906 lines
19 KiB
C
906 lines
19 KiB
C
/* $NetBSD: ip_sfil.c,v 1.5 2000/05/23 06:16:44 veego Exp $ */
|
|
|
|
/*
|
|
* Copyright (C) 1993-2000 by Darren Reed.
|
|
*
|
|
* Redistribution and use in source and binary forms are permitted
|
|
* provided that this notice is preserved and due credit is given
|
|
* to the original author and the contributors.
|
|
*
|
|
* I hate legaleese, don't you ?
|
|
*/
|
|
#if !defined(lint)
|
|
static const char sccsid[] = "%W% %G% (C) 1993-2000 Darren Reed";
|
|
static const char rcsid[] = "@(#)Id: ip_sfil.c,v 2.23.2.2 2000/05/22 10:26:14 darrenr Exp";
|
|
#endif
|
|
|
|
#include <sys/types.h>
|
|
#include <sys/errno.h>
|
|
#include <sys/param.h>
|
|
#include <sys/cpuvar.h>
|
|
#include <sys/open.h>
|
|
#include <sys/ioctl.h>
|
|
#include <sys/filio.h>
|
|
#include <sys/systm.h>
|
|
#include <sys/cred.h>
|
|
#include <sys/ddi.h>
|
|
#include <sys/sunddi.h>
|
|
#include <sys/ksynch.h>
|
|
#include <sys/kmem.h>
|
|
#include <sys/mkdev.h>
|
|
#include <sys/protosw.h>
|
|
#include <sys/socket.h>
|
|
#include <sys/dditypes.h>
|
|
#include <sys/cmn_err.h>
|
|
#include <net/if.h>
|
|
#include <net/af.h>
|
|
#include <net/route.h>
|
|
#include <netinet/in.h>
|
|
#include <netinet/in_systm.h>
|
|
#include <netinet/ip.h>
|
|
#include <netinet/ip_var.h>
|
|
#include <netinet/tcp.h>
|
|
#include <netinet/udp.h>
|
|
#include <netinet/tcpip.h>
|
|
#include <netinet/ip_icmp.h>
|
|
#include "ip_compat.h"
|
|
#ifdef USE_INET6
|
|
# include <netinet/icmp6.h>
|
|
#endif
|
|
#include "ip_fil.h"
|
|
#include "ip_state.h"
|
|
#include "ip_nat.h"
|
|
#include "ip_frag.h"
|
|
#include "ip_auth.h"
|
|
#include <inet/ip_ire.h>
|
|
#ifndef MIN
|
|
#define MIN(a,b) (((a)<(b))?(a):(b))
|
|
#endif
|
|
|
|
|
|
extern fr_flags, fr_active;
|
|
|
|
int fr_running = 0;
|
|
int ipl_unreach = ICMP_UNREACH_HOST;
|
|
u_long ipl_frouteok[2] = {0, 0};
|
|
static int frzerostats __P((caddr_t));
|
|
|
|
static int frrequest __P((minor_t, int, caddr_t, int));
|
|
kmutex_t ipl_mutex, ipf_authmx, ipf_rw, ipf_hostmap;
|
|
KRWLOCK_T ipf_mutex, ipfs_mutex, ipf_solaris;
|
|
KRWLOCK_T ipf_frag, ipf_state, ipf_nat, ipf_natfrag, ipf_auth;
|
|
kcondvar_t iplwait, ipfauthwait;
|
|
|
|
|
|
int ipldetach()
|
|
{
|
|
int i;
|
|
|
|
#ifdef IPFDEBUG
|
|
cmn_err(CE_CONT, "ipldetach()\n");
|
|
#endif
|
|
#ifdef IPFILTER_LOG
|
|
for (i = IPL_LOGMAX; i >= 0; i--)
|
|
ipflog_clear(i);
|
|
#endif
|
|
i = FR_INQUE|FR_OUTQUE;
|
|
(void) frflush(IPL_LOGIPF, FR_INQUE|FR_OUTQUE);
|
|
ipfr_unload();
|
|
fr_stateunload();
|
|
ip_natunload();
|
|
cv_destroy(&iplwait);
|
|
cv_destroy(&ipfauthwait);
|
|
mutex_destroy(&ipf_hostmap);
|
|
mutex_destroy(&ipf_authmx);
|
|
mutex_destroy(&ipl_mutex);
|
|
mutex_destroy(&ipf_rw);
|
|
RW_DESTROY(&ipf_mutex);
|
|
RW_DESTROY(&ipf_frag);
|
|
RW_DESTROY(&ipf_state);
