPF_B ) .
Arithmetic overflow when calculating a variable offset terminates
the filter program and the packet is ignored.
The semantics of all the recognized BPF_LD instructions follow.
.Bl -column "BPF_LD_BPF_W_BPF_ABS" "A <- P[k:4]" -offset indent
.It Sy BPF_LD+BPF_W+BPF_ABS Ta A <- P[k:4]
.It Sy BPF_LD+BPF_H+BPF_ABS Ta A <- P[k:2]
.It Sy BPF_LD+BPF_B+BPF_ABS Ta A <- P[k:1]
.It Sy BPF_LD+BPF_W+BPF_IND Ta A <- P[X+k:4]
.It Sy BPF_LD+BPF_H+BPF_IND Ta A <- P[X+k:2]
.It Sy BPF_LD+BPF_B+BPF_IND Ta A <- P[X+k:1]
.It Sy BPF_LD+BPF_W+BPF_LEN Ta A <- len
.It Sy BPF_LD+BPF_IMM Ta A <- k
.It Sy BPF_LD+BPF_MEM Ta A <- M[k]
.El
.It Sy BPF_LDX
These instructions load a value into the index register.
Note that the addressing modes are more restricted than those of
the accumulator loads, but they include
.Sy BPF_MSH ,
a hack for efficiently loading the IP header length.
.Bl -column "BPF_LDX_BPF_W_BPF_MEM" "X <- k" -offset indent
.It Sy BPF_LDX+BPF_W+BPF_IMM Ta X <- k
.It Sy BPF_LDX+BPF_W+BPF_MEM Ta X <- M[k]
.It Sy BPF_LDX+BPF_W+BPF_LEN Ta X <- len
.It Sy BPF_LDX+BPF_B+BPF_MSH Ta X <- 4*(P[k:1]&0xf)
.El
.It Sy BPF_ST
This instruction stores the accumulator into the scratch memory.
We do not need an addressing mode since there is only one possibility
for the destination.
.Bl -column "BPF_ST" "M[k] <- A" -offset indent
.It Sy BPF_ST Ta M[k] <- A
.El
.It Sy BPF_STX
This instruction stores the index register in the scratch memory store.
.Bl -column "BPF_STX" "M[k] <- X" -offset indent
.It Sy BPF_STX Ta M[k] <- X
.El
.It Sy BPF_ALU
The alu instructions perform operations between the accumulator and
index register or constant, and store the result back in the accumulator.
For binary operations, a source mode is required
.Sy ( BPF_K
or
.Sy BPF_X ) .
.Bl -column "BPF_ALU_BPF_ADD_BPF_K" "A <- A + k" -offset indent
.It Sy BPF_ALU+BPF_ADD+BPF_K Ta A <- A + k
.It Sy BPF_ALU+BPF_SUB+BPF_K Ta A <- A - k
.It Sy BPF_ALU+BPF_MUL+BPF_K Ta A <- A * k
.It Sy BPF_ALU+BPF_DIV+BPF_K Ta A <- A / k
.It Sy BPF_ALU+BPF_AND+BPF_K Ta A <- A & k
.It Sy BPF_ALU+BPF_OR+BPF_K Ta A <- A | k
.It Sy BPF_ALU+BPF_LSH+BPF_K Ta A <- A << k
.It Sy BPF_ALU+BPF_RSH+BPF_K Ta A <- A >> k
.It Sy BPF_ALU+BPF_ADD+BPF_X Ta A <- A + X
.It Sy BPF_ALU+BPF_SUB+BPF_X Ta A <- A - X
.It Sy BPF_ALU+BPF_MUL+BPF_X Ta A <- A * X
.It Sy BPF_ALU+BPF_DIV+BPF_X Ta A <- A / X
.It Sy BPF_ALU+BPF_AND+BPF_X Ta A <- A & X
.It Sy BPF_ALU+BPF_OR+BPF_X Ta A <- A | X
.It Sy BPF_ALU+BPF_LSH+BPF_X Ta A <- A << X
.It Sy BPF_ALU+BPF_RSH+BPF_X Ta A <- A >> X
.It Sy BPF_ALU+BPF_NEG Ta A <- -A
.El
.It Sy BPF_JMP
The jump instructions alter flow of control.
Conditional jumps compare the accumulator against a constant
.Sy ( BPF_K )
or the index register
.Sy ( BPF_X ) .
If the result is true (or non-zero),
the true branch is taken, otherwise the false branch is taken.
Jump offsets are encoded in 8 bits so the longest jump is 256 instructions.
However, the jump always
.Sy ( BPF_JA )
opcode uses the 32 bit
.Va k
field as the offset, allowing arbitrarily distant destinations.
All conditionals use unsigned comparison conventions.
