mirror of
https://github.com/freebsd/freebsd-src.git
synced 2024-11-30 08:43:23 +00:00
6f9cba8f8b
Local changes: - In contrib/libpcap/pcap/bpf.h, do not include pcap/dlt.h. Our system net/dlt.h is pulled in from net/bpf.h. - sys/net/dlt.h: Incorporate changes from libpcap 1.10.3. - lib/libpcap/Makefile: Update for libpcap 1.10.3. Changelog: https://git.tcpdump.org/libpcap/blob/95691ebe7564afa3faa5c6ba0dbd17e351be455a:/CHANGES Reviewed by: emaste Obtained from: https://www.tcpdump.org/release/libpcap-1.10.3.tar.gz Sponsored by: The FreeBSD Foundation
513 lines
13 KiB
C
513 lines
13 KiB
C
/*
|
|
* Copyright (c) 1990, 1991, 1992, 1993, 1994, 1995, 1996
|
|
* 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.
|
|
*
|
|
* Modifications made to accommodate the new SunOS4.0 NIT facility by
|
|
* Micky Liu, micky@cunixc.cc.columbia.edu, Columbia University in May, 1989.
|
|
* This module now handles the STREAMS based NIT.
|
|
*/
|
|
|
|
#ifdef HAVE_CONFIG_H
|
|
#include <config.h>
|
|
#endif
|
|
|
|
#include <sys/types.h>
|
|
#include <sys/time.h>
|
|
#include <sys/timeb.h>
|
|
#include <sys/dir.h>
|
|
#include <sys/fcntlcom.h>
|
|
#include <sys/file.h>
|
|
#include <sys/ioctl.h>
|
|
#include <sys/socket.h>
|
|
#include <sys/stropts.h>
|
|
|
|
#include <net/if.h>
|
|
#include <net/nit.h>
|
|
#include <net/nit_if.h>
|
|
#include <net/nit_pf.h>
|
|
#include <net/nit_buf.h>
|
|
|
|
#include <netinet/in.h>
|
|
#include <netinet/in_systm.h>
|
|
#include <netinet/ip.h>
|
|
#include <netinet/if_ether.h>
|
|
#include <netinet/ip_var.h>
|
|
#include <netinet/udp.h>
|
|
#include <netinet/udp_var.h>
|
|
#include <netinet/tcp.h>
|
|
#include <netinet/tcpip.h>
|
|
|
|
#include <errno.h>
|
|
#include <stdio.h>
|
|
#include <string.h>
|
|
#include <unistd.h>
|
|
|
|
#include "pcap-int.h"
|
|
|
|
#ifdef HAVE_OS_PROTO_H
|
|
#include "os-proto.h"
|
|
#endif
|
|
|
|
/*
|
|
* The chunk size for NIT. This is the amount of buffering
|
|
* done for read calls.
|
|
*/
|
|
#define CHUNKSIZE (2*1024)
|
|
|
|
/*
|
|
* The total buffer space used by NIT.
|
|
*/
|
|
#define BUFSPACE (4*CHUNKSIZE)
|
|
|
|
/* Forwards */
|
|
static int nit_setflags(int, int, int, char *);
|
|
|
|
/*
|
|
* Private data for capturing on STREAMS NIT devices.
|
|
*/
|
|
struct pcap_snit {
|
|
struct pcap_stat stat;
|
|
};
|
|
|
|
static int
|
|
pcap_stats_snit(pcap_t *p, struct pcap_stat *ps)
|
|
{
|
|
struct pcap_snit *psn = p->priv;
|
|
|
|
/*
|
|
* "ps_recv" counts packets handed to the filter, not packets
|
|
* that passed the filter. As filtering is done in userland,
|
|
* this does not include packets dropped because we ran out
|
|
* of buffer space.
|
|
*
|
|
* "ps_drop" counts packets dropped inside the "/dev/nit"
|
|
* device because of flow control requirements or resource
|
|
* exhaustion; it doesn't count packets dropped by the
|
|
* interface driver, or packets dropped upstream. As filtering
|
|
* is done in userland, it counts packets regardless of whether
|
|
* they would've passed the filter.
|
|
*
|
|
* These statistics don't include packets not yet read from the
|
|
* kernel by libpcap or packets not yet read from libpcap by the
|
|
* application.
