MAN page from Fedora 5 libpcap-0.9.4-4.fc5.x86_64.rpm
Section: C Library Functions (3)
Updated: 27 February 2004Index
pcap - Packet Capture library
#include <pcap.h>char errbuf[PCAP_ERRBUF_SIZE];pcap_t *pcap_open_live(const char *device, int snaplen,
int promisc, int to_ms, char *errbuf)pcap_t *pcap_open_dead(int linktype, int snaplen)pcap_t *pcap_open_offline(const char *fname, char *errbuf)pcap_t *pcap_fopen_offline(FILE *fp, char *errbuf)pcap_dumper_t *pcap_dump_open(pcap_t *p, const char *fname)pcap_dumper_t *pcap_dump_fopen(pcap_t *p, FILE *fp)int pcap_setnonblock(pcap_t *p, int nonblock, char *errbuf);int pcap_getnonblock(pcap_t *p, char *errbuf);int pcap_findalldevs(pcap_if_t **alldevsp, char *errbuf)void pcap_freealldevs(pcap_if_t *alldevs)char *pcap_lookupdev(char *errbuf)int pcap_lookupnet(const char *device, bpf_u_int32 *netp,
bpf_u_int32 *maskp, char *errbuf)typedef void (*pcap_handler)(u_char *user, const struct pcap_pkthdr *h,
const u_char *bytes);int pcap_dispatch(pcap_t *p, int cnt,
pcap_handler callback, u_char *user)int pcap_loop(pcap_t *p, int cnt,
pcap_handler callback, u_char *user)void pcap_dump(u_char *user, struct pcap_pkthdr *h,
u_char *sp)int pcap_compile(pcap_t *p, struct bpf_program *fp,
char *str, int optimize, bpf_u_int32 netmask)int pcap_setfilter(pcap_t *p, struct bpf_program *fp)void pcap_freecode(struct bpf_program *)int pcap_setdirection(pcap_t *p, pcap_direction_t d)const u_char *pcap_next(pcap_t *p, struct pcap_pkthdr *h)int pcap_next_ex(pcap_t *p, struct pcap_pkthdr **pkt_header,
const u_char **pkt_data)void pcap_breakloop(pcap_t *)int pcap_inject(pcap_t *p, const void *buf, size_t size)int pcap_sendpacket(pcap_t *p, const u_char *buf, int size)int pcap_datalink(pcap_t *p)int pcap_list_datalinks(pcap_t *p, int **dlt_buf);int pcap_set_datalink(pcap_t *p, int dlt);int pcap_datalink_name_to_val(const char *name);const char *pcap_datalink_val_to_name(int dlt);const char *pcap_datalink_val_to_description(int dlt);int pcap_snapshot(pcap_t *p)int pcap_is_swapped(pcap_t *p)int pcap_major_version(pcap_t *p)int pcap_minor_version(pcap_t *p)int pcap_stats(pcap_t *p, struct pcap_stat *ps)FILE *pcap_file(pcap_t *p)int pcap_fileno(pcap_t *p)int pcap_get_selectable_fd(pcap_t *p);void pcap_perror(pcap_t *p, char *prefix)char *pcap_geterr(pcap_t *p)char *pcap_strerror(int error)const char *pcap_lib_version(void)void pcap_close(pcap_t *p)int pcap_dump_flush(pcap_dumper_t *p)long pcap_dump_ftell(pcap_dumper_t *p)FILE *pcap_dump_file(pcap_dumper_t *p)void pcap_dump_close(pcap_dumper_t *p)
The Packet Capture libraryprovides a high level interface to packet capture systems. All packetson the network, even those destined for other hosts, are accessiblethrough this mechanism.