|
|
RW_DESTROY(&ipf_natfrag);
|
|
RW_DESTROY(&ipf_nat);
|
|
RW_DESTROY(&ipf_auth);
|
|
RW_DESTROY(&ipfs_mutex);
|
|
/* NOTE: This lock is acquired in ipf_detach */
|
|
RWLOCK_EXIT(&ipf_solaris);
|
|
RW_DESTROY(&ipf_solaris);
|
|
return 0;
|
|
}
|
|
|
|
|
|
int iplattach __P((void))
|
|
{
|
|
#ifdef IPFDEBUG
|
|
cmn_err(CE_CONT, "iplattach()\n");
|
|
#endif
|
|
bzero((char *)frcache, sizeof(frcache));
|
|
mutex_init(&ipf_rw, "ipf rw mutex", MUTEX_DRIVER, NULL);
|
|
mutex_init(&ipl_mutex, "ipf log mutex", MUTEX_DRIVER, NULL);
|
|
mutex_init(&ipf_authmx, "ipf auth log mutex", MUTEX_DRIVER, NULL);
|
|
mutex_init(&ipf_hostmap, "ipf hostmap mutex", MUTEX_DRIVER, NULL);
|
|
RWLOCK_INIT(&ipf_solaris, "ipf filter load/unload mutex", NULL);
|
|
RWLOCK_INIT(&ipf_mutex, "ipf filter rwlock", NULL);
|
|
RWLOCK_INIT(&ipfs_mutex, "ipf solaris mutex", NULL);
|
|
RWLOCK_INIT(&ipf_frag, "ipf fragment rwlock", NULL);
|
|
RWLOCK_INIT(&ipf_state, "ipf IP state rwlock", NULL);
|
|
RWLOCK_INIT(&ipf_nat, "ipf IP NAT rwlock", NULL);
|
|
RWLOCK_INIT(&ipf_natfrag, "ipf IP NAT-Frag rwlock", NULL);
|
|
RWLOCK_INIT(&ipf_auth, "ipf IP User-Auth rwlock", NULL);
|
|
cv_init(&iplwait, "ipl condvar", CV_DRIVER, NULL);
|
|
cv_init(&ipfauthwait, "ipf auth condvar", CV_DRIVER, NULL);
|
|
#ifdef IPFILTER_LOG
|
|
ipflog_init();
|
|
#endif
|
|
if (nat_init() == -1)
|
|
return -1;
|
|
if (fr_stateinit() == -1)
|
|
return -1;
|
|
if (appr_init() == -1)
|
|
return -1;
|
|
return 0;
|
|
}
|
|
|
|
|
|
static int frzerostats(data)
|
|
caddr_t data;
|
|
{
|
|
friostat_t fio;
|
|
int error;
|
|
|
|
fr_getstat(&fio);
|
|
error = IWCOPYPTR((caddr_t)&fio, data, sizeof(fio));
|
|
if (error)
|
|
return EFAULT;
|
|
|
|
bzero((char *)frstats, sizeof(*frstats) * 2);
|
|
|
|
return 0;
|
|
}
|
|
|
|
|
|
/*
|
|
* Filter ioctl interface.
|
|
*/
|
|
int iplioctl(dev, cmd, data, mode, cp, rp)
|
|
dev_t dev;
|
|
int cmd;
|
|
#if SOLARIS2 >= 7
|
|
intptr_t data;
|
|
#else
|
|
int *data;
|
|
#endif
|
|
int mode;
|
|
cred_t *cp;
|
|
int *rp;
|
|
{
|
|
int error = 0, tmp;
|
|
minor_t unit;
|
|
|
|
#ifdef IPFDEBUG
|
|
cmn_err(CE_CONT, "iplioctl(%x,%x,%x,%d,%x,%d)\n",
|
|
dev, cmd, data, mode, cp, rp);
|
|
#endif
|
|
unit = getminor(dev);
|
|
if (IPL_LOGMAX < unit)
|
|
return ENXIO;
|
|
|
|
if (fr_running <= 0)
|
|
return 0;
|
|
|
|
READ_ENTER(&ipf_solaris);
|
|
if (unit == IPL_LOGNAT) {
|
|
error = nat_ioctl((caddr_t)data, cmd, mode);
|
|
RWLOCK_EXIT(&ipf_solaris);
|
|
return error;
|
|
}
|
|
if (unit == IPL_LOGSTATE) {
|
|
error = fr_state_ioctl((caddr_t)data, cmd, mode);
|
|
RWLOCK_EXIT(&ipf_solaris);
|
|
return error;
|
|
}
|
|
if (unit == IPL_LOGAUTH) {
|
|
error = fr_auth_ioctl((caddr_t)data, cmd, NULL, NULL);
|
|
RWLOCK_EXIT(&ipf_solaris);
|
|
return error;
|
|
}
|
|
|
|
switch (cmd) {
|
|
case SIOCFRENB :
