Opening a capture handle for reading

To obtain a list of devices that can be opened for a live capture, call pcap_findalldevs(); to free the list returned by pcap_findalldevs(), call pcap_freealldevs(). pcap_lookupdev() will return the first device on that list that is not a ``loopback`` network interface.

To open a handle for a ``savefile'' from which to read packets, given the pathname of the ``savefile'', call pcap_open_offline(); to set up a handle for a ``savefile'', given a FILE * referring to a file already opened for reading, call pcap_fopen_offline().

In order to get a ``fake'' pcap_t for use in routines that require a pcap_t as an argument, such as routines to open a ``savefile'' for writing and to compile a filter expression, call pcap_open_dead().

pcap_create(), pcap_open_offline(), pcap_fopen_offline(), and pcap_open_dead() return a pointer to a pcap_t, which is the handle used for reading packets from the capture stream or the ``savefile'', and for finding out information about the capture stream or ``savefile''. To close a handle, use pcap_close().

The options that can be set on a capture handle include

snapshot length

If, when capturing, you capture the entire contents of the packet, that requires more CPU time to copy the packet to your application, more disk and possibly network bandwidth to write the packet data to a file, and more disk space to save the packet. If you don't need the entire contents of the packet - for example, if you are only interested in the TCP headers of packets - you can set the "snapshot length" for the capture to an appropriate value. If the snapshot length is set to snaplen, and snaplen is less than the size of a packet that is captured, only the first snaplen bytes of that packet will be captured and provided as packet data.

A snapshot length of 65535 should be sufficient, on most if not all networks, to capture all the data available from the packet.

The snapshot length is set with pcap_set_snaplen().

promiscuous mode

On broadcast LANs such as Ethernet, if the network isn't switched, or if the adapter is connected to a "mirror port" on a switch to which all packets passing through the switch are sent, a network adapter receives all packets on the LAN, including unicast or multicast packets not sent to a network address that the network adapter isn't configured to recognize.

Normally, the adapter will discard those packets; however, many network adapters support "promiscuous mode", which is a mode in which all packets, even if they are not sent to an address that the adapter recognizes, are provided to the host. This is useful for passively capturing traffic between two or more other hosts for analysis.

Note that even if an application does not set promiscuous mode, the adapter could well be in promiscuous mode for some other reason.

For now, this doesn't work on the "any" device; if an argument of "any" or NULL is supplied, the setting of promiscuous mode is ignored.

Promiscuous mode is set with pcap_set_promisc().

monitor mode

On IEEE 802.11 wireless LANs, even if an adapter is in promiscuous mode, it will supply to the host only frames for the network with which it's associated. It might also supply only data frames, not management or control frames, and might not provide the 802.11 header or radio information pseudo-header for those frames.

In "monitor mode", sometimes also called "rfmon mode" (for "Radio Frequency MONitor"), the adapter will supply all frames that it receives, with 802.11 headers, and might supply a pseudo-header with radio information about the frame as well.

Note that in monitor mode the adapter might disassociate from the network with which it's associated, so that you will not be able to use any wireless networks with that adapter. This could prevent accessing files on a network server, or resolving host names or network addresses, if you are capturing in monitor mode and are not connected to another network with another adapter.

Monitor mode is set with pcap_set_rfmon(), and pcap_can_set_rfmon() can be used to determine whether an adapter can be put into monitor mode.

packet buffer timeout

If, when capturing, packets are delivered as soon as they arrive, the application capturing the packets will be woken up for each packet as it arrives, and might have to make one or more calls to the operating system to fetch each packet.

If, instead, packets are not delivered as soon as they arrive, but are delivered after a short delay (called a "packet buffer timeout"), more than one packet can be accumulated before the packets are delivered, so that a single wakeup would be done for multiple packets, and each set of calls made to the operating system would supply multiple packets, rather than a single packet. This reduces the per-packet CPU overhead if packets are arriving at a high rate, increasing the number of packets per second that can be captured.

The packet buffer timeout is required so that an application won't wait for the operating system's capture buffer to fill up before packets are delivered; if packets are arriving slowly, that wait could take an arbitrarily long period of time.

Not all platforms support a packet buffer timeout; on platforms that don't, the packet buffer timeout is ignored. A zero value for the timeout, on platforms that support a packet buffer timeout, will cause a read to wait forever to allow enough packets to arrive, with no timeout. A negative value is invalid; the result of setting the timeout to a negative value is unpredictable.

