freebsd-skq/sys/net/bpf.c

1324 lines
28 KiB
C
Raw Normal View History

1994-05-24 10:09:53 +00:00
/*
* Copyright (c) 1990, 1991, 1993
* The Regents of the University of California. All rights reserved.
*
* This code is derived from the Stanford/CMU enet packet filter,
* (net/enet.c) distributed as part of 4.3BSD, and code contributed
1995-05-30 08:16:23 +00:00
* to Berkeley by Steven McCanne and Van Jacobson both of Lawrence
1994-05-24 10:09:53 +00:00
* Berkeley Laboratory.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the University of
* California, Berkeley and its contributors.
* 4. 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 BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* @(#)bpf.c 8.2 (Berkeley) 3/28/94
*
1999-08-28 01:08:13 +00:00
* $FreeBSD$
1994-05-24 10:09:53 +00:00
*/
1999-07-06 19:23:32 +00:00
#include "bpf.h"
1994-05-24 10:09:53 +00:00
#ifndef __GNUC__
#define inline
#else
#define inline __inline
#endif
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/conf.h>
#include <sys/malloc.h>
1994-05-24 10:09:53 +00:00
#include <sys/mbuf.h>
#include <sys/time.h>
#include <sys/proc.h>
#include <sys/signalvar.h>
#include <sys/filio.h>
#include <sys/sockio.h>
#include <sys/ttycom.h>
#include <sys/filedesc.h>
1994-05-24 10:09:53 +00:00
#if defined(sparc) && BSD < 199103
#include <sys/stream.h>
#endif
#include <sys/poll.h>
1994-05-24 10:09:53 +00:00
#include <sys/socket.h>
#include <sys/vnode.h>
1994-05-24 10:09:53 +00:00
#include <net/if.h>
1994-05-24 10:09:53 +00:00
#include <net/bpf.h>
#include <net/bpfdesc.h>
#include <netinet/in.h>
#include <netinet/if_ether.h>
#include <sys/kernel.h>
#include <sys/sysctl.h>
MALLOC_DEFINE(M_BPF, "BPF", "BPF data");
1999-07-06 19:23:32 +00:00
#if NBPF > 0
1994-05-24 10:09:53 +00:00
/*
* Older BSDs don't have kernel malloc.
*/
#if BSD < 199103
extern bcopy();
static caddr_t bpf_alloc();
#include <net/bpf_compat.h>
#define BPF_BUFSIZE (MCLBYTES-8)
#define UIOMOVE(cp, len, code, uio) uiomove(cp, len, code, uio)
#else
#define BPF_BUFSIZE 4096
#define UIOMOVE(cp, len, code, uio) uiomove(cp, len, uio)
#endif
#define PRINET 26 /* interruptible */
/*
* The default read buffer size is patchable.
*/
static int bpf_bufsize = BPF_BUFSIZE;
SYSCTL_INT(_debug, OID_AUTO, bpf_bufsize, CTLFLAG_RW,
&bpf_bufsize, 0, "");
static int bpf_maxbufsize = BPF_MAXBUFSIZE;
SYSCTL_INT(_debug, OID_AUTO, bpf_maxbufsize, CTLFLAG_RW,
&bpf_maxbufsize, 0, "");
1994-05-24 10:09:53 +00:00
/*
* bpf_iflist is the list of interfaces; each corresponds to an ifnet
*/
static struct bpf_if *bpf_iflist;
1994-05-24 10:09:53 +00:00
static int bpf_allocbufs __P((struct bpf_d *));
static void bpf_attachd __P((struct bpf_d *d, struct bpf_if *bp));
static void bpf_detachd __P((struct bpf_d *d));
1994-05-24 10:09:53 +00:00
static void bpf_freed __P((struct bpf_d *));
static void bpf_mcopy __P((const void *, void *, size_t));
1994-05-24 10:09:53 +00:00
static int bpf_movein __P((struct uio *, int,
struct mbuf **, struct sockaddr *, int *));
static int bpf_setif __P((struct bpf_d *, struct ifreq *));
static inline void
bpf_wakeup __P((struct bpf_d *));
static void catchpacket __P((struct bpf_d *, u_char *, u_int,
u_int, void (*)(const void *, void *, size_t)));
1994-05-24 10:09:53 +00:00
static void reset_d __P((struct bpf_d *));
static int bpf_setf __P((struct bpf_d *, struct bpf_program *));
1994-05-24 10:09:53 +00:00
static d_open_t bpfopen;
static d_close_t bpfclose;
static d_read_t bpfread;
static d_write_t bpfwrite;
static d_ioctl_t bpfioctl;
static d_poll_t bpfpoll;
#define CDEV_MAJOR 23
static struct cdevsw bpf_cdevsw = {
/* open */ bpfopen,
/* close */ bpfclose,
/* read */ bpfread,
/* write */ bpfwrite,
/* ioctl */ bpfioctl,
/* poll */ bpfpoll,
/* mmap */ nommap,
/* strategy */ nostrategy,
/* name */ "bpf",
/* maj */ CDEV_MAJOR,
/* dump */ nodump,
/* psize */ nopsize,
/* flags */ 0,
/* bmaj */ -1
};
1994-05-24 10:09:53 +00:00
static int
bpf_movein(uio, linktype, mp, sockp, datlen)
register struct uio *uio;
int linktype, *datlen;
register struct mbuf **mp;
register struct sockaddr *sockp;
{
struct mbuf *m;
int error;
int len;
int hlen;
/*
* Build a sockaddr based on the data link layer type.