is assumed to be able to hold at leastPCAP_ERRBUF_SIZE
pcap_open_live()is used to obtain a packet capture descriptor to lookat packets on the network.deviceis a string that specifies the network device to open; on Linux systemswith 2.2 or later kernels, adeviceargument of "any" orNULLcan be used to capture packets from all interfaces.snaplenspecifies the maximum number of bytes to capture. If this value is lessthan the size of a packet that is captured, only the firstsnaplenbytes of that packet will be captured and provided as packet data. Avalue of 65535 should be sufficient, on most if not all networks, tocapture all the data available from the packet.promiscspecifies if the interface is to be put into promiscuous mode.(Note that even if this parameter is false, the interfacecould well be in promiscuous mode for some other reason.) For now, thisdoesn't work on the "any" device; if an argument of "any" or NULL issupplied, thepromiscflag is ignored.to_msspecifies the read timeout in milliseconds. The read timeout is used toarrange that the read not necessarily return immediately when a packetis seen, but that it wait for some amount of time to allow more packetsto arrive and to read multiple packets from the OS kernel in oneoperation. Not all platforms support a read timeout; on platforms thatdon't, the read timeout is ignored. A zero value forto_ms,on platforms that support a read timeout,will cause a read to wait forever to allow enough packets toarrive, with no timeout.errbufis used to return error or warning text. It will be set to error text whenpcap_open_live()fails and returnsNULL.errbufmay also be set to warning text whenpcap_open_live()succeds; to detect this case the caller should store a zero-length string inerrbufbefore callingpcap_open_live()and display the warning to the user iferrbufis no longer a zero-length string.
pcap_open_dead()is used for creating apcap_tstructure to use when calling the other functions in libpcap. It istypically used when just using libpcap for compiling BPF code.
pcap_open_offline()is called to open a ``savefile'' for reading.fnamespecifies the name of the file to open. The file hasthe same format as those used bytcpdump(8)andtcpslice(8).The name "-" in a synonym forstdin.Alternatively, you may callpcap_fopen_offline()to read dumped data from an existing open streamfp.Note that on Windows, that stream should be opened in binary mode.errbufis used to return error text and is only set whenpcap_open_offline()orpcap_fopen_offline()fails and returnsNULL.
pcap_dump_open()is called to open a ``savefile'' for writing. The name "-" in a synonymforstdout.NULLis returned on failure.pis apcapstruct as returned bypcap_open_offline()orpcap_open_live().fnamespecifies the name of the file to open. Alternatively, you may callpcap_dump_fopen()to write data to an existing open streamfp.Note that on Windows, that stream should be opened in binary mode.IfNULLis returned,pcap_geterr()can be used to get the error text.
pcap_setnonblock()puts a capture descriptor, opened withpcap_open_live(),into ``non-blocking'' mode, or takes it out of ``non-blocking'' mode,depending on whether thenonblockargument is non-zero or zero. It has no effect on ``savefiles''.If there is an error, -1 is returned anderrbufis filled in with an appropriate error message; otherwise, 0 isreturned.In``non-blocking'' mode, an attempt to read from the capture descriptorwithpcap_dispatch()will, if no packets are currently available to be read, return 0immediately rather than blocking waiting for packets to arrive.pcap_loop()andpcap_next()will not work in ``non-blocking'' mode.
pcap_getnonblock()returns the current ``non-blocking'' state of the capture descriptor; italways returns 0 on ``savefiles''.If there is an error, -1 is returned anderrbufis filled in with an appropriate error message.
pcap_findalldevs()constructs a list of network devices that can be opened withpcap_open_live().(Note that there may be network devices that cannot be opened withpcap_open_live()by theprocess callingpcap_findalldevs(),because, for example, that process might not have sufficient privilegesto open them for capturing; if so, those devices will not appear on thelist.)alldevspis set to point to the first element of the list; each element of thelist is of typepcap_if_t,and has the following members:
- if notNULL,a pointer to the next element in the list;NULLfor the last element of the list
- a pointer to a string giving a name for the device to pass topcap_open_live()
- if notNULL,a pointer to a string giving a human-readable description of the device
- a pointer to the first element of a list of addresses for the interface
- interface flags:
- set if the interface is a loopback interface
Each element of the list of addresses is of typepcap_addr_t,and has the following members:
- if notNULL,a pointer to the next element in the list;NULLfor the last element of the list
- a pointer to astruct sockaddrcontaining an address
- if notNULL,a pointer to astruct sockaddrthat contains the netmask corresponding to the address pointed to byaddr
- if notNULL,a pointer to astruct sockaddrthat contains the broadcast address corresponding to the address pointedto byaddr;may be null if the interface doesn't support broadcasts
- if notNULL,a pointer to astruct sockaddrthat contains the destination address corresponding to the address pointedto byaddr;may be null if the interface isn't a point-to-point interface
Note that not all the addresses in the list of addresses arenecessarily IPv4 or IPv6 addresses - you must check thesa_familymember of thestruct sockaddrbefore interpreting the contents of the address.