|
|
*/
|
|
*ps = psn->stat;
|
|
return (0);
|
|
}
|
|
|
|
static int
|
|
pcap_read_snit(pcap_t *p, int cnt, pcap_handler callback, u_char *user)
|
|
{
|
|
struct pcap_snit *psn = p->priv;
|
|
register int cc, n;
|
|
register u_char *bp, *cp, *ep;
|
|
register struct nit_bufhdr *hdrp;
|
|
register struct nit_iftime *ntp;
|
|
register struct nit_iflen *nlp;
|
|
register struct nit_ifdrops *ndp;
|
|
register int caplen;
|
|
|
|
cc = p->cc;
|
|
if (cc == 0) {
|
|
cc = read(p->fd, (char *)p->buffer, p->bufsize);
|
|
if (cc < 0) {
|
|
if (errno == EWOULDBLOCK)
|
|
return (0);
|
|
pcap_fmt_errmsg_for_errno(p->errbuf, sizeof(p->errbuf),
|
|
errno, "pcap_read");
|
|
return (-1);
|
|
}
|
|
bp = (u_char *)p->buffer;
|
|
} else
|
|
bp = p->bp;
|
|
|
|
/*
|
|
* loop through each snapshot in the chunk
|
|
*
|
|
* This assumes that a single buffer of packets will have
|
|
* <= INT_MAX packets, so the packet count doesn't overflow.
|
|
*/
|
|
n = 0;
|
|
ep = bp + cc;
|
|
while (bp < ep) {
|
|
/*
|
|
* Has "pcap_breakloop()" been called?
|
|
* If so, return immediately - if we haven't read any
|
|
* packets, clear the flag and return -2 to indicate
|
|
* that we were told to break out of the loop, otherwise
|
|
* leave the flag set, so that the *next* call will break
|
|
* out of the loop without having read any packets, and
|
|
* return the number of packets we've processed so far.
|
|
*/
|
|
if (p->break_loop) {
|
|
if (n == 0) {
|
|
p->break_loop = 0;
|
|
return (-2);
|
|
} else {
|
|
p->bp = bp;
|
|
p->cc = ep - bp;
|
|
return (n);
|
|
}
|
|
}
|
|
|
|
++psn->stat.ps_recv;
|
|
cp = bp;
|
|
|
|
/* get past NIT buffer */
|
|
hdrp = (struct nit_bufhdr *)cp;
|
|
cp += sizeof(*hdrp);
|
|
|
|
/* get past NIT timer */
|
|
ntp = (struct nit_iftime *)cp;
|
|
cp += sizeof(*ntp);
|
|
|
|
ndp = (struct nit_ifdrops *)cp;
|
|
psn->stat.ps_drop = ndp->nh_drops;
|
|
cp += sizeof *ndp;
|
|
|
|
/* get past packet len */
|
|
nlp = (struct nit_iflen *)cp;
|
|
cp += sizeof(*nlp);
|
|
|
|
/* next snapshot */
|
|
bp += hdrp->nhb_totlen;
|
|
|
|
caplen = nlp->nh_pktlen;
|
|
if (caplen > p->snapshot)
|
|
caplen = p->snapshot;
|
|
|
|
if (pcap_filter(p->fcode.bf_insns, cp, nlp->nh_pktlen, caplen)) {
|
|
struct pcap_pkthdr h;
|
|
h.ts = ntp->nh_timestamp;
|
|
h.len = nlp->nh_pktlen;
|
|
h.caplen = caplen;
|
|
(*callback)(user, &h, cp);
|
|
if (++n >= cnt && !PACKET_COUNT_IS_UNLIMITED(cnt)) {
|
|
p->cc = ep - bp;
|
|
p->bp = bp;
|
|
return (n);
|
|
}
|
|
}
|
|
}
|
|
p->cc = 0;
|
|
return (n);
|
|
}
|
|
|
|
static int
|
|
pcap_inject_snit(pcap_t *p, const void *buf, int size)
|
|
{
|
|
struct strbuf ctl, data;
|
|
|
|
/*
|
|
* XXX - can we just do
|
|
*
|
|
ret = write(pd->f, buf, size);
|
|
*/
|
|
ctl.len = sizeof(*sa); /* XXX - what was this? */
|
|
ctl.buf = (char *)sa;
|
|
data.buf = buf;
|
|
data.len = size;
|
|
ret = putmsg(p->fd, &ctl, &data);
|
|
if (ret == -1) {
|
|
pcap_fmt_errmsg_for_errno(p->errbuf, PCAP_ERRBUF_SIZE,
|
|
errno, "send");
|
|
return (-1);
|
|
}
|
|
return (ret);
|
|
}
|
|
|
|
static int
|
|
nit_setflags(pcap_t *p)
|
|
{
|
|
bpf_u_int32 flags;
|
|
struct strioctl si;
|
|
u_int zero = 0;
|
|
struct timeval timeout;
|
|
|
|
if (p->opt.immediate) {
|
|
/*
|
|
* Set the chunk size to zero, so that chunks get sent
|
|
* up immediately.