* We do this at this level because the ethernet header
* is copied directly into the data field of the sockaddr.
* In the case of SLIP, there is no header and the packet
* is forwarded as is.
* Also, we are careful to leave room at the front of the mbuf
* for the link level header.
*/
switch (linktype) {
case DLT_SLIP:
sockp->sa_family = AF_INET;
hlen = 0;
break;
case DLT_EN10MB:
sockp->sa_family = AF_UNSPEC;
/* XXX Would MAXLINKHDR be better? */
hlen = sizeof(struct ether_header);
break;
case DLT_FDDI:
#if defined(__FreeBSD__) || defined(__bsdi__)
sockp->sa_family = AF_IMPLINK;
hlen = 0;
#else
1994-05-24 10:09:53 +00:00
sockp->sa_family = AF_UNSPEC;
/* XXX 4(FORMAC)+6(dst)+6(src)+3(LLC)+5(SNAP) */
hlen = 24;
#endif
1994-05-24 10:09:53 +00:00
break;
1998-08-18 10:13:11 +00:00
case DLT_RAW:
1994-05-24 10:09:53 +00:00
case DLT_NULL:
sockp->sa_family = AF_UNSPEC;
hlen = 0;
break;
#ifdef __FreeBSD__
case DLT_ATM_RFC1483:
/*
* en atm driver requires 4-byte atm pseudo header.
* though it isn't standard, vpi:vci needs to be
* specified anyway.
*/
sockp->sa_family = AF_UNSPEC;
hlen = 12; /* XXX 4(ATM_PH) + 3(LLC) + 5(SNAP) */
break;
#endif
1994-05-24 10:09:53 +00:00
default:
return (EIO);
}
len = uio->uio_resid;
*datlen = len - hlen;
if ((unsigned)len > MCLBYTES)
return (EIO);
MGETHDR(m, M_WAIT, MT_DATA);
1994-05-24 10:09:53 +00:00
if (m == 0)
return (ENOBUFS);
if (len > MHLEN) {
1994-05-24 10:09:53 +00:00
#if BSD >= 199103
MCLGET(m, M_WAIT);
if ((m->m_flags & M_EXT) == 0) {
#else
MCLGET(m);
if (m->m_len != MCLBYTES) {
#endif
error = ENOBUFS;
goto bad;
}
}
m->m_pkthdr.len = m->m_len = len;
m->m_pkthdr.rcvif = NULL;
1994-05-24 10:09:53 +00:00
*mp = m;
/*
* Make room for link header.
*/
if (hlen != 0) {
m->m_pkthdr.len -= hlen;
1994-05-24 10:09:53 +00:00
m->m_len -= hlen;
#if BSD >= 199103
m->m_data += hlen; /* XXX */
#else
m->m_off += hlen;
#endif
error = UIOMOVE((caddr_t)sockp->sa_data, hlen, UIO_WRITE, uio);
if (error)
goto bad;
}
error = UIOMOVE(mtod(m, caddr_t), len - hlen, UIO_WRITE, uio);
if (!error)
return (0);
bad:
m_freem(m);
return (error);
}
/*
* Attach file to the bpf interface, i.e. make d listen on bp.
* Must be called at splimp.
*/
static void
bpf_attachd(d, bp)
struct bpf_d *d;
struct bpf_if *bp;
{
/*
* Point d at bp, and add d to the interface's list of listeners.
* Finally, point the driver's bpf cookie at the interface so
* it will divert packets to bpf.
*/
d->bd_bif = bp;
d->bd_next = bp->bif_dlist;
bp->bif_dlist = d;
bp->bif_ifp->if_bpf = bp;
1994-05-24 10:09:53 +00:00
}
/*
* Detach a file from its interface.
*/
static void
bpf_detachd(d)
struct bpf_d *d;
{
struct bpf_d **p;
struct bpf_if *bp;
bp = d->bd_bif;
/*
* Check if this descriptor had requested promiscuous mode.
* If so, turn it off.
*/
if (d->bd_promisc) {
d->bd_promisc = 0;
if (ifpromisc(bp->bif_ifp, 0))
/*
* Something is really wrong if we were able to put
* the driver into promiscuous mode, but can't
* take it out.
*/
panic("bpf: ifpromisc failed");
}
/* Remove d from the interface's descriptor list. */
p = &bp->bif_dlist;
while (*p != d) {
p = &(*p)->bd_next;
if (*p == 0)
panic("bpf_detachd: descriptor not in list");
}
*p = (*p)->bd_next;
if (bp->bif_dlist == 0)
/*
* Let the driver know that there are no more listeners.
*/
d->bd_bif->bif_ifp->if_bpf = 0;
1994-05-24 10:09:53 +00:00
d->bd_bif = 0;
}
/*
* Open ethernet device. Returns ENXIO for illegal minor device number,
* EBUSY if file is open by another process.