-1is returned on failure, in which caseerrbufis filled in with an appropriate error message;0is returned on success.
pcap_freealldevs()is used to free a list allocated bypcap_findalldevs().
pcap_lookupdev()returns a pointer to a network device suitable for use withpcap_open_live()andpcap_lookupnet().If there is an error,NULLis returned anderrbufis filled in with an appropriate error message.
pcap_lookupnet()is used to determine the network number and maskassociated with the network devicedevice.Bothnetpandmaskparebpf_u_int32pointers.A return of -1 indicates an error in which caseerrbufis filled in with an appropriate error message.
pcap_dispatch()is used to collect and process packets.cntspecifies the maximum number of packets to process before returning.This is not a minimum number; when reading a live capture, only onebufferful of packets is read at a time, so fewer thancntpackets may be processed. Acntof -1 processes all the packets received in one buffer when reading alive capture, or all the packets in the file when reading a``savefile''.callbackspecifies a routine to be called with three arguments:au_charpointer which is passed in frompcap_dispatch(),aconst struct pcap_pkthdrpointer to a structure with the following members:
- astruct timevalcontaining the time when the packet was captured
- abpf_u_int32giving the number of bytes of the packet that are available from thecapture
- abpf_u_int32giving the length of the packet, in bytes (which might be more than thenumber of bytes available from the capture, if the length of the packetis larger than the maximum number of bytes to capture)
and aconst u_charpointer to the firstcaplen(as given in thestruct pcap_pkthdra pointer to which is passed to the callback routine)bytes of data from the packet (which won't necessarily be the entirepacket; to capture the entire packet, you will have to provide a valueforsnaplenin your call topcap_open_live()that is sufficiently large to get all of the packet's data - a value of65535 should be sufficient on most if not all networks).
The number of packets read is returned.0 is returned if no packets were read from a live capture (if, forexample, they were discarded because they didn't pass the packet filter,or if, on platforms that support a read timeout that starts before anypackets arrive, the timeout expires before any packets arrive, or if thefile descriptor for the capture device is in non-blocking mode and nopackets were available to be read) or if no more packets are availablein a ``savefile.'' A return of -1 indicatesan error in which casepcap_perror()orpcap_geterr()may be used to display the error text.A return of -2 indicates that the loop terminated due to a call topcap_breakloop()before any packets were processed.If your application uses pcap_breakloop(),make sure that you explicitly check for -1 and -2, rather than justchecking for a return value < 0.
NOTE:when reading a live capture,pcap_dispatch()will not necessarily return when the read times out; on some platforms,the read timeout isn't supported, and, on other platforms, the timerdoesn't start until at least one packet arrives. This means that theread timeout shouldNOTbe used in, for example, an interactive application, to allow the packetcapture loop to ``poll'' for user input periodically, as there's noguarantee thatpcap_dispatch()will return after the timeout expires.
pcap_loop()is similar topcap_dispatch()except it keeps reading packets untilcntpackets are processed or an error occurs.It doesnotreturn when live read timeouts occur.Rather, specifying a non-zero read timeout topcap_open_live()and then callingpcap_dispatch()allows the reception and processing of any packets that arrive when thetimeout occurs.A negativecntcausespcap_loop()to loop forever (or at least until an error occurs). -1 is returned onan error; 0 is returned ifcntis exhausted; -2 is returned if the loop terminated due to a call topcap_breakloop()before any packets were processed.If your application uses pcap_breakloop(),make sure that you explicitly check for -1 and -2, rather than justchecking for a return value < 0.