|
|
{
|
|
u_int enable;
|
|
|
|
if (!(mode & FWRITE))
|
|
error = EPERM;
|
|
else
|
|
error = IRCOPY((caddr_t)data, (caddr_t)&enable,
|
|
sizeof(enable));
|
|
break;
|
|
}
|
|
case SIOCSETFF :
|
|
if (!(mode & FWRITE))
|
|
error = EPERM;
|
|
else {
|
|
WRITE_ENTER(&ipf_mutex);
|
|
error = IRCOPY((caddr_t)data, (caddr_t)&fr_flags,
|
|
sizeof(fr_flags));
|
|
RWLOCK_EXIT(&ipf_mutex);
|
|
}
|
|
break;
|
|
case SIOCGETFF :
|
|
error = IWCOPY((caddr_t)&fr_flags, (caddr_t)data,
|
|
sizeof(fr_flags));
|
|
break;
|
|
case SIOCINAFR :
|
|
case SIOCRMAFR :
|
|
case SIOCADAFR :
|
|
case SIOCZRLST :
|
|
if (!(mode & FWRITE))
|
|
error = EPERM;
|
|
else
|
|
error = frrequest(unit, cmd, (caddr_t)data, fr_active);
|
|
break;
|
|
case SIOCINIFR :
|
|
case SIOCRMIFR :
|
|
case SIOCADIFR :
|
|
if (!(mode & FWRITE))
|
|
error = EPERM;
|
|
else
|
|
error = frrequest(unit, cmd, (caddr_t)data,
|
|
1 - fr_active);
|
|
break;
|
|
case SIOCSWAPA :
|
|
if (!(mode & FWRITE))
|
|
error = EPERM;
|
|
else {
|
|
WRITE_ENTER(&ipf_mutex);
|
|
bzero((char *)frcache, sizeof(frcache[0]) * 2);
|
|
error = IWCOPY((caddr_t)&fr_active, (caddr_t)data,
|
|
sizeof(fr_active));
|
|
fr_active = 1 - fr_active;
|
|
RWLOCK_EXIT(&ipf_mutex);
|
|
}
|
|
break;
|
|
case SIOCGETFS :
|
|
{
|
|
friostat_t fio;
|
|
|
|
READ_ENTER(&ipf_mutex);
|
|
fr_getstat(&fio);
|
|
RWLOCK_EXIT(&ipf_mutex);
|
|
error = IWCOPYPTR((caddr_t)&fio, (caddr_t)data, sizeof(fio));
|
|
if (error)
|
|
error = EFAULT;
|
|
break;
|
|
}
|
|
case SIOCFRZST :
|
|
if (!(mode & FWRITE))
|
|
error = EPERM;
|
|
else
|
|
error = frzerostats((caddr_t)data);
|
|
break;
|
|
case SIOCIPFFL :
|
|
if (!(mode & FWRITE))
|
|
error = EPERM;
|
|
else {
|
|
error = IRCOPY((caddr_t)data, (caddr_t)&tmp,
|
|
sizeof(tmp));
|
|
if (!error) {
|
|
tmp = frflush(unit, tmp);
|
|
error = IWCOPY((caddr_t)&tmp, (caddr_t)data,
|
|
sizeof(tmp));
|
|
}
|
|
}
|
|
break;
|
|
case SIOCSTLCK :
|
|
error = IRCOPY((caddr_t)data, (caddr_t)&tmp, sizeof(tmp));
|
|
if (!error) {
|
|
fr_state_lock = tmp;
|
|
fr_nat_lock = tmp;
|
|
fr_frag_lock = tmp;
|
|
fr_auth_lock = tmp;
|
|
} else
|
|
error = EFAULT;
|
|
break;
|
|
#ifdef IPFILTER_LOG
|
|
case SIOCIPFFB :
|
|
if (!(mode & FWRITE))
|
|
error = EPERM;
|
|
else {
|
|
tmp = ipflog_clear(unit);
|
|
error = IWCOPY((caddr_t)&tmp, (caddr_t)data,
|
|
sizeof(tmp));
|
|
}
|
|
break;
|
|
#endif /* IPFILTER_LOG */
|
|
case SIOCFRSYN :
|
|
if (!(mode & FWRITE))
|
|
error = EPERM;
|
|
else
|
|
error = ipfsync();
|
|
break;
|
|
case SIOCGFRST :
|
|
error = IWCOPYPTR((caddr_t)ipfr_fragstats(), (caddr_t)data,
|
|
sizeof(ipfrstat_t));
|
|
if (error)
|
|
error = EFAULT;
|
|
break;
|
|
case FIONREAD :
|
|
{
|
|
#ifdef IPFILTER_LOG
|
|
int copy = (int)iplused[IPL_LOGIPF];
|
|
|
|
error = IWCOPY((caddr_t)©, (caddr_t)data, sizeof(copy));
|
|
#endif
|