|
|
*/
|
|
si.ic_cmd = NIOCSCHUNK;
|
|
si.ic_len = sizeof(zero);
|
|
si.ic_dp = (char *)&zero;
|
|
if (ioctl(p->fd, I_STR, (char *)&si) < 0) {
|
|
pcap_fmt_errmsg_for_errno(p->errbuf, PCAP_ERRBUF_SIZE,
|
|
errno, "NIOCSCHUNK");
|
|
return (-1);
|
|
}
|
|
}
|
|
si.ic_timout = INFTIM;
|
|
if (p->opt.timeout != 0) {
|
|
timeout.tv_sec = p->opt.timeout / 1000;
|
|
timeout.tv_usec = (p->opt.timeout * 1000) % 1000000;
|
|
si.ic_cmd = NIOCSTIME;
|
|
si.ic_len = sizeof(timeout);
|
|
si.ic_dp = (char *)&timeout;
|
|
if (ioctl(p->fd, I_STR, (char *)&si) < 0) {
|
|
pcap_fmt_errmsg_for_errno(p->errbuf, PCAP_ERRBUF_SIZE,
|
|
errno, "NIOCSTIME");
|
|
return (-1);
|
|
}
|
|
}
|
|
flags = NI_TIMESTAMP | NI_LEN | NI_DROPS;
|
|
if (p->opt.promisc)
|
|
flags |= NI_PROMISC;
|
|
si.ic_cmd = NIOCSFLAGS;
|
|
si.ic_len = sizeof(flags);
|
|
si.ic_dp = (char *)&flags;
|
|
if (ioctl(p->fd, I_STR, (char *)&si) < 0) {
|
|
pcap_fmt_errmsg_for_errno(p->errbuf, PCAP_ERRBUF_SIZE,
|
|
errno, "NIOCSFLAGS");
|
|
return (-1);
|
|
}
|
|
return (0);
|
|
}
|
|
|
|
static int
|
|
pcap_activate_snit(pcap_t *p)
|
|
{
|
|
struct strioctl si; /* struct for ioctl() */
|
|
struct ifreq ifr; /* interface request struct */
|
|
int chunksize = CHUNKSIZE;
|
|
int fd;
|
|
static const char dev[] = "/dev/nit";
|
|
int err;
|
|
|
|
if (p->opt.rfmon) {
|
|
/*
|
|
* No monitor mode on SunOS 4.x (no Wi-Fi devices on
|
|
* hardware supported by SunOS 4.x).
|
|
*/
|
|
return (PCAP_ERROR_RFMON_NOTSUP);
|
|
}
|
|
|
|
/*
|
|
* Turn a negative snapshot value (invalid), a snapshot value of
|
|
* 0 (unspecified), or a value bigger than the normal maximum
|
|
* value, into the maximum allowed value.
|
|
*
|
|
* If some application really *needs* a bigger snapshot
|
|
* length, we should just increase MAXIMUM_SNAPLEN.
|
|
*/
|
|
if (p->snapshot <= 0 || p->snapshot > MAXIMUM_SNAPLEN)
|
|
p->snapshot = MAXIMUM_SNAPLEN;
|
|
|
|
if (p->snapshot < 96)
|
|
/*
|
|
* NIT requires a snapshot length of at least 96.