.Bl -column "BPF_JMP+BPF_JSET+BPF_K" "pc += (A \*[Ge] k) ? jt : jf" -offset indent
.It Sy BPF_JMP+BPF_JA Ta pc += k
.It Sy BPF_JMP+BPF_JGT+BPF_K Ta "pc += (A > k) ? jt : jf"
.It Sy BPF_JMP+BPF_JGE+BPF_K Ta "pc += (A \*[Ge] k) ? jt : jf"
.It Sy BPF_JMP+BPF_JEQ+BPF_K Ta "pc += (A == k) ? jt : jf"
.It Sy BPF_JMP+BPF_JSET+BPF_K Ta "pc += (A & k) ? jt : jf"
.It Sy BPF_JMP+BPF_JGT+BPF_X Ta "pc += (A > X) ? jt : jf"
.It Sy BPF_JMP+BPF_JGE+BPF_X Ta "pc += (A \*[Ge] X) ? jt : jf"
.It Sy BPF_JMP+BPF_JEQ+BPF_X Ta "pc += (A == X) ? jt : jf"
.It Sy BPF_JMP+BPF_JSET+BPF_X Ta "pc += (A & X) ? jt : jf"
.El
.It Sy BPF_RET
The return instructions terminate the filter program and specify the amount
of packet to accept (i.e., they return the truncation amount).
A return value of zero indicates that the packet should be ignored.
The return value is either a constant
.Sy ( BPF_K )
or the accumulator
.Sy ( BPF_A ) .
.Bl -column "BPF_RET+BPF_A" "accept A bytes" -offset indent
.It Sy BPF_RET+BPF_A Ta accept A bytes
.It Sy BPF_RET+BPF_K Ta accept k bytes
.El
.It Sy BPF_MISC
The miscellaneous category was created for anything that doesn't
fit into the above classes, and for any new instructions that might need to
be added.
Currently, these are the register transfer instructions
that copy the index register to the accumulator or vice versa.
.Bl -column "BPF_MISC+BPF_TAX" "X <- A" -offset indent
.It Sy BPF_MISC+BPF_TAX Ta X <- A
.It Sy BPF_MISC+BPF_TXA Ta A <- X
.El
.Pp
Also, two instructions to call a "coprocessor" if initialized by the kernel
component.
There is no coprocessor by default.
.Bl -column "BPF_MISC+BPF_COPX" "A <- funcs[X](...)" -offset indent
.It Sy BPF_MISC+BPF_COP Ta A <- funcs[k](..)
.It Sy BPF_MISC+BPF_COPX Ta A <- funcs[X](..)
.El
.Pp
If the coprocessor is not set or the function index is out of range, these
instructions will abort the program and return zero.
.El
.Pp
The BPF interface provides the following macros to facilitate
array initializers:
.Bd -unfilled -offset indent
.Fn BPF_STMT opcode operand
.Fn BPF_JUMP opcode operand true_offset false_offset
.Ed
.Sh SYSCTLS
The following sysctls are available when
.Nm
is enabled:
.Bl -tag -width "XnetXbpfXmaxbufsizeXX"
.It Li net.bpf.maxbufsize
Sets the maximum buffer size available for
.Nm
peers.
.It Li net.bpf.stats
Shows
.Nm
statistics.
They can be retrieved with the
.Xr netstat 1
utility.
.It Li net.bpf.peers
Shows the current
.Nm
peers.
This is only available to the super user and can also be retrieved with the
.Xr netstat 1
utility.
.El
.Pp
On architectures with
.Xr bpfjit 4
support, the additional sysctl is available:
.Bl -tag -width "XnetXbpfXjitXX"
.It Li net.bpf.jit
Toggle
.Sy Just-In-Time
compilation of new filter programs.
In order to enable Just-In-Time compilation,
the bpfjit kernel module must be loaded.
Changing a value of this sysctl doesn't affect
existing filter programs.
.El
.Sh FILES
.Pa /dev/bpf
.Sh EXAMPLES
The following filter is taken from the Reverse ARP Daemon.
It accepts only Reverse ARP requests.
.Bd -literal -offset indent
struct bpf_insn insns[] = {
	BPF_STMT(BPF_LD+BPF_H+BPF_ABS, 12),
	BPF_JUMP(BPF_JMP+BPF_JEQ+BPF_K, ETHERTYPE_REVARP, 0, 3),
	BPF_STMT(BPF_LD+BPF_H+BPF_ABS, 20),
	BPF_JUMP(BPF_JMP+BPF_JEQ+BPF_K, REVARP_REQUEST, 0, 1),
	BPF_STMT(BPF_RET+BPF_K, sizeof(struct ether_arp) +
	    sizeof(struct ether_header)),
	BPF_STMT(BPF_RET+BPF_K, 0),
};
.Ed
.Pp
This filter accepts only IP packets between host 128.3.112.15 and
128.3.112.35.