*/
/* ARGSUSED */
static int
bpfopen(dev, flags, fmt, p)
1994-05-24 10:09:53 +00:00
dev_t dev;
int flags;
int fmt;
struct proc *p;
1994-05-24 10:09:53 +00:00
{
register struct bpf_d *d;
This Implements the mumbled about "Jail" feature. This is a seriously beefed up chroot kind of thing. The process is jailed along the same lines as a chroot does it, but with additional tough restrictions imposed on what the superuser can do. For all I know, it is safe to hand over the root bit inside a prison to the customer living in that prison, this is what it was developed for in fact: "real virtual servers". Each prison has an ip number associated with it, which all IP communications will be coerced to use and each prison has its own hostname. Needless to say, you need more RAM this way, but the advantage is that each customer can run their own particular version of apache and not stomp on the toes of their neighbors. It generally does what one would expect, but setting up a jail still takes a little knowledge. A few notes: I have no scripts for setting up a jail, don't ask me for them. The IP number should be an alias on one of the interfaces. mount a /proc in each jail, it will make ps more useable. /proc/<pid>/status tells the hostname of the prison for jailed processes. Quotas are only sensible if you have a mountpoint per prison. There are no privisions for stopping resource-hogging. Some "#ifdef INET" and similar may be missing (send patches!) If somebody wants to take it from here and develop it into more of a "virtual machine" they should be most welcome! Tools, comments, patches & documentation most welcome. Have fun... Sponsored by: http://www.rndassociates.com/ Run for almost a year by: http://www.servetheweb.com/
1999-04-28 11:38:52 +00:00
if (p->p_prison)
return (EPERM);
d = dev->si_drv1;
1994-05-24 10:09:53 +00:00
/*
* Each minor can be opened by only one process. If the requested
1994-05-24 10:09:53 +00:00
* minor is in use, return EBUSY.
*/
if (d)
1994-05-24 10:09:53 +00:00
return (EBUSY);
make_dev(&bpf_cdevsw, minor(dev), 0, 0, 0600, "bpf%d", lminor(dev));
MALLOC(d, struct bpf_d *, sizeof(*d), M_BPF, M_WAITOK);
bzero(d, sizeof(*d));
dev->si_drv1 = d;
1994-05-24 10:09:53 +00:00
d->bd_bufsize = bpf_bufsize;
d->bd_sig = SIGIO;
1994-05-24 10:09:53 +00:00
return (0);
}
/*
* Close the descriptor by detaching it from its interface,
* deallocating its buffers, and marking it free.
*/
/* ARGSUSED */
static int
bpfclose(dev, flags, fmt, p)
1994-05-24 10:09:53 +00:00
dev_t dev;
int flags;
int fmt;
struct proc *p;
1994-05-24 10:09:53 +00:00
{
register struct bpf_d *d = dev->si_drv1;
1994-05-24 10:09:53 +00:00
register int s;
funsetown(d->bd_sigio);
1994-05-24 10:09:53 +00:00
s = splimp();
if (d->bd_bif)
bpf_detachd(d);
splx(s);
bpf_freed(d);
dev->si_drv1 = 0;
FREE(d, M_BPF);
1994-05-24 10:09:53 +00:00
return (0);
}
/*
* Support for SunOS, which does not have tsleep.
*/
#if BSD < 199103
static
bpf_timeout(arg)
caddr_t arg;
{
struct bpf_d *d = (struct bpf_d *)arg;
d->bd_timedout = 1;
wakeup(arg);
}
#define BPF_SLEEP(chan, pri, s, t) bpf_sleep((struct bpf_d *)chan)
int
bpf_sleep(d)
register struct bpf_d *d;
{
register int rto = d->bd_rtout;
register int st;
if (rto != 0) {
d->bd_timedout = 0;
timeout(bpf_timeout, (caddr_t)d, rto);
}
st = sleep((caddr_t)d, PRINET|PCATCH);
if (rto != 0) {
if (d->bd_timedout == 0)
untimeout(bpf_timeout, (caddr_t)d);
else if (st == 0)
return EWOULDBLOCK;
}
return (st != 0) ? EINTR : 0;
}
#else
#define BPF_SLEEP tsleep
#endif
/*
* Rotate the packet buffers in descriptor d. Move the store buffer
* into the hold slot, and the free buffer into the store slot.
* Zero the length of the new store buffer.
*/
#define ROTATE_BUFFERS(d) \
(d)->bd_hbuf = (d)->bd_sbuf; \
(d)->bd_hlen = (d)->bd_slen; \
(d)->bd_sbuf = (d)->bd_fbuf; \
(d)->bd_slen = 0; \
(d)->bd_fbuf = 0;
/*
* bpfread - read next chunk of packets from buffers
*/
static int
bpfread(dev, uio, ioflag)
1994-05-24 10:09:53 +00:00
dev_t dev;
register struct uio *uio;
int ioflag;
1994-05-24 10:09:53 +00:00
{
register struct bpf_d *d = dev->si_drv1;
1994-05-24 10:09:53 +00:00
int error;
int s;
/*
* Restrict application to use a buffer the same size as
* as kernel buffers.
*/
if (uio->uio_resid != d->bd_bufsize)
return (EINVAL);
s = splimp();
/*
* If the hold buffer is empty, then do a timed sleep, which
* ends when the timeout expires or when enough packets
* have arrived to fill the store buffer.
*/
while (d->bd_hbuf == 0) {
if (d->bd_immediate && d->bd_slen != 0) {
/*
* A packet(s) either arrived since the previous
* read or arrived while we were asleep.