|
|
*/
|
|
p->snapshot = 96;
|
|
|
|
/*
|
|
* Initially try a read/write open (to allow the inject
|
|
* method to work). If that fails due to permission
|
|
* issues, fall back to read-only. This allows a
|
|
* non-root user to be granted specific access to pcap
|
|
* capabilities via file permissions.
|
|
*
|
|
* XXX - we should have an API that has a flag that
|
|
* controls whether to open read-only or read-write,
|
|
* so that denial of permission to send (or inability
|
|
* to send, if sending packets isn't supported on
|
|
* the device in question) can be indicated at open
|
|
* time.
|
|
*/
|
|
p->fd = fd = open(dev, O_RDWR);
|
|
if (fd < 0 && errno == EACCES)
|
|
p->fd = fd = open(dev, O_RDONLY);
|
|
if (fd < 0) {
|
|
if (errno == EACCES) {
|
|
err = PCAP_ERROR_PERM_DENIED;
|
|
snprintf(p->errbuf, PCAP_ERRBUF_SIZE,
|
|
"Attempt to open %s failed with EACCES - root privileges may be required",
|
|
dev);
|
|
} else {
|
|
err = PCAP_ERROR;
|
|
pcap_fmt_errmsg_for_errno(p->errbuf, PCAP_ERRBUF_SIZE,
|
|
errno, "%s", dev);
|
|
}
|
|
goto bad;
|
|
}
|
|
|
|
/* arrange to get discrete messages from the STREAM and use NIT_BUF */
|
|
if (ioctl(fd, I_SRDOPT, (char *)RMSGD) < 0) {
|
|
pcap_fmt_errmsg_for_errno(p->errbuf, PCAP_ERRBUF_SIZE,
|
|
errno, "I_SRDOPT");
|
|
err = PCAP_ERROR;
|
|
goto bad;
|
|
}
|
|
if (ioctl(fd, I_PUSH, "nbuf") < 0) {
|
|
pcap_fmt_errmsg_for_errno(p->errbuf, PCAP_ERRBUF_SIZE,
|
|
errno, "push nbuf");
|
|
err = PCAP_ERROR;
|
|
goto bad;
|
|
}
|
|
/* set the chunksize */
|
|
si.ic_cmd = NIOCSCHUNK;
|
|
si.ic_timout = INFTIM;
|
|
si.ic_len = sizeof(chunksize);
|
|
si.ic_dp = (char *)&chunksize;
|
|
if (ioctl(fd, I_STR, (char *)&si) < 0) {
|
|
pcap_fmt_errmsg_for_errno(p->errbuf, PCAP_ERRBUF_SIZE,
|
|
errno, "NIOCSCHUNK");
|
|
err = PCAP_ERROR;
|
|
goto bad;
|
|
}
|
|
|
|
/* request the interface */
|
|
strncpy(ifr.ifr_name, p->opt.device, sizeof(ifr.ifr_name));
|
|
ifr.ifr_name[sizeof(ifr.ifr_name) - 1] = '\0';
|
|
si.ic_cmd = NIOCBIND;
|
|
si.ic_len = sizeof(ifr);
|
|
si.ic_dp = (char *)𝔦
|
|
if (ioctl(fd, I_STR, (char *)&si) < 0) {
|
|
/*
|
|
* XXX - is there an error that means "no such device"?
|
|
* Is there one that means "that device doesn't support
|
|
* STREAMS NIT"?
|
|
*/
|
|
pcap_fmt_errmsg_for_errno(p->errbuf, PCAP_ERRBUF_SIZE,
|
|
errno, "NIOCBIND: %s", ifr.ifr_name);
|
|
err = PCAP_ERROR;
|
|
goto bad;
|
|
}
|
|
|
|
/* set the snapshot length */
|
|
si.ic_cmd = NIOCSSNAP;
|
|
si.ic_len = sizeof(p->snapshot);
|
|
si.ic_dp = (char *)&p->snapshot;
|
|
if (ioctl(fd, I_STR, (char *)&si) < 0) {
|
|
pcap_fmt_errmsg_for_errno(p->errbuf, PCAP_ERRBUF_SIZE,
|
|
errno, "NIOCSSNAP");
|
|
err = PCAP_ERROR;
|
|
goto bad;
|
|
}
|
|
if (nit_setflags(p) < 0) {
|
|
err = PCAP_ERROR;
|
|
goto bad;
|
|
}
|
|
|
|
(void)ioctl(fd, I_FLUSH, (char *)FLUSHR);
|
|
/*
|
|
* NIT supports only ethernets.