pcap_next()reads the next packet (by callingpcap_dispatch()with acntof 1) and returns au_charpointer to the data in that packet. (Thepcap_pkthdrstruct for that packet is not supplied.)NULLis returned if an error occured, or if no packets were read from a livecapture (if, for example, they were discarded because they didn't passthe packet filter, or if, on platforms that support a read timeout thatstarts before any packets arrive, the timeout expires before any packetsarrive, or if the file descriptor for the capture device is innon-blocking mode and no packets were available to be read), or if nomore packets are available in a ``savefile.'' Unfortunately, there isno way to determine whether an error occured or not.
pcap_next_ex()reads the next packet and returns a success/failure indication:
- the packet was read without problems
- packets are being read from a live capture, and the timeout expired
- an error occurred while reading the packet
- packets are being read from a ``savefile'', and there are no morepackets to read from the savefile.
If the packet was read without problems, the pointer pointed to by thepkt_headerargument is set to point to thepcap_pkthdrstruct for the packet, and thepointer pointed to by thepkt_dataargument is set to point to the data in the packet.
pcap_breakloop()sets a flag that will forcepcap_dispatch()orpcap_loop()to return rather than looping; they will return the number of packetsthat have been processed so far, or -2 if no packets have beenprocessed so far.
This routine is safe to use inside a signal handler on UNIX or a consolecontrol handler on Windows, as it merely sets a flag that is checkedwithin the loop.
The flag is checked in loops reading packets from the OS - a signal byitself will not necessarily terminate those loops - as well as in loopsprocessing a set of packets returned by the OS.Note that if you are catching signals on UNIX systems that supportrestarting system calls after a signal, and calling pcap_breakloop()in the signal handler, you must specify, when catching those signals,that system calls should NOT be restarted by that signal. Otherwise,if the signal interrupted a call reading packets in a live capture,when your signal handler returns after calling pcap_breakloop(), thecall will be restarted, and the loop will not terminate until morepackets arrive and the call completes.
Note also that, in a multi-threaded application, if one thread isblocked inpcap_dispatch(),pcap_loop(),pcap_next(),orpcap_next_ex(),a call topcap_breakloop()in a different thread will not unblock that thread; you will need to usewhatever mechanism the OS provides for breaking a thread out of blockingcalls in order to unblock the thread, such as thread cancellation insystems that support POSIX threads.
Note thatpcap_next()will, on some platforms, loop reading packets from the OS; that loopwill not necessarily be terminated by a signal, sopcap_breakloop()should be used to terminate packet processing even ifpcap_next()is being used.
pcap_breakloop()does not guarantee that no further packets will be processed bypcap_dispatch()orpcap_loop()after it is called; at most one more packet might be processed.
If -2 is returned frompcap_dispatch()orpcap_loop(),the flag is cleared, so a subsequent call will resume reading packets. If a positive number is returned, the flag is not cleared, so asubsequent call will return -2 and clear the flag.
pcap_inject()sends a raw packet through the network interface;bufpoints to the data of the packet, including the link-layer header, andsizeis the number of bytes in the packet.It returns the number of bytes written on success. A return of -1indicates an error in which casepcap_perror()orpcap_geterr()may be used to display the error text.Note that, even if you successfully open the network interface, youmight not have permission to send packets on it, or it might not supportsending packets; aspcap_open_live()doesn't have a flag to indicate whether to open for capturing, sending,or capturing and sending, you cannot request an open that supportssending and be notified at open time whether sending will be possible.Note also that some devices might not support sending packets.
Note that, on some platforms, the link-layer header of the packet that'ssent might not be the same as the link-layer header of the packetsupplied topcap_inject(),as the source link-layer address, if the header contains such anaddress, might be changed to be the address assigned to the interface onwhich the packet it sent, if the platform doesn't support sendingcompletely raw and unchanged packets. Even worse, some drivers on someplatforms might change the link-layer type field to whatever valuelibpcap used when attaching to the device, even on platforms thatdonominally support sending completely raw and unchanged packets.
pcap_sendpacket()is likepcap_inject(),but it returns 0 on success and -1 on failure.(pcap_inject()comes from OpenBSD;pcap_sendpacket()comes from WinPcap. Both are provided for compatibility.)
pcap_dump()outputs a packet to the ``savefile'' opened withpcap_dump_open().Note that its calling arguments are suitable for use withpcap_dispatch()orpcap_loop().If called directly, the userparameter is of type pcap_dumper_tas returned bypcap_dump_open().