* Rotate the buffers and return what's here.
*/
ROTATE_BUFFERS(d);
break;
}
if (ioflag & IO_NDELAY)
error = EWOULDBLOCK;
else
error = BPF_SLEEP((caddr_t)d, PRINET|PCATCH, "bpf",
d->bd_rtout);
1994-05-24 10:09:53 +00:00
if (error == EINTR || error == ERESTART) {
splx(s);
return (error);
}
if (error == EWOULDBLOCK) {
/*
* On a timeout, return what's in the buffer,
* which may be nothing. If there is something
* in the store buffer, we can rotate the buffers.
*/
if (d->bd_hbuf)
/*
* We filled up the buffer in between
* getting the timeout and arriving
* here, so we don't need to rotate.
*/
break;
if (d->bd_slen == 0) {
splx(s);
return (0);
}
ROTATE_BUFFERS(d);
break;
}
}
/*
* At this point, we know we have something in the hold slot.
*/
splx(s);
/*
* Move data from hold buffer into user space.
* We know the entire buffer is transferred since
* we checked above that the read buffer is bpf_bufsize bytes.
*/
error = UIOMOVE(d->bd_hbuf, d->bd_hlen, UIO_READ, uio);
s = splimp();
d->bd_fbuf = d->bd_hbuf;
d->bd_hbuf = 0;
d->bd_hlen = 0;
splx(s);
return (error);
}
/*
* If there are processes sleeping on this descriptor, wake them up.
*/
static inline void
bpf_wakeup(d)
register struct bpf_d *d;
{
wakeup((caddr_t)d);
if (d->bd_async && d->bd_sig && d->bd_sigio)
pgsigio(d->bd_sigio, d->bd_sig, 0);
1994-05-24 10:09:53 +00:00
#if BSD >= 199103
selwakeup(&d->bd_sel);
/* XXX */
d->bd_sel.si_pid = 0;
#else
if (d->bd_selproc) {
selwakeup(d->bd_selproc, (int)d->bd_selcoll);
d->bd_selcoll = 0;
d->bd_selproc = 0;
}
#endif
}
static int
bpfwrite(dev, uio, ioflag)
1994-05-24 10:09:53 +00:00
dev_t dev;
struct uio *uio;
int ioflag;
1994-05-24 10:09:53 +00:00
{
register struct bpf_d *d = dev->si_drv1;
1994-05-24 10:09:53 +00:00
struct ifnet *ifp;
struct mbuf *m;
int error, s;
static struct sockaddr dst;
int datlen;
if (d->bd_bif == 0)
return (ENXIO);
ifp = d->bd_bif->bif_ifp;
if (uio->uio_resid == 0)
return (0);
error = bpf_movein(uio, (int)d->bd_bif->bif_dlt, &m, &dst, &datlen);
if (error)
return (error);
if (datlen > ifp->if_mtu)
return (EMSGSIZE);
if (d->bd_hdrcmplt)
dst.sa_family = pseudo_AF_HDRCMPLT;
1994-05-24 10:09:53 +00:00
s = splnet();
#if BSD >= 199103
error = (*ifp->if_output)(ifp, m, &dst, (struct rtentry *)0);
#else
error = (*ifp->if_output)(ifp, m, &dst);
#endif
splx(s);
/*
* The driver frees the mbuf.
*/
return (error);
}
/*
* Reset a descriptor by flushing its packet buffer and clearing the
* receive and drop counts. Should be called at splimp.
*/
static void
reset_d(d)
struct bpf_d *d;
{
if (d->bd_hbuf) {
/* Free the hold buffer. */
d->bd_fbuf = d->bd_hbuf;
d->bd_hbuf = 0;
}
d->bd_slen = 0;
d->bd_hlen = 0;
d->bd_rcount = 0;
d->bd_dcount = 0;
}
/*
* FIONREAD Check for read packet available.
* SIOCGIFADDR Get interface address - convenient hook to driver.
* BIOCGBLEN Get buffer len [for read()].
* BIOCSETF Set ethernet read filter.
* BIOCFLUSH Flush read packet buffer.
* BIOCPROMISC Put interface into promiscuous mode.
* BIOCGDLT Get link layer type.
* BIOCGETIF Get interface name.
* BIOCSETIF Set interface.
* BIOCSRTIMEOUT Set read timeout.
* BIOCGRTIMEOUT Get read timeout.
* BIOCGSTATS Get packet stats.
* BIOCIMMEDIATE Set immediate mode.
* BIOCVERSION Get filter language version.
* BIOCGHDRCMPLT Get "header already complete" flag
* BIOCSHDRCMPLT Set "header already complete" flag
1994-05-24 10:09:53 +00:00
*/
/* ARGSUSED */
static int
bpfioctl(dev, cmd, addr, flags, p)
1994-05-24 10:09:53 +00:00
dev_t dev;
u_long cmd;
1994-05-24 10:09:53 +00:00
caddr_t addr;
int flags;
struct proc *p;
1994-05-24 10:09:53 +00:00
{
register struct bpf_d *d = dev->si_drv1;
1994-05-24 10:09:53 +00:00
int s, error = 0;
switch (cmd) {
default:
error = EINVAL;
break;
/*
* Check for read packet available.