|
|
*/
|
|
p->linktype = DLT_EN10MB;
|
|
|
|
p->bufsize = BUFSPACE;
|
|
p->buffer = malloc(p->bufsize);
|
|
if (p->buffer == NULL) {
|
|
pcap_fmt_errmsg_for_errno(p->errbuf, PCAP_ERRBUF_SIZE,
|
|
errno, "malloc");
|
|
err = PCAP_ERROR;
|
|
goto bad;
|
|
}
|
|
|
|
/*
|
|
* "p->fd" is an FD for a STREAMS device, so "select()" and
|
|
* "poll()" should work on it.
|
|
*/
|
|
p->selectable_fd = p->fd;
|
|
|
|
/*
|
|
* This is (presumably) a real Ethernet capture; give it a
|
|
* link-layer-type list with DLT_EN10MB and DLT_DOCSIS, so
|
|
* that an application can let you choose it, in case you're
|
|
* capturing DOCSIS traffic that a Cisco Cable Modem
|
|
* Termination System is putting out onto an Ethernet (it
|
|
* doesn't put an Ethernet header onto the wire, it puts raw
|
|
* DOCSIS frames out on the wire inside the low-level
|
|
* Ethernet framing).
|
|
*/
|
|
p->dlt_list = (u_int *) malloc(sizeof(u_int) * 2);
|
|
/*
|
|
* If that fails, just leave the list empty.
|
|
*/
|
|
if (p->dlt_list != NULL) {
|
|
p->dlt_list[0] = DLT_EN10MB;
|
|
p->dlt_list[1] = DLT_DOCSIS;
|
|
p->dlt_count = 2;
|
|
}
|
|
|
|
p->read_op = pcap_read_snit;
|
|
p->inject_op = pcap_inject_snit;
|
|
p->setfilter_op = install_bpf_program; /* no kernel filtering */
|
|
p->setdirection_op = NULL; /* Not implemented. */
|
|
p->set_datalink_op = NULL; /* can't change data link type */
|
|
p->getnonblock_op = pcap_getnonblock_fd;
|
|
p->setnonblock_op = pcap_setnonblock_fd;
|
|
p->stats_op = pcap_stats_snit;
|
|
|
|
return (0);
|
|
bad:
|
|
pcap_cleanup_live_common(p);
|
|
return (err);
|
|
}
|
|
|
|
pcap_t *
|
|
pcap_create_interface(const char *device _U_, char *ebuf)
|
|
{
|
|
pcap_t *p;
|
|
|
|
p = PCAP_CREATE_COMMON(ebuf, struct pcap_snit);
|
|
if (p == NULL)
|
|
return (NULL);
|
|
|
|
p->activate_op = pcap_activate_snit;
|
|
return (p);
|
|
}
|
|
|
|
/*
|
|
* XXX - there's probably a NIOCBIND error that means "that device
|
|
* doesn't support NIT"; if so, we should try an NIOCBIND and use that.
|
|
*/
|
|
static int
|
|
can_be_bound(const char *name _U_)
|
|
{
|
|
return (1);
|
|
}
|
|
|
|
static int
|
|
get_if_flags(const char *name _U_, bpf_u_int32 *flags _U_, char *errbuf _U_)
|
|
{
|
|
/*
|
|
* Nothing we can do.
|
|
* XXX - is there a way to find out whether an adapter has
|
|
* something plugged into it?
|
|
*/
|
|
return (0);
|
|
}
|
|
|
|
int
|
|
pcap_platform_finddevs(pcap_if_list_t *devlistp, char *errbuf)
|
|
{
|
|
return (pcap_findalldevs_interfaces(devlistp, errbuf, can_be_bound,
|
|
get_if_flags));
|
|
}
|
|
|
|
/*
|
|
* Libpcap version string.
|
|
*/
|
|
const char *
|
|
pcap_lib_version(void)
|
|
{
|
|
return (PCAP_VERSION_STRING);
|
|
}
|