|
break;
|
|
}
|
|
default :
|
|
error = EINVAL;
|
|
break;
|
|
}
|
|
RWLOCK_EXIT(&ipf_solaris);
|
|
return error;
|
|
}
|
|
|
|
|
|
ill_t *get_unit(name, v)
|
|
char *name;
|
|
int v;
|
|
{
|
|
size_t len = strlen(name) + 1; /* includes \0 */
|
|
ill_t *il;
|
|
int sap;
|
|
|
|
if (v == 4)
|
|
sap = 0x0800;
|
|
else if (v == 6)
|
|
sap = 0x86dd;
|
|
else
|
|
return NULL;
|
|
for (il = ill_g_head; il; il = il->ill_next)
|
|
if ((len == il->ill_name_length) && (il->ill_sap == sap) &&
|
|
!strncmp(il->ill_name, name, len))
|
|
return il;
|
|
return NULL;
|
|
}
|
|
|
|
|
|
static int frrequest(unit, req, data, set)
|
|
minor_t unit;
|
|
int req, set;
|
|
caddr_t data;
|
|
{
|
|
register frentry_t *fp, *f, **fprev;
|
|
register frentry_t **ftail;
|
|
frentry_t fr;
|
|
frdest_t *fdp;
|
|
frgroup_t *fg = NULL;
|
|
u_int *p, *pp;
|
|
int error = 0, in;
|
|
u_32_t group;
|
|
ill_t *ill;
|
|
ipif_t *ipif;
|
|
ire_t *ire;
|
|
|
|
fp = &fr;
|
|
error = IRCOPYPTR(data, (caddr_t)fp, sizeof(*fp));
|
|
if (error)
|
|
return EFAULT;
|
|
fp->fr_ref = 0;
|
|
#if SOLARIS2 >= 8
|
|
if (fp->fr_v == 4)
|
|
fp->fr_sap = IP_DL_SAP;
|
|
else if (fp->fr_v == 6)
|
|
fp->fr_sap = IP6_DL_SAP;
|
|
else
|
|
return EINVAL;
|
|
#else
|
|
fp->fr_sap = 0;
|
|
#endif
|
|
|
|
WRITE_ENTER(&ipf_mutex);
|
|
/*
|
|
* Check that the group number does exist and that if a head group
|
|
* has been specified, doesn't exist.
|
|
*/
|
|
if ((req != SIOCZRLST) && fp->fr_grhead &&
|
|
fr_findgroup(fp->fr_grhead, fp->fr_flags, unit, set, NULL)) {
|
|
error = EEXIST;
|
|
goto out;
|
|
}
|
|
if ((req != SIOCZRLST) && fp->fr_group &&
|
|
!fr_findgroup(fp->fr_group, fp->fr_flags, unit, set, NULL)) {
|
|
error = ESRCH;
|
|
goto out;
|
|
}
|
|
|
|
in = (fp->fr_flags & FR_INQUE) ? 0 : 1;
|
|
|
|
if (unit == IPL_LOGAUTH)
|
|
ftail = fprev = &ipauth;
|
|
else if ((fp->fr_flags & FR_ACCOUNT) && (fp->fr_v == 4))
|
|
ftail = fprev = &ipacct[in][set];
|
|
else if ((fp->fr_flags & (FR_OUTQUE|FR_INQUE)) && (fp->fr_v == 4))
|
|
ftail = fprev = &ipfilter[in][set];
|
|
#ifdef USE_INET6
|
|
else if ((fp->fr_flags & FR_ACCOUNT) && (fp->fr_v == 6))
|
|
ftail = fprev = &ipacct6[in][set];
|
|
else if ((fp->fr_flags & (FR_OUTQUE|FR_INQUE)) && (fp->fr_v == 6))
|
|
ftail = fprev = &ipfilter6[in][set];
|
|
#endif
|
|
else {
|
|
error = ESRCH;
|
|
goto out;
|
|
}
|
|
|
|
group = fp->fr_group;
|
|
if (group != NULL) {
|
|
fg = fr_findgroup(group, fp->fr_flags, unit, set, NULL);
|
|
if (fg == NULL) {
|
|
error = ESRCH;
|
|
goto out;
|
|
}
|
|
ftail = fprev = fg->fg_start;
|
|
}
|
|
|
|
bzero((char *)frcache, sizeof(frcache[0]) * 2);
|
|
|
|
if (*fp->fr_ifname) {
|
|
fp->fr_ifa = (void *)get_unit((char *)fp->fr_ifname,
|
|
(int)fp->fr_v);
|
|
if (!fp->fr_ifa)
|
|
fp->fr_ifa = (struct ifnet *)-1;
|
|
}
|
|
|
|
fdp = &fp->fr_dif;
|
|