*/
case FIONREAD:
{
int n;
s = splimp();
n = d->bd_slen;
if (d->bd_hbuf)
n += d->bd_hlen;
splx(s);
*(int *)addr = n;
break;
}
case SIOCGIFADDR:
{
struct ifnet *ifp;
if (d->bd_bif == 0)
error = EINVAL;
else {
ifp = d->bd_bif->bif_ifp;
error = (*ifp->if_ioctl)(ifp, cmd, addr);
}
break;
}
/*
* Get buffer len [for read()].
*/
case BIOCGBLEN:
*(u_int *)addr = d->bd_bufsize;
break;
/*
* Set buffer length.
*/
case BIOCSBLEN:
#if BSD < 199103
error = EINVAL;
#else
if (d->bd_bif != 0)
error = EINVAL;
else {
register u_int size = *(u_int *)addr;
if (size > bpf_maxbufsize)
*(u_int *)addr = size = bpf_maxbufsize;
1994-05-24 10:09:53 +00:00
else if (size < BPF_MINBUFSIZE)
*(u_int *)addr = size = BPF_MINBUFSIZE;
d->bd_bufsize = size;
}
#endif
break;
/*
* Set link layer read filter.
*/
case BIOCSETF:
error = bpf_setf(d, (struct bpf_program *)addr);
break;
/*
* Flush read packet buffer.
*/
case BIOCFLUSH:
s = splimp();
reset_d(d);
splx(s);
break;
/*
* Put interface into promiscuous mode.
*/
case BIOCPROMISC:
if (d->bd_bif == 0) {
/*
* No interface attached yet.
*/
error = EINVAL;
break;
}
s = splimp();
if (d->bd_promisc == 0) {
error = ifpromisc(d->bd_bif->bif_ifp, 1);
if (error == 0)
d->bd_promisc = 1;
}
splx(s);
break;
/*
* Get device parameters.
*/
case BIOCGDLT:
if (d->bd_bif == 0)
error = EINVAL;
else
*(u_int *)addr = d->bd_bif->bif_dlt;
break;
/*
* Get interface name.
1994-05-24 10:09:53 +00:00
*/
case BIOCGETIF:
if (d->bd_bif == 0)
error = EINVAL;
else {
struct ifnet *const ifp = d->bd_bif->bif_ifp;
struct ifreq *const ifr = (struct ifreq *)addr;
snprintf(ifr->ifr_name, sizeof(ifr->ifr_name),
"%s%d", ifp->if_name, ifp->if_unit);
}
1994-05-24 10:09:53 +00:00
break;
/*
* Set interface.
*/
case BIOCSETIF:
error = bpf_setif(d, (struct ifreq *)addr);
break;
/*
* Set read timeout.
*/
case BIOCSRTIMEOUT:
{
struct timeval *tv = (struct timeval *)addr;
/*
* Subtract 1 tick from tvtohz() since this isn't
* a one-shot timer.
*/
if ((error = itimerfix(tv)) == 0)
d->bd_rtout = tvtohz(tv) - 1;
1994-05-24 10:09:53 +00:00
break;
}
/*
* Get read timeout.
*/
case BIOCGRTIMEOUT:
{
struct timeval *tv = (struct timeval *)addr;
tv->tv_sec = d->bd_rtout / hz;
tv->tv_usec = (d->bd_rtout % hz) * tick;
1994-05-24 10:09:53 +00:00
break;
}
/*
* Get packet stats.
*/
case BIOCGSTATS:
{
struct bpf_stat *bs = (struct bpf_stat *)addr;
bs->bs_recv = d->bd_rcount;
bs->bs_drop = d->bd_dcount;
break;
}
/*
* Set immediate mode.
*/
case BIOCIMMEDIATE:
d->bd_immediate = *(u_int *)addr;
break;
case BIOCVERSION:
{
struct bpf_version *bv = (struct bpf_version *)addr;
bv->bv_major = BPF_MAJOR_VERSION;
bv->bv_minor = BPF_MINOR_VERSION;
break;
}
/*
* Get "header already complete" flag
*/
case BIOCGHDRCMPLT:
*(u_int *)addr = d->bd_hdrcmplt;
break;
/*
* Set "header already complete" flag
*/
case BIOCSHDRCMPLT:
d->bd_hdrcmplt = *(u_int *)addr ? 1 : 0;
break;
case FIONBIO: /* Non-blocking I/O */
break;
case FIOASYNC: /* Send signal on receive packets */
d->bd_async = *(int *)addr;
break;
case FIOSETOWN:
error = fsetown(*(int *)addr, &d->bd_sigio);
break;
case FIOGETOWN:
*(int *)addr = fgetown(d->bd_sigio);
break;
/* This is deprecated, FIOSETOWN should be used instead. */
case TIOCSPGRP:
error = fsetown(-(*(int *)addr), &d->bd_sigio);
break;
/* This is deprecated, FIOGETOWN should be used instead. */
case TIOCGPGRP:
*(int *)addr = -fgetown(d->bd_sigio);
break;
case BIOCSRSIG: /* Set receive signal */
{
u_int sig;
sig = *(u_int *)addr;
if (sig >= NSIG)
error = EINVAL;
else
d->bd_sig = sig;
break;
}
case BIOCGRSIG:
*(u_int *)addr = d->bd_sig;
break;
1994-05-24 10:09:53 +00:00
}
return (error);
}
/*
* Set d's packet filter program to fp. If this file already has a filter,
* free it and replace it. Returns EINVAL for bogus requests.