NOTE: the packet buffer timeout cannot be used to cause calls that read packets to return within a limited period of time, because, on some platforms, the packet buffer timeout isn't supported, and, on other platforms, the timer doesn't start until at least one packet arrives. This means that the packet buffer timeout should NOT be used, for example, in an interactive application to allow the packet capture loop to ``poll'' for user input periodically, as there's no guarantee that a call reading packets will return after the timeout expires even if no packets have arrived.

The packet buffer timeout is set with pcap_set_timeout().

immediate mode

In immediate mode, packets are always delivered as soon as they arrive, with no buffering. Immediate mode is set with pcap_set_immediate_mode().

buffer size

Packets that arrive for a capture are stored in a buffer, so that they do not have to be read by the application as soon as they arrive. On some platforms, the buffer's size can be set; a size that's too small could mean that, if too many packets are being captured and the snapshot length doesn't limit the amount of data that's buffered, packets could be dropped if the buffer fills up before the application can read packets from it, while a size that's too large could use more non-pageable operating system memory than is necessary to prevent packets from being dropped.

The buffer size is set with pcap_set_buffer_size().

timestamp type

On some platforms, the time stamp given to packets on live captures can come from different sources that can have different resolutions or that can have different relationships to the time values for the current time supplied by routines on the native operating system. See pcap-tstamp(7) for a list of time stamp types.

The time stamp type is set with pcap_set_tstamp_type().

Reading packets from a network interface may require that you have special privileges:

Under SunOS 3.x or 4.x with NIT or BPF:

You must have read access to /dev/nit or /dev/bpf*.

Under Solaris with DLPI:

You must have read/write access to the network pseudo device, e.g. /dev/le. On at least some versions of Solaris, however, this is not sufficient to allow tcpdump to capture in promiscuous mode; on those versions of Solaris, you must be root, or the application capturing packets must be installed setuid to root, in order to capture in promiscuous mode. Note that, on many (perhaps all) interfaces, if you don't capture in promiscuous mode, you will not see any outgoing packets, so a capture not done in promiscuous mode may not be very useful.

In newer versions of Solaris, you must have been given the net_rawaccess privilege; this is both necessary and sufficient to give you access to the network pseudo-device - there is no need to change the privileges on that device. A user can be given that privilege by, for example, adding that privilege to the user's defaultpriv key with the usermod (8) command.

Under HP-UX with DLPI:

You must be root or the application capturing packets must be installed setuid to root.

Under IRIX with snoop:

You must be root or the application capturing packets must be installed setuid to root.

Under Linux:

You must be root or the application capturing packets must be installed setuid to root (unless your distribution has a kernel that supports capability bits such as CAP_NET_RAW and code to allow those capability bits to be given to particular accounts and to cause those bits to be set on a user's initial processes when they log in, in which case you must have CAP_NET_RAW in order to capture and CAP_NET_ADMIN to enumerate network devices with, for example, the -D flag).

Under ULTRIX and Digital UNIX/Tru64 UNIX:

Any user may capture network traffic. However, no user (not even the super-user) can capture in promiscuous mode on an interface unless the super-user has enabled promiscuous-mode operation on that interface using pfconfig(8), and no user (not even the super-user) can capture unicast traffic received by or sent by the machine on an interface unless the super-user has enabled copy-all-mode operation on that interface using pfconfig, so useful packet capture on an interface probably requires that either promiscuous-mode or copy-all-mode operation, or both modes of operation, be enabled on that interface.

Under BSD (this includes macOS):

You must have read access to /dev/bpf* on systems that don't have a cloning BPF device, or to /dev/bpf on systems that do. On BSDs with a devfs (this includes macOS), this might involve more than just having somebody with super-user access setting the ownership or permissions on the BPF devices - it might involve configuring devfs to set the ownership or permissions every time the system is booted, if the system even supports that; if it doesn't support that, you might have to find some other way to make that happen at boot time.

Reading a saved packet file doesn't require special privileges.

The packets read from the handle may include a ``pseudo-header'' containing various forms of packet meta-data, and probably includes a link-layer header whose contents can differ for different network interfaces. To determine the format of the packets supplied by the handle, call pcap_datalink(); https://www.tcpdump.org/linktypes.html lists the values it returns and describes the packet formats that correspond to those values.