fp->fr_flags &= ~FR_DUP;
|
|
if (*fdp->fd_ifname) {
|
|
ill = get_unit(fdp->fd_ifname, (int)fp->fr_v);
|
|
if (!ill)
|
|
ire = (ire_t *)-1;
|
|
else if ((ipif = ill->ill_ipif) && (fp->fr_v == 4)) {
|
|
#if SOLARIS2 > 5
|
|
ire = ire_ctable_lookup(ipif->ipif_local_addr, 0,
|
|
IRE_LOCAL, NULL, NULL,
|
|
MATCH_IRE_TYPE);
|
|
#else
|
|
ire = ire_lookup_myaddr(ipif->ipif_local_addr);
|
|
#endif
|
|
if (!ire)
|
|
ire = (ire_t *)-1;
|
|
else
|
|
fp->fr_flags |= FR_DUP;
|
|
}
|
|
#ifdef USE_INET6
|
|
else if ((ipif = ill->ill_ipif) && (fp->fr_v == 6)) {
|
|
ire = ire_ctable_lookup_v6(&ipif->ipif_v6lcl_addr, 0,
|
|
IRE_LOCAL, NULL, NULL,
|
|
MATCH_IRE_TYPE);
|
|
if (!ire)
|
|
ire = (ire_t *)-1;
|
|
else
|
|
fp->fr_flags |= FR_DUP;
|
|
}
|
|
#endif
|
|
fdp->fd_ifp = (struct ifnet *)ire;
|
|
}
|
|
|
|
fdp = &fp->fr_tif;
|
|
if (*fdp->fd_ifname) {
|
|
ill = get_unit(fdp->fd_ifname, (int)fp->fr_v);
|
|
if (!ill)
|
|
ire = (ire_t *)-1;
|
|
else if ((ipif = ill->ill_ipif) && (fp->fr_v == 4)) {
|
|
#if SOLARIS2 > 5
|
|
ire = ire_ctable_lookup(ipif->ipif_local_addr, 0,
|
|
IRE_LOCAL, NULL, NULL,
|
|
MATCH_IRE_TYPE);
|
|
#else
|
|
ire = ire_lookup_myaddr(ipif->ipif_local_addr);
|
|
#endif
|
|
if (!ire)
|
|
ire = (ire_t *)-1;
|
|
}
|
|
#ifdef USE_INET6
|
|
else if ((ipif = ill->ill_ipif) && (fp->fr_v == 6)) {
|
|
ire = ire_ctable_lookup_v6(&ipif->ipif_v6lcl_addr, 0,
|
|
IRE_LOCAL, NULL, NULL,
|
|
MATCH_IRE_TYPE);
|
|
if (!ire)
|
|
ire = (ire_t *)-1;
|
|
}
|
|
#endif
|
|
fdp->fd_ifp = (struct ifnet *)ire;
|
|
}
|
|
|
|
/*
|
|
* Look for a matching filter rule, but don't include the next or
|
|
* interface pointer in the comparison (fr_next, fr_ifa).
|
|
*/
|
|
for (fp->fr_cksum = 0, p = (u_int *)&fp->fr_ip, pp = &fp->fr_cksum;
|
|
p != pp; p++)
|
|
fp->fr_cksum += *p;
|
|
|
|
for (; (f = *ftail); ftail = &f->fr_next)
|
|
if ((fp->fr_cksum == f->fr_cksum) &&
|
|
!bcmp((char *)&f->fr_ip, (char *)&fp->fr_ip, FR_CMPSIZ))
|
|
break;
|
|
|
|
/*
|
|
* If zero'ing statistics, copy current to caller and zero.
|
|
*/
|
|
if (req == SIOCZRLST) {
|
|
if (!f) {
|
|
error = ESRCH;
|
|
goto out;
|
|
}
|
|
MUTEX_DOWNGRADE(&ipf_mutex);
|
|
error = IWCOPYPTR((caddr_t)f, data, sizeof(*f));
|
|
if (error) {
|
|
error = EFAULT;
|
|
goto out;
|
|
}
|
|
f->fr_hits = 0;
|
|
f->fr_bytes = 0;
|
|
goto out;
|
|
}
|
|
|
|
if (!f) {
|
|
if (req != SIOCINAFR && req != SIOCINIFR)
|
|
while ((f = *ftail))
|
|
ftail = &f->fr_next;
|
|
else {
|
|
if (fp->fr_hits) {
|
|
ftail = fprev;
|
|
while (--fp->fr_hits && (f = *ftail))
|
|
ftail = &f->fr_next;
|
|
}
|
|
f = NULL;
|
|
}
|
|
}
|
|
|
|
if (req == SIOCRMAFR || req == SIOCRMIFR) {
|
|
if (!f)
|
|
error = ESRCH;
|
|
else {
|
|
/*
|
|
* Only return EBUSY if there is a group list, else
|
|
* it's probably just state information referencing
|
|
* the rule.