.Bd -literal -offset indent
struct bpf_insn insns[] = {
	BPF_STMT(BPF_LD+BPF_H+BPF_ABS, 12),
	BPF_JUMP(BPF_JMP+BPF_JEQ+BPF_K, ETHERTYPE_IP, 0, 8),
	BPF_STMT(BPF_LD+BPF_W+BPF_ABS, 26),
	BPF_JUMP(BPF_JMP+BPF_JEQ+BPF_K, 0x8003700f, 0, 2),
	BPF_STMT(BPF_LD+BPF_W+BPF_ABS, 30),
	BPF_JUMP(BPF_JMP+BPF_JEQ+BPF_K, 0x80037023, 3, 4),
	BPF_JUMP(BPF_JMP+BPF_JEQ+BPF_K, 0x80037023, 0, 3),
	BPF_STMT(BPF_LD+BPF_W+BPF_ABS, 30),
	BPF_JUMP(BPF_JMP+BPF_JEQ+BPF_K, 0x8003700f, 0, 1),
	BPF_STMT(BPF_RET+BPF_K, (u_int)-1),
	BPF_STMT(BPF_RET+BPF_K, 0),
};
.Ed
.Pp
Finally, this filter returns only TCP finger packets.
We must parse the IP header to reach the TCP header.
The
.Sy BPF_JSET
instruction checks that the IP fragment offset is 0 so we are sure
that we have a TCP header.
.Bd -literal -offset indent
struct bpf_insn insns[] = {
	BPF_STMT(BPF_LD+BPF_H+BPF_ABS, 12),
	BPF_JUMP(BPF_JMP+BPF_JEQ+BPF_K, ETHERTYPE_IP, 0, 10),
	BPF_STMT(BPF_LD+BPF_B+BPF_ABS, 23),
	BPF_JUMP(BPF_JMP+BPF_JEQ+BPF_K, IPPROTO_TCP, 0, 8),
	BPF_STMT(BPF_LD+BPF_H+BPF_ABS, 20),
	BPF_JUMP(BPF_JMP+BPF_JSET+BPF_K, 0x1fff, 6, 0),
	BPF_STMT(BPF_LDX+BPF_B+BPF_MSH, 14),
	BPF_STMT(BPF_LD+BPF_H+BPF_IND, 14),
	BPF_JUMP(BPF_JMP+BPF_JEQ+BPF_K, 79, 2, 0),
	BPF_STMT(BPF_LD+BPF_H+BPF_IND, 16),
	BPF_JUMP(BPF_JMP+BPF_JEQ+BPF_K, 79, 0, 1),
	BPF_STMT(BPF_RET+BPF_K, (u_int)-1),
	BPF_STMT(BPF_RET+BPF_K, 0),
};
.Ed
.Sh SEE ALSO
.Xr ioctl 2 ,
.Xr read 2 ,
.Xr select 2 ,
.Xr signal 3 ,
.Xr bpfjit 4 ,
.Xr tcpdump 8
.Rs
.%T "The BSD Packet Filter: A New Architecture for User-level Packet Capture"
.%A S. McCanne
.%A V. Jacobson
.%J Proceedings of the 1993 Winter USENIX
.%C Technical Conference, San Diego, CA
.Re
.Sh HISTORY
The Enet packet filter was created in 1980 by Mike Accetta and
Rick Rashid at Carnegie-Mellon University.
Jeffrey Mogul, at Stanford, ported the code to BSD and continued
its development from 1983 on.
Since then, it has evolved into the ULTRIX Packet Filter
at DEC, a STREAMS NIT module under SunOS 4.1, and BPF.
.Sh AUTHORS
.An -nosplit
.An Steven McCanne ,
of Lawrence Berkeley Laboratory, implemented BPF in Summer 1990.
The design was in collaboration with
.An Van Jacobson ,
also of Lawrence Berkeley Laboratory.
.Sh BUGS
The read buffer must be of a fixed size (returned by the
.Dv BIOCGBLEN
ioctl).
.Pp
A file that does not request promiscuous mode may receive promiscuously
received packets as a side effect of another file requesting this
mode on the same hardware interface.
This could be fixed in the kernel with additional processing overhead.
However, we favor the model where
all files must assume that the interface is promiscuous, and if
so desired, must use a filter to reject foreign packets.
.Pp
Under SunOS, if a BPF application reads more than 2^31 bytes of
data, read will fail in
.Er EINVAL .
You can either fix the bug in SunOS,
or lseek to 0 when read fails for this reason.
.Pp
.Dq Immediate mode
and the
.Dq read timeout
are misguided features.
This functionality can be emulated with non-blocking mode and
.Xr select 2 .