pcap_compile()is used to compile the stringstrinto a filter program.programis a pointer to abpf_programstruct and is filled in bypcap_compile().optimizecontrols whether optimization on the resulting code is performed.netmaskspecifies the IPv4 netmask of the network on which packets are beingcaptured; it is used only when checking for IPv4 broadcast addresses inthe filter program. If the netmask of the network on which packets arebeing captured isn't known to the program, or if packets are beingcaptured on the Linux "any" pseudo-interface that can capture on morethan one network, a value of 0 can be supplied; tests for IPv4 broadcastaddreses won't be done correctly, but all other tests in the filterprogram will be OK. A return of -1 indicates an error in which casepcap_geterr()may be used to display the error text.
pcap_compile_nopcap()is similar topcap_compile()except that instead of passing a pcap structure, one passes thesnaplen and linktype explicitly. It is intended to be used forcompiling filters for direct BPF usage, without necessarily havingcalledpcap_open().A return of -1 indicates an error; the error text is unavailable.(pcap_compile_nopcap()is a wrapper aroundpcap_open_dead(),pcap_compile(),andpcap_close();the latter three routines can be used directly in order to get the errortext for a compilation error.)
pcap_setfilter()is used to specify a filter program.fpis a pointer to abpf_programstruct, usually the result of a call topcap_compile().-1is returned on failure, in which casepcap_geterr()may be used to display the error text;0is returned on success.
pcap_freecode()is used to free up allocated memory pointed to by abpf_programstruct generated bypcap_compile()when that BPF program is no longer needed, for example after ithas been made the filter program for a pcap structure by a call topcap_setfilter().
pcap_setdirection()is used to specify a direction that packets will be captured.pcap_direction_t is one of the constantsPCAP_D_IN,PCAP_D_OUTor PCAP_D_INOUT.PCAP_D_INwill only capture packets received by the device,PCAP_D_OUTwill only capture packets sent by the device andPCAP_D_INOUTwill capture packets received by or sent by the device.PCAP_D_INOUTis the default setting if this function is not called. This isn'tnecessarily supported on all platforms; some platforms might return anerror, and some other platforms might not supportPCAP_D_OUT.This operation is not supported if a ``savefile'' is being read.-1is returned on failure,0is returned on success.
pcap_datalink()returns the link layer type; link layer types it can return include:
- BSD loopback encapsulation; the link layer header is a 4-byte field, inhostbyte order, containing a PF_ value fromsocket.hfor the network-layer protocol of the packet.
- Note that ``host byte order'' is the byte order of the machine on whichthe packets are captured, and the PF_ values are for the OS of themachine on which the packets are captured; if a live capture is beingdone, ``host byte order'' is the byte order of the machine capturing thepackets, and the PF_ values are those of the OS of the machine capturingthe packets, but if a ``savefile'' is being read, the byte order and PF_values arenotnecessarily those of the machine reading the capture file.
- Ethernet (10Mb, 100Mb, 1000Mb, and up)
- IEEE 802.5 Token Ring
- SLIP; the link layer header contains, in order:
a 1-byte flag, which is 0 for packets received by the machine and 1 forpackets sent by the machine;
a 1-byte field, the upper 4 bits of which indicate the type of packet,as per RFC 1144:
- an unmodified IP datagram (TYPE_IP);
- an uncompressed-TCP IP datagram (UNCOMPRESSED_TCP), with that byte beingthe first byte of the raw IP header on the wire, containing theconnection number in the protocol field;
- a compressed-TCP IP datagram (COMPRESSED_TCP), with that byte being thefirst byte of the compressed TCP/IP datagram header;
for UNCOMPRESSED_TCP, the rest of the modified IP header, and forCOMPRESSED_TCP, the compressed TCP/IP datagram header;
for a total of 16 bytes; the uncompressed IP datagram follows the header.
- PPP; if the first 2 bytes are 0xff and 0x03, it's PPP in HDLC-likeframing, with the PPP header following those two bytes, otherwise it'sPPP without framing, and the packet begins with the PPP header.