|
|
*/
|
|
if ((f->fr_ref > 1) && f->fr_grp) {
|
|
error = EBUSY;
|
|
goto out;
|
|
}
|
|
if (fg && fg->fg_head)
|
|
fg->fg_head->fr_ref--;
|
|
if (unit == IPL_LOGAUTH) {
|
|
error = fr_auth_ioctl(data, req, fp, ftail);
|
|
goto out;
|
|
}
|
|
if (f->fr_grhead)
|
|
fr_delgroup(f->fr_grhead, fp->fr_flags,
|
|
unit, set);
|
|
fixskip(fprev, f, -1);
|
|
*ftail = f->fr_next;
|
|
f->fr_next = NULL;
|
|
if (f->fr_ref == 0)
|
|
KFREE(f);
|
|
}
|
|
} else {
|
|
if (f) {
|
|
error = EEXIST;
|
|
} else {
|
|
if (unit == IPL_LOGAUTH) {
|
|
error = fr_auth_ioctl(data, req, fp, ftail);
|
|
goto out;
|
|
}
|
|
KMALLOC(f, frentry_t *);
|
|
if (f != NULL) {
|
|
if (fg && fg->fg_head)
|
|
fg->fg_head->fr_ref++;
|
|
bcopy((char *)fp, (char *)f, sizeof(*f));
|
|
f->fr_ref = 1;
|
|
f->fr_hits = 0;
|
|
f->fr_next = *ftail;
|
|
*ftail = f;
|
|
if (req == SIOCINIFR || req == SIOCINAFR)
|
|
fixskip(fprev, f, 1);
|
|
f->fr_grp = NULL;
|
|
group = f->fr_grhead;
|
|
if (group != NULL)
|
|
fg = fr_addgroup(group, f, unit, set);
|
|
} else
|
|
error = ENOMEM;
|
|
}
|
|
}
|
|
out:
|
|
RWLOCK_EXIT(&ipf_mutex);
|
|
return (error);
|
|
}
|
|
|
|
|
|
/*
|
|
* routines below for saving IP headers to buffer
|
|
*/
|
|
int iplopen(devp, flags, otype, cred)
|
|
dev_t *devp;
|
|
int flags, otype;
|
|
cred_t *cred;
|
|
{
|
|
minor_t min = getminor(*devp);
|
|
|
|
#ifdef IPFDEBUG
|
|
cmn_err(CE_CONT, "iplopen(%x,%x,%x,%x)\n", devp, flags, otype, cred);
|
|
#endif
|
|
if ((fr_running <= 0) || !(otype & OTYP_CHR))
|
|
return ENXIO;
|
|
min = (IPL_LOGMAX < min) ? ENXIO : 0;
|
|
return min;
|
|
}
|
|
|
|
|
|
int iplclose(dev, flags, otype, cred)
|
|
dev_t dev;
|
|
int flags, otype;
|
|
cred_t *cred;
|
|
{
|
|
minor_t min = getminor(dev);
|
|
|
|
#ifdef IPFDEBUG
|
|
cmn_err(CE_CONT, "iplclose(%x,%x,%x,%x)\n", dev, flags, otype, cred);
|
|
#endif
|
|
min = (IPL_LOGMAX < min) ? ENXIO : 0;
|
|
return min;
|
|
}
|
|
|
|
#ifdef IPFILTER_LOG
|
|
/*
|
|
* iplread/ipllog
|
|
* both of these must operate with at least splnet() lest they be
|
|
* called during packet processing and cause an inconsistancy to appear in
|
|
* the filter lists.
|
|
*/
|
|
int iplread(dev, uio, cp)
|
|
dev_t dev;
|
|
register struct uio *uio;
|
|
cred_t *cp;
|
|
{
|
|
#ifdef IPFDEBUG
|
|
cmn_err(CE_CONT, "iplread(%x,%x,%x)\n", dev, uio, cp);
|
|
#endif
|
|
return ipflog_read(getminor(dev), uio);
|
|
}
|
|
#endif /* IPFILTER_LOG */
|
|
|
|
|
|
/*
|
|
* send_reset - this could conceivably be a call to tcp_respond(), but that
|
|
* requires a large amount of setting up and isn't any more efficient.