*/
static int
1994-05-24 10:09:53 +00:00
bpf_setf(d, fp)
struct bpf_d *d;
struct bpf_program *fp;
{
struct bpf_insn *fcode, *old;
u_int flen, size;
int s;
old = d->bd_filter;
if (fp->bf_insns == 0) {
if (fp->bf_len != 0)
return (EINVAL);
s = splimp();
d->bd_filter = 0;
reset_d(d);
splx(s);
if (old != 0)
free((caddr_t)old, M_BPF);
1994-05-24 10:09:53 +00:00
return (0);
}
flen = fp->bf_len;
if (flen > BPF_MAXINSNS)
return (EINVAL);
size = flen * sizeof(*fp->bf_insns);
fcode = (struct bpf_insn *)malloc(size, M_BPF, M_WAITOK);
1994-05-24 10:09:53 +00:00
if (copyin((caddr_t)fp->bf_insns, (caddr_t)fcode, size) == 0 &&
bpf_validate(fcode, (int)flen)) {
s = splimp();
d->bd_filter = fcode;
reset_d(d);
splx(s);
if (old != 0)
free((caddr_t)old, M_BPF);
1994-05-24 10:09:53 +00:00
return (0);
}
free((caddr_t)fcode, M_BPF);
1994-05-24 10:09:53 +00:00
return (EINVAL);
}
/*
* Detach a file from its current interface (if attached at all) and attach
* to the interface indicated by the name stored in ifr.
* Return an errno or 0.
*/
static int
bpf_setif(d, ifr)
struct bpf_d *d;
struct ifreq *ifr;
{
struct bpf_if *bp;
int s, error;
struct ifnet *theywant;
theywant = ifunit(ifr->ifr_name);
if (theywant == 0)
return ENXIO;
1994-05-24 10:09:53 +00:00
/*
* Look through attached interfaces for the named one.
*/
for (bp = bpf_iflist; bp != 0; bp = bp->bif_next) {
struct ifnet *ifp = bp->bif_ifp;
if (ifp == 0 || ifp != theywant)
1994-05-24 10:09:53 +00:00
continue;
/*
* We found the requested interface.
* If it's not up, return an error.
* Allocate the packet buffers if we need to.
* If we're already attached to requested interface,
* just flush the buffer.
*/
if ((ifp->if_flags & IFF_UP) == 0)
return (ENETDOWN);
if (d->bd_sbuf == 0) {
error = bpf_allocbufs(d);
if (error != 0)
return (error);
}
s = splimp();
if (bp != d->bd_bif) {
if (d->bd_bif)
/*
* Detach if attached to something else.
*/
bpf_detachd(d);
bpf_attachd(d, bp);
}
reset_d(d);
splx(s);
return (0);
}
/* Not found. */
return (ENXIO);
}
/*
* Support for select() and poll() system calls
1994-05-24 10:09:53 +00:00
*
* Return true iff the specific operation will not block indefinitely.
* Otherwise, return false but make a note that a selwakeup() must be done.
*/
int
bpfpoll(dev, events, p)
1994-05-24 10:09:53 +00:00
register dev_t dev;
int events;
1994-05-24 10:09:53 +00:00
struct proc *p;
{
register struct bpf_d *d;
register int s;
int revents = 0;
1994-05-24 10:09:53 +00:00
/*
* An imitation of the FIONREAD ioctl code.
*/
d = dev->si_drv1;
1994-05-24 10:09:53 +00:00
s = splimp();
This Implements the mumbled about "Jail" feature. This is a seriously beefed up chroot kind of thing. The process is jailed along the same lines as a chroot does it, but with additional tough restrictions imposed on what the superuser can do. For all I know, it is safe to hand over the root bit inside a prison to the customer living in that prison, this is what it was developed for in fact: "real virtual servers". Each prison has an ip number associated with it, which all IP communications will be coerced to use and each prison has its own hostname. Needless to say, you need more RAM this way, but the advantage is that each customer can run their own particular version of apache and not stomp on the toes of their neighbors. It generally does what one would expect, but setting up a jail still takes a little knowledge. A few notes: I have no scripts for setting up a jail, don't ask me for them. The IP number should be an alias on one of the interfaces. mount a /proc in each jail, it will make ps more useable. /proc/<pid>/status tells the hostname of the prison for jailed processes. Quotas are only sensible if you have a mountpoint per prison. There are no privisions for stopping resource-hogging. Some "#ifdef INET" and similar may be missing (send patches!) If somebody wants to take it from here and develop it into more of a "virtual machine" they should be most welcome! Tools, comments, patches & documentation most welcome. Have fun... Sponsored by: http://www.rndassociates.com/ Run for almost a year by: http://www.servetheweb.com/
1999-04-28 11:38:52 +00:00
if (events & (POLLIN | POLLRDNORM)) {
if (d->bd_hlen != 0 || (d->bd_immediate && d->bd_slen != 0))
revents |= events & (POLLIN | POLLRDNORM);
else
selrecord(p, &d->bd_sel);
This Implements the mumbled about "Jail" feature. This is a seriously beefed up chroot kind of thing. The process is jailed along the same lines as a chroot does it, but with additional tough restrictions imposed on what the superuser can do. For all I know, it is safe to hand over the root bit inside a prison to the customer living in that prison, this is what it was developed for in fact: "real virtual servers". Each prison has an ip number associated with it, which all IP communications will be coerced to use and each prison has its own hostname. Needless to say, you need more RAM this way, but the advantage is that each customer can run their own particular version of apache and not stomp on the toes of their neighbors. It generally does what one would expect, but setting up a jail still takes a little knowledge. A few notes: I have no scripts for setting up a jail, don't ask me for them. The IP number should be an alias on one of the interfaces. mount a /proc in each jail, it will make ps more useable. /proc/<pid>/status tells the hostname of the prison for jailed processes. Quotas are only sensible if you have a mountpoint per prison. There are no privisions for stopping resource-hogging. Some "#ifdef INET" and similar may be missing (send patches!) If somebody wants to take it from here and develop it into more of a "virtual machine" they should be most welcome! Tools, comments, patches & documentation most welcome. Have fun... Sponsored by: http://www.rndassociates.com/ Run for almost a year by: http://www.servetheweb.com/
1999-04-28 11:38:52 +00:00
}
1994-05-24 10:09:53 +00:00
splx(s);
return (revents);
1994-05-24 10:09:53 +00:00
}
/*
* Incoming linkage from device drivers. Process the packet pkt, of length
* pktlen, which is stored in a contiguous buffer. The packet is parsed
* by each process' filter, and if accepted, stashed into the corresponding
* buffer.