Do NOT assume that the packets for a given capture or ``savefile`` will have any given link-layer header type, such as DLT_EN10MB for Ethernet. For example, the "any" device on Linux will have a link-layer header type of DLT_LINUX_SLL even if all devices on the system at the time the "any" device is opened have some other data link type, such as DLT_EN10MB for Ethernet.

To obtain the FILE * corresponding to a pcap_t opened for a ``savefile'', call pcap_file().

Routines

Selecting a link-layer header type for a live capture

Routines

Reading packets

The callback is also supplied a const u_char pointer to the first caplen (as given in the struct pcap_pkthdr mentioned above) bytes of data from the packet. This won't necessarily be the entire packet; to capture the entire packet, you will have to provide a value for snaplen in your call to pcap_set_snaplen() that is sufficiently large to get all of the packet's data - a value of 65535 should be sufficient on most if not all networks). When reading from a ``savefile'', the snapshot length specified when the capture was performed will limit the amount of packet data available.

pcap_next() is passed an argument that points to a struct pcap_pkthdr structure, and fills it in with the time stamp and length values for the packet. It returns a const u_char to the first caplen bytes of the packet on success, and NULL on error.

pcap_next_ex() is passed two pointer arguments, one of which points to a structpcap_pkthdr* and one of which points to a const u_char*. It sets the first pointer to point to a struct pcap_pkthdr structure with the time stamp and length values for the packet, and sets the second pointer to point to the first caplen bytes of the packet.

To force the loop in pcap_dispatch() or pcap_loop() to terminate, call pcap_breakloop().

By default, when reading packets from an interface opened for a live capture, pcap_dispatch(), pcap_next(), and pcap_next_ex() will, if no packets are currently available to be read, block waiting for packets to become available. On some, but not all, platforms, if a packet buffer timeout was specified, the wait will terminate after the packet buffer timeout expires; applications should be prepared for this, as it happens on some platforms, but should not rely on it, as it does not happen on other platforms. Note that the wait might, or might not, terminate even if no packets are available; applications should be prepared for this to happen, but must not rely on it happening.

A handle can be put into ``non-blocking mode'', so that those routines will, rather than blocking, return an indication that no packets are available to read. Call pcap_setnonblock() to put a handle into non-blocking mode or to take it out of non-blocking mode; call pcap_getnonblock() to determine whether a handle is in non-blocking mode. Note that non-blocking mode does not work correctly in Mac OS X 10.6.

Non-blocking mode is often combined with routines such as select(2) or poll(2) or other routines a platform offers to wait for any of a set of descriptors to be ready to read. To obtain, for a handle, a descriptor that can be used in those routines, call pcap_get_selectable_fd(). If the routine indicates that data is available to read on the descriptor, an attempt should be made to read from the device.

Not all handles have such a descriptor available; pcap_get_selectable_fd() will return PCAP_ERROR if no such descriptor is available. If no such descriptor is available, this may be because the device must be polled periodically for packets; in that case, pcap_get_required_select_timeout() will return a pointer to a struct timeval whose value can be used as a timeout in those routines. When the routine returns, an attmept should be made to read packets from the device. If pcap_get_required_select_timeout() returns NULL, no such timeout is available, and those routines cannot be used with the device.

In addition, for various reasons, one or more of those routines will not work properly with the descriptor; the documentation for pcap_get_selectable_fd() gives details. Note that, just as an attempt to read packets from a pcap_t may not return any packets if the packet buffer timeout expires, a select(), poll(), or other such call may, if the packet buffer timeout expires, indicate that a descriptor is ready to read even if there are no packets available to read.

Routines

Filters

A filter can be specified as a text string; the syntax and semantics of the string are as described by pcap-filter(7). A filter string is compiled into a program in a pseudo-machine-language by pcap_compile() and the resulting program can be made a filter for a handle with pcap_setfilter(). The result of pcap_compile() can be freed with a call to pcap_freecode(). pcap_compile() may require a network mask for certain expressions in the filter string; pcap_lookupnet() can be used to find the network address and network mask for a given capture device.

A compiled filter can also be applied directly to a packet that has been read using pcap_offline_filter().

Routines

Incoming and outgoing packets

Routines

Capture statistics

Routines

Opening a handle for writing captured packets

Routines

Writing packets

Routines

Injecting packets

Routines

Reporting errors

Routines

Getting library version information

Routines