|
|
*/
|
|
int send_reset(oip, fin)
|
|
ip_t *oip;
|
|
fr_info_t *fin;
|
|
{
|
|
tcphdr_t *tcp, *tcp2;
|
|
int tlen = 0, hlen;
|
|
mblk_t *m;
|
|
#ifdef USE_INET6
|
|
ip6_t *ip6, *oip6 = (ip6_t *)oip;
|
|
#endif
|
|
ip_t *ip;
|
|
|
|
tcp = (struct tcphdr *)fin->fin_dp;
|
|
if (tcp->th_flags & TH_RST)
|
|
return -1;
|
|
if (tcp->th_flags & TH_SYN)
|
|
tlen = 1;
|
|
#ifdef USE_INET6
|
|
if (fin->fin_v == 6)
|
|
hlen = sizeof(ip6_t);
|
|
else
|
|
#endif
|
|
hlen = sizeof(ip_t);
|
|
hlen += sizeof(*tcp2);
|
|
if ((m = (mblk_t *)allocb(hlen + 16, BPRI_HI)) == NULL)
|
|
return -1;
|
|
|
|
m->b_rptr += 16;
|
|
MTYPE(m) = M_DATA;
|
|
m->b_wptr = m->b_rptr + hlen;
|
|
bzero((char *)m->b_rptr, hlen);
|
|
tcp2 = (struct tcphdr *)(m->b_rptr + hlen - sizeof(*tcp2));
|
|
tcp2->th_dport = tcp->th_sport;
|
|
tcp2->th_sport = tcp->th_dport;
|
|
tcp2->th_ack = htonl(ntohl(tcp->th_seq) + tlen);
|
|
tcp2->th_seq = tcp->th_ack;
|
|
tcp2->th_off = sizeof(struct tcphdr) >> 2;
|
|
tcp2->th_flags = TH_RST|TH_ACK;
|
|
|
|
/*
|
|
* This is to get around a bug in the Solaris 2.4/2.5 TCP checksum
|
|
* computation that is done by their put routine.
|
|
*/
|
|
tcp2->th_sum = htons(0x14);
|
|
#ifdef USE_INET6
|
|
if (fin->fin_v == 6) {
|
|
ip6 = (ip6_t *)m->b_rptr;
|
|
ip6->ip6_src = oip6->ip6_dst;
|
|
ip6->ip6_dst = oip6->ip6_src;
|
|
ip6->ip6_plen = htons(sizeof(*tcp));
|
|
ip6->ip6_nxt = IPPROTO_TCP;
|
|
} else
|
|
#endif
|
|
{
|
|
ip = (ip_t *)m->b_rptr;
|
|
ip->ip_src.s_addr = oip->ip_dst.s_addr;
|
|
ip->ip_dst.s_addr = oip->ip_src.s_addr;
|
|
ip->ip_hl = sizeof(*ip) >> 2;
|
|
ip->ip_p = IPPROTO_TCP;
|
|
ip->ip_len = htons(sizeof(*ip) + sizeof(*tcp));
|
|
ip->ip_tos = oip->ip_tos;
|
|
}
|
|
return send_ip(fin, m);
|
|
}
|
|
|
|
|
|
int send_ip(fin, m)
|
|
fr_info_t *fin;
|
|
mblk_t *m;
|
|
{
|
|
RWLOCK_EXIT(&ipfs_mutex);
|
|
RWLOCK_EXIT(&ipf_solaris);
|
|
#ifdef USE_INET6
|
|
if (fin->fin_v == 6) {
|
|
ip6_t *ip6;
|
|
|
|
ip6 = (ip6_t *)m->b_rptr;
|
|
ip6->ip6_flow = 0;
|
|
ip6->ip6_vfc = 0x60;
|
|
ip6->ip6_hlim = 127;
|
|
ip_wput_v6(((qif_t *)fin->fin_qif)->qf_ill->ill_wq, m);
|
|
} else
|
|
#endif
|
|
{
|
|
ip_t *ip;
|
|
|
|
ip = (ip_t *)m->b_rptr;
|
|
ip->ip_v = IPVERSION;
|
|
ip->ip_ttl = 60;
|
|
ip_wput(((qif_t *)fin->fin_qif)->qf_ill->ill_wq, m);
|
|
}
|
|
READ_ENTER(&ipf_solaris);
|
|
READ_ENTER(&ipfs_mutex);
|
|
return 0;
|
|
}
|
|
|
|
|
|
int send_icmp_err(oip, type, fin, dst)
|
|
ip_t *oip;
|
|
int type;
|
|
fr_info_t *fin;
|
|
int dst;
|
|
{
|
|
struct in_addr dst4;
|
|
struct icmp *icmp;
|
|
mblk_t *m, *mb;
|
|
int hlen, code;
|
|
qif_t *qif;
|
|
u_short sz;
|
|
#ifdef USE_INET6
|
|
ip6_t *ip6, *oip6;
|
|
#endif
|
|
ip_t *ip;
|
|
|
|
if ((type < 0) || (type > ICMP_MAXTYPE))
|
|
return -1;
|
|
|
|
code = fin->fin_icode;