*/
void
bpf_tap(ifp, pkt, pktlen)
struct ifnet *ifp;
1994-05-24 10:09:53 +00:00
register u_char *pkt;
register u_int pktlen;
{
struct bpf_if *bp;
register struct bpf_d *d;
register u_int slen;
/*
* Note that the ipl does not have to be raised at this point.
* The only problem that could arise here is that if two different
* interfaces shared any data. This is not the case.
*/
bp = ifp->if_bpf;
1994-05-24 10:09:53 +00:00
for (d = bp->bif_dlist; d != 0; d = d->bd_next) {
++d->bd_rcount;
slen = bpf_filter(d->bd_filter, pkt, pktlen, pktlen);
if (slen != 0)
catchpacket(d, pkt, pktlen, slen, bcopy);
}
}
/*
* Copy data from an mbuf chain into a buffer. This code is derived
* from m_copydata in sys/uipc_mbuf.c.
*/
static void
bpf_mcopy(src_arg, dst_arg, len)
const void *src_arg;
void *dst_arg;
register size_t len;
1994-05-24 10:09:53 +00:00
{
register const struct mbuf *m;
register u_int count;
u_char *dst;
m = src_arg;
dst = dst_arg;
while (len > 0) {
if (m == 0)
panic("bpf_mcopy");
count = min(m->m_len, len);
bcopy(mtod(m, void *), dst, count);
1994-05-24 10:09:53 +00:00
m = m->m_next;
dst += count;
len -= count;
}
}
/*
* Incoming linkage from device drivers, when packet is in an mbuf chain.
*/
void
bpf_mtap(ifp, m)
struct ifnet *ifp;
1994-05-24 10:09:53 +00:00
struct mbuf *m;
{
struct bpf_if *bp = ifp->if_bpf;
1994-05-24 10:09:53 +00:00
struct bpf_d *d;
u_int pktlen, slen;
struct mbuf *m0;
pktlen = 0;
for (m0 = m; m0 != 0; m0 = m0->m_next)
pktlen += m0->m_len;
for (d = bp->bif_dlist; d != 0; d = d->bd_next) {
++d->bd_rcount;
slen = bpf_filter(d->bd_filter, (u_char *)m, pktlen, 0);
if (slen != 0)
catchpacket(d, (u_char *)m, pktlen, slen, bpf_mcopy);
}
}
/*
* Move the packet data from interface memory (pkt) into the
* store buffer. Return 1 if it's time to wakeup a listener (buffer full),
* otherwise 0. "copy" is the routine called to do the actual data
* transfer. bcopy is passed in to copy contiguous chunks, while
* bpf_mcopy is passed in to copy mbuf chains. In the latter case,
* pkt is really an mbuf.
*/
static void
catchpacket(d, pkt, pktlen, snaplen, cpfn)
register struct bpf_d *d;
register u_char *pkt;
register u_int pktlen, snaplen;
register void (*cpfn) __P((const void *, void *, size_t));
1994-05-24 10:09:53 +00:00
{
register struct bpf_hdr *hp;
register int totlen, curlen;
register int hdrlen = d->bd_bif->bif_hdrlen;
/*
* Figure out how many bytes to move. If the packet is
* greater or equal to the snapshot length, transfer that
* much. Otherwise, transfer the whole packet (unless
* we hit the buffer size limit).
*/
totlen = hdrlen + min(snaplen, pktlen);
if (totlen > d->bd_bufsize)
totlen = d->bd_bufsize;
/*
* Round up the end of the previous packet to the next longword.
*/
curlen = BPF_WORDALIGN(d->bd_slen);
if (curlen + totlen > d->bd_bufsize) {
/*
* This packet will overflow the storage buffer.
* Rotate the buffers if we can, then wakeup any
* pending reads.
*/
if (d->bd_fbuf == 0) {
/*
* We haven't completed the previous read yet,
* so drop the packet.