- RFC 1483 LLC/SNAP-encapsulated ATM; the packet begins with an IEEE 802.2LLC header.
- raw IP; the packet begins with an IP header.
- PPP in HDLC-like framing, as per RFC 1662, or Cisco PPP with HDLCframing, as per section 4.3.1 of RFC 1547; the first byte will be 0xFFfor PPP in HDLC-like framing, and will be 0x0F or 0x8F for Cisco PPPwith HDLC framing.
- PPPoE; the packet begins with a PPPoE header, as per RFC 2516.
- Cisco PPP with HDLC framing, as per section 4.3.1 of RFC 1547.
- IEEE 802.11 wireless LAN
- Frame Relay
- OpenBSD loopback encapsulation; the link layer header is a 4-byte field, innetworkbyte order, containing a PF_ value from OpenBSD'ssocket.hfor the network-layer protocol of the packet.
- Note that, if a ``savefile'' is being read, those PF_ values arenotnecessarily those of the machine reading the capture file.
- Linux "cooked" capture encapsulation; the link layer header contains, inorder:
a 2-byte "packet type", in network byte order, which is one of:
- packet was sent to us by somebody else
- packet was broadcast by somebody else
- packet was multicast, but not broadcast, by somebody else
- packet was sent by somebody else to somebody else
- packet was sent by us
a 2-byte field, in network byte order, containing a Linux ARPHRD_ valuefor the link layer device type;
a 2-byte field, in network byte order, containing the length of thelink layer address of the sender of the packet (which could be 0);
an 8-byte field containing that number of bytes of the link layer header(if there are more than 8 bytes, only the first 8 are present);
a 2-byte field containing an Ethernet protocol type, in network byteorder, or containing 1 for Novell 802.3 frames without an 802.2 LLCheader or 4 for frames beginning with an 802.2 LLC header.
- Apple LocalTalk; the packet begins with an AppleTalk LLAP header.
- OpenBSD pflog; the link layer header contains, in order:
a 1-byte header length, in host byte order;
a 4-byte PF_ value, in host byte order;
a 2-byte action code, in network byte order, which is one of:
a 2-byte reason code, in network byte order, which is one of:
- bad offset
a 16-character interface name;
a 16-character ruleset name (only meaningful if subrule is set);
a 4-byte rule number, in network byte order;
a 4-byte subrule number, in network byte order;
a 1-byte direction, in network byte order, which is one of:
- incoming or outgoing
- Prism monitor mode information followed by an 802.11 header.
- RFC 2625 IP-over-Fibre Channel, with the link-layer header being theNetwork_Header as described in that RFC.
- SunATM devices; the link layer header contains, in order:
a 1-byte flag field, containing a direction flag in the uppermost bit,which is set for packets transmitted by the machine and clear forpackets received by the machine, and a 4-byte traffic type in thelow-order 4 bits, which is one of:
- raw traffic
- LANE traffic
- LLC-encapsulated traffic
- MARS traffic
- IFMP traffic
- ILMI traffic
- Q.2931 traffic
a 1-byte VPI value;
a 2-byte VCI field, in network byte order.
- link-layer information followed by an 802.11 header - seehttp://www.shaftnet.org/~pizza/software/capturefrm.txt for a descriptionof the link-layer information.
- ARCNET, with no exception frames, reassembled packets rather than rawframes, and an extra 16-bit offset field between the destination hostand type bytes.
- Linux-IrDA packets, with aDLT_LINUX_SLLheader followed by the IrLAP header.
pcap_list_datalinks()is used to get a list of the supported data link types of the interfaceassociated with the pcap descriptor.pcap_list_datalinks()allocates an array to hold the list and sets*dlt_buf.The caller is responsible for freeing the array.-1is returned on failure;otherwise, the number of data link types in the array is returned.
pcap_set_datalink()is used to set the current data link type of the pcap descriptorto the type specified bydlt.-1is returned on failure.
pcap_datalink_name_to_val()translates a data link type name, which is aDLT_name with theDLT_removed, to the corresponding data link type value. The translationis case-insensitive. -1is returned on failure.