|
|
#ifdef USE_INET6
|
|
if ((code < 0) || (code > sizeof(icmptoicmp6unreach)/sizeof(int)))
|
|
return -1;
|
|
#endif
|
|
|
|
qif = fin->fin_qif;
|
|
m = fin->fin_qfm;
|
|
|
|
#ifdef USE_INET6
|
|
if (oip->ip_v == 6) {
|
|
oip6 = (ip6_t *)oip;
|
|
sz = sizeof(ip6_t);
|
|
sz += MIN(m->b_wptr - m->b_rptr, 512);
|
|
hlen = sizeof(ip6_t);
|
|
type = icmptoicmp6types[type];
|
|
if (type == ICMP6_DST_UNREACH)
|
|
code = icmptoicmp6unreach[code];
|
|
} else
|
|
#endif
|
|
{
|
|
if ((oip->ip_p == IPPROTO_ICMP) &&
|
|
!(fin->fin_fi.fi_fl & FI_SHORT))
|
|
switch (ntohs(fin->fin_data[0]) >> 8)
|
|
{
|
|
case ICMP_ECHO :
|
|
case ICMP_TSTAMP :
|
|
case ICMP_IREQ :
|
|
case ICMP_MASKREQ :
|
|
break;
|
|
default :
|
|
return 0;
|
|
}
|
|
|
|
sz = sizeof(ip_t) * 2;
|
|
sz += 8; /* 64 bits of data */
|
|
hlen = sz;
|
|
}
|
|
|
|
sz += offsetof(struct icmp, icmp_ip);
|
|
if ((mb = (mblk_t *)allocb((size_t)sz + 16, BPRI_HI)) == NULL)
|
|
return -1;
|
|
MTYPE(mb) = M_DATA;
|
|
mb->b_rptr += 16;
|
|
mb->b_wptr = mb->b_rptr + sz;
|
|
bzero((char *)mb->b_rptr, (size_t)sz);
|
|
icmp = (struct icmp *)(mb->b_rptr + sizeof(*ip));
|
|
icmp->icmp_type = type;
|
|
icmp->icmp_code = code;
|
|
icmp->icmp_cksum = 0;
|
|
|
|
#ifdef USE_INET6
|
|
if (oip->ip_v == 6) {
|
|
struct in6_addr dst6;
|
|
int csz;
|
|
|
|
if (dst == 0) {
|
|
if (fr_ifpaddr(6, ((qif_t *)fin->fin_qif)->qf_ill,
|
|
(struct in_addr *)&dst6) == -1)
|
|
return -1;
|
|
} else
|
|
dst6 = oip6->ip6_dst;
|
|
|
|
csz = sz;
|
|
sz -= sizeof(ip6_t);
|
|
ip6 = (ip6_t *)mb->b_rptr;
|
|
ip6->ip6_flow = 0;
|
|
ip6->ip6_vfc = 0x60;
|
|
ip6->ip6_hlim = 127;
|
|
ip6->ip6_plen = htons(sz);
|
|
ip6->ip6_nxt = IPPROTO_ICMPV6;
|
|
ip6->ip6_src = dst6;
|
|
ip6->ip6_dst = oip6->ip6_src;
|
|
sz -= offsetof(struct icmp, icmp_ip);
|
|
bcopy((char *)m->b_rptr, (char *)&icmp->icmp_ip, sz);
|
|
icmp->icmp_cksum = csz - sizeof(ip6_t);
|
|
} else
|
|
#endif
|
|
{
|
|
ip = (ip_t *)mb->b_rptr;
|
|
ip->ip_v = IPVERSION;
|
|
ip->ip_hl = (sizeof(*ip) >> 2);
|
|
ip->ip_p = IPPROTO_ICMP;
|
|
ip->ip_id = oip->ip_id;
|
|
ip->ip_sum = 0;
|
|
ip->ip_ttl = 60;
|
|
ip->ip_tos = oip->ip_tos;
|
|
ip->ip_len = (u_short)htons(sz);
|
|
if (dst == 0) {
|
|
if (fr_ifpaddr(4, ((qif_t *)fin->fin_qif)->qf_ill,
|
|
&dst4) == -1)
|
|
return -1;
|
|
} else
|
|
dst4 = oip->ip_dst;
|
|
ip->ip_src = dst4;
|
|
ip->ip_dst = oip->ip_src;
|
|
bcopy((char *)oip, (char *)&icmp->icmp_ip, sizeof(*oip));
|
|
bcopy((char *)oip + (oip->ip_hl << 2),
|
|
(char *)&icmp->icmp_ip + sizeof(*oip), 8);
|
|
icmp->icmp_cksum = ipf_cksum((u_short *)icmp,
|
|
sizeof(*icmp) + 8);
|
|
}
|
|
|
|
/*
|
|
* Need to exit out of these so we don't recursively call rw_enter
|
|
* from fr_qout.
|
|
*/
|
|
return send_ip(fin, mb);
|
|
}
|