*/
++d->bd_dcount;
return;
}
ROTATE_BUFFERS(d);
bpf_wakeup(d);
curlen = 0;
}
else if (d->bd_immediate)
/*
* Immediate mode is set. A packet arrived so any
* reads should be woken up.
*/
bpf_wakeup(d);
/*
* Append the bpf header.
*/
hp = (struct bpf_hdr *)(d->bd_sbuf + curlen);
#if BSD >= 199103
microtime(&hp->bh_tstamp);
#elif defined(sun)
uniqtime(&hp->bh_tstamp);
#else
hp->bh_tstamp = time;
#endif
hp->bh_datalen = pktlen;
hp->bh_hdrlen = hdrlen;
/*
* Copy the packet data into the store buffer and update its length.
*/
(*cpfn)(pkt, (u_char *)hp + hdrlen, (hp->bh_caplen = totlen - hdrlen));
d->bd_slen = curlen + totlen;
}
/*
* Initialize all nonzero fields of a descriptor.
*/
static int
bpf_allocbufs(d)
register struct bpf_d *d;
{
d->bd_fbuf = (caddr_t)malloc(d->bd_bufsize, M_BPF, M_WAITOK);
1994-05-24 10:09:53 +00:00
if (d->bd_fbuf == 0)
return (ENOBUFS);
d->bd_sbuf = (caddr_t)malloc(d->bd_bufsize, M_BPF, M_WAITOK);
1994-05-24 10:09:53 +00:00
if (d->bd_sbuf == 0) {
free(d->bd_fbuf, M_BPF);
1994-05-24 10:09:53 +00:00
return (ENOBUFS);
}
d->bd_slen = 0;
d->bd_hlen = 0;
return (0);
}
/*
* Free buffers currently in use by a descriptor.
* Called on close.
*/
static void
bpf_freed(d)
register struct bpf_d *d;
{
/*
* We don't need to lock out interrupts since this descriptor has
* been detached from its interface and it yet hasn't been marked
* free.
*/
if (d->bd_sbuf != 0) {
free(d->bd_sbuf, M_BPF);
1994-05-24 10:09:53 +00:00
if (d->bd_hbuf != 0)
free(d->bd_hbuf, M_BPF);
1994-05-24 10:09:53 +00:00
if (d->bd_fbuf != 0)
free(d->bd_fbuf, M_BPF);
1994-05-24 10:09:53 +00:00
}
if (d->bd_filter)
free((caddr_t)d->bd_filter, M_BPF);
1994-05-24 10:09:53 +00:00
}
/*
* Attach an interface to bpf. driverp is a pointer to a (struct bpf_if *)
* in the driver's softc; dlt is the link layer type; hdrlen is the fixed
* size of the link header (variable length headers not yet supported).
*/
void
bpfattach(ifp, dlt, hdrlen)
1994-05-24 10:09:53 +00:00
struct ifnet *ifp;
u_int dlt, hdrlen;
{
struct bpf_if *bp;
bp = (struct bpf_if *)malloc(sizeof(*bp), M_BPF, M_DONTWAIT);
1994-05-24 10:09:53 +00:00
if (bp == 0)
panic("bpfattach");
bp->bif_dlist = 0;
bp->bif_ifp = ifp;
bp->bif_dlt = dlt;
bp->bif_next = bpf_iflist;
bpf_iflist = bp;
bp->bif_ifp->if_bpf = 0;
1994-05-24 10:09:53 +00:00
/*
* Compute the length of the bpf header. This is not necessarily
* equal to SIZEOF_BPF_HDR because we want to insert spacing such
* that the network layer header begins on a longword boundary (for
* performance reasons and to alleviate alignment restrictions).
*/
bp->bif_hdrlen = BPF_WORDALIGN(hdrlen + SIZEOF_BPF_HDR) - hdrlen;
if (bootverbose)
printf("bpf: %s%d attached\n", ifp->if_name, ifp->if_unit);
1994-05-24 10:09:53 +00:00
}
1997-09-16 11:44:05 +00:00
static void bpf_drvinit __P((void *unused));
1997-09-16 11:44:05 +00:00
static void
bpf_drvinit(unused)
void *unused;
{
cdevsw_add(&bpf_cdevsw);
}
SYSINIT(bpfdev,SI_SUB_DRIVERS,SI_ORDER_MIDDLE+CDEV_MAJOR,bpf_drvinit,NULL)
1999-07-06 19:23:32 +00:00
#else /* !BPF */
/*
* NOP stubs to allow bpf-using drivers to load and function.
*
* A 'better' implementation would allow the core bpf functionality
* to be loaded at runtime.
*/
void
bpf_tap(ifp, pkt, pktlen)
struct ifnet *ifp;
register u_char *pkt;
register u_int pktlen;
{
}
void
bpf_mtap(ifp, m)
struct ifnet *ifp;
struct mbuf *m;
{
}
void
bpfattach(ifp, dlt, hdrlen)
struct ifnet *ifp;
u_int dlt, hdrlen;
{
}
u_int
bpf_filter(pc, p, wirelen, buflen)
register const struct bpf_insn *pc;
register u_char *p;
u_int wirelen;
register u_int buflen;
{
return -1; /* "no filter" behaviour */
}
#endif /* !BPF */