pcap_datalink_val_to_name()translates a data link type value to the corresponding data link typename. NULL is returned on failure.
pcap_datalink_val_to_description()translates a data link type value to a short description of that datalink type. NULL is returned on failure.
pcap_snapshot()returns the snapshot length specified whenpcap_open_live()was called.
pcap_is_swapped()returns true if the current ``savefile'' uses a different byte orderthan the current system.
pcap_major_version()returns the major number of the file format of the savefile;pcap_minor_version()returns the minor number of the file format of the savefile. Theversion number is stored in the header of the savefile.
pcap_file()returns the standard I/O stream of the ``savefile,'' if a ``savefile''was opened withpcap_open_offline(),or NULL, if a network device was opened withpcap_open_live().
pcap_stats()returns 0 and fills in apcap_statstruct. The values represent packet statistics from the start of therun to the time of the call. If there is an error or the underlyingpacket capture doesn't support packet statistics, -1 is returned andthe error text can be obtained withpcap_perror()orpcap_geterr().pcap_stats()is supported only on live captures, not on ``savefiles''; no statisticsare stored in ``savefiles'', so no statistics are available when readingfrom a ``savefile''.
pcap_fileno()returns the file descriptor number from which captured packets are read,if a network device was opened withpcap_open_live(),or -1, if a ``savefile'' was opened withpcap_open_offline().
pcap_get_selectable_fd()returns, on UNIX, a file descriptor number for a file descriptor onwhich one cando aselect()orpoll()to wait for it to be possible to read packets without blocking, if sucha descriptor exists, or -1, if no such descriptor exists. Some networkdevices opened withpcap_open_live()do not supportselect()orpoll()(for example, regular network devices on FreeBSD 4.3 and 4.4, and EndaceDAG devices), so -1 is returned for those devices.
Note that on most versions of most BSDs (including Mac OS X)select()andpoll()do not work correctly on BPF devices;pcap_get_selectable_fd()will return a file descriptor on most of those versions (the exceptionsbeing FreeBSD 4.3 and 4.4), a simpleselect()orpoll()will not return even after a timeout specified inpcap_open_live()expires. To work around this, an application that usesselect()orpoll()to wait for packets to arrive must put thepcap_tin non-blocking mode, and must arrange that theselect()orpoll()have a timeout less than or equal to the timeout specified inpcap_open_live(),and must try to read packets after that timeout expires, regardless ofwhetherselect()orpoll()indicated that the file descriptor for thepcap_tis ready to be read or not. (That workaround will not work in FreeBSD4.3 and later; however, in FreeBSD 4.6 and later,select()andpoll()work correctly on BPF devices, so the workaround isn't necessary,although it does no harm.)
pcap_get_selectable_fd()is not available on Windows.
pcap_perror()prints the text of the last pcap library error onstderr,prefixed byprefix.
pcap_geterr()returns the error text pertaining to the last pcap library error.NOTE:the pointer it returns will no longer point to a valid error messagestring after thepcap_tpassed to it is closed; you must use or copy the string before closingthepcap_t.
pcap_strerror()is provided in casestrerror(1)isn't available.
pcap_lib_version()returns a pointer to a string giving information about the version ofthe libpcap library being used; note that it contains more informationthan just a version number.
pcap_close()closes the files associated withpand deallocates resources.
pcap_dump_file()returns the standard I/O stream of the ``savefile'' opened bypcap_dump_open().
pcap_dump_flush()flushes the output buffer to the ``savefile,'' so that any packetswritten withpcap_dump()but not yet written to the ``savefile'' will be written.-1is returned on error, 0 on success.
pcap_dump_ftell()returns the current file position for the ``savefile'', representing thenumber of bytes written bypcap_dump_open()andpcap_dump().-1is returned on error.
pcap_dump_close()closes the ``savefile.''
The original authors are:
Van Jacobson,Craig Leres andSteven McCanne, all of theLawrence Berkeley National Laboratory, University of California, Berkeley, CA.
The current version is available from "The Tcpdump Group"'s Web site at
Please send problems, bugs, questions, desirable enhancements, etc. to:
Please send source code contributions, etc. to:
- SEE ALSO
This document was created byman2html,using the manual pages.