2c1ec831c9
in implicitly via if.h -> if_var.h pollution. Sponsored by: Netflix Sponsored by: Nginx, Inc.
2198 lines
54 KiB
C
2198 lines
54 KiB
C
/*-
|
|
* Copyright (c) 2002 Cedric Berger
|
|
* All rights reserved.
|
|
*
|
|
* Redistribution and use in source and binary forms, with or without
|
|
* modification, are permitted provided that the following conditions
|
|
* are met:
|
|
*
|
|
* - Redistributions of source code must retain the above copyright
|
|
* notice, this list of conditions and the following disclaimer.
|
|
* - 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.
|
|
*
|
|
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS 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
|
|
* COPYRIGHT HOLDERS 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.
|
|
*
|
|
* $OpenBSD: pf_table.c,v 1.79 2008/10/08 06:24:50 mcbride Exp $
|
|
*/
|
|
|
|
#include <sys/cdefs.h>
|
|
__FBSDID("$FreeBSD$");
|
|
|
|
#include "opt_inet.h"
|
|
#include "opt_inet6.h"
|
|
|
|
#include <sys/param.h>
|
|
#include <sys/kernel.h>
|
|
#include <sys/lock.h>
|
|
#include <sys/malloc.h>
|
|
#include <sys/mbuf.h>
|
|
#include <sys/mutex.h>
|
|
#include <sys/refcount.h>
|
|
#include <sys/rwlock.h>
|
|
#include <sys/socket.h>
|
|
#include <vm/uma.h>
|
|
|
|
#include <net/if.h>
|
|
#include <net/vnet.h>
|
|
#include <net/pfvar.h>
|
|
|
|
#define ACCEPT_FLAGS(flags, oklist) \
|
|
do { \
|
|
if ((flags & ~(oklist)) & \
|
|
PFR_FLAG_ALLMASK) \
|
|
return (EINVAL); \
|
|
} while (0)
|
|
|
|
#define FILLIN_SIN(sin, addr) \
|
|
do { \
|
|
(sin).sin_len = sizeof(sin); \
|
|
(sin).sin_family = AF_INET; \
|
|
(sin).sin_addr = (addr); \
|
|
} while (0)
|
|
|
|
#define FILLIN_SIN6(sin6, addr) \
|
|
do { \
|
|
(sin6).sin6_len = sizeof(sin6); \
|
|
(sin6).sin6_family = AF_INET6; \
|
|
(sin6).sin6_addr = (addr); \
|
|
} while (0)
|
|
|
|
#define SWAP(type, a1, a2) \
|
|
do { \
|
|
type tmp = a1; \
|
|
a1 = a2; \
|
|
a2 = tmp; \
|
|
} while (0)
|
|
|
|
#define SUNION2PF(su, af) (((af)==AF_INET) ? \
|
|
(struct pf_addr *)&(su)->sin.sin_addr : \
|
|
(struct pf_addr *)&(su)->sin6.sin6_addr)
|
|
|
|
#define AF_BITS(af) (((af)==AF_INET)?32:128)
|
|
#define ADDR_NETWORK(ad) ((ad)->pfra_net < AF_BITS((ad)->pfra_af))
|
|
#define KENTRY_NETWORK(ke) ((ke)->pfrke_net < AF_BITS((ke)->pfrke_af))
|
|
#define KENTRY_RNF_ROOT(ke) \
|
|
((((struct radix_node *)(ke))->rn_flags & RNF_ROOT) != 0)
|
|
|
|
#define NO_ADDRESSES (-1)
|
|
#define ENQUEUE_UNMARKED_ONLY (1)
|
|
#define INVERT_NEG_FLAG (1)
|
|
|
|
struct pfr_walktree {
|
|
enum pfrw_op {
|
|
PFRW_MARK,
|
|
PFRW_SWEEP,
|
|
PFRW_ENQUEUE,
|
|
PFRW_GET_ADDRS,
|
|
PFRW_GET_ASTATS,
|
|
PFRW_POOL_GET,
|
|
PFRW_DYNADDR_UPDATE
|
|
} pfrw_op;
|
|
union {
|
|
struct pfr_addr *pfrw1_addr;
|
|
struct pfr_astats *pfrw1_astats;
|
|
struct pfr_kentryworkq *pfrw1_workq;
|
|
struct pfr_kentry *pfrw1_kentry;
|
|
struct pfi_dynaddr *pfrw1_dyn;
|
|
} pfrw_1;
|
|
int pfrw_free;
|
|
};
|
|
#define pfrw_addr pfrw_1.pfrw1_addr
|
|
#define pfrw_astats pfrw_1.pfrw1_astats
|
|
#define pfrw_workq pfrw_1.pfrw1_workq
|
|
#define pfrw_kentry pfrw_1.pfrw1_kentry
|
|
#define pfrw_dyn pfrw_1.pfrw1_dyn
|
|
#define pfrw_cnt pfrw_free
|
|
|
|
#define senderr(e) do { rv = (e); goto _bad; } while (0)
|
|
|
|
static MALLOC_DEFINE(M_PFTABLE, "pf_table", "pf(4) tables structures");
|
|
static VNET_DEFINE(uma_zone_t, pfr_kentry_z);
|
|
#define V_pfr_kentry_z VNET(pfr_kentry_z)
|
|
static VNET_DEFINE(uma_zone_t, pfr_kcounters_z);
|
|
#define V_pfr_kcounters_z VNET(pfr_kcounters_z)
|
|
|
|
static struct pf_addr pfr_ffaddr = {
|
|
.addr32 = { 0xffffffff, 0xffffffff, 0xffffffff, 0xffffffff }
|
|
};
|
|
|
|
static void pfr_copyout_addr(struct pfr_addr *,
|
|
struct pfr_kentry *ke);
|
|
static int pfr_validate_addr(struct pfr_addr *);
|
|
static void pfr_enqueue_addrs(struct pfr_ktable *,
|
|
struct pfr_kentryworkq *, int *, int);
|
|
static void pfr_mark_addrs(struct pfr_ktable *);
|
|
static struct pfr_kentry
|
|
*pfr_lookup_addr(struct pfr_ktable *,
|
|
struct pfr_addr *, int);
|
|
static struct pfr_kentry *pfr_create_kentry(struct pfr_addr *);
|
|
static void pfr_destroy_kentries(struct pfr_kentryworkq *);
|
|
static void pfr_destroy_kentry(struct pfr_kentry *);
|
|
static void pfr_insert_kentries(struct pfr_ktable *,
|
|
struct pfr_kentryworkq *, long);
|
|
static void pfr_remove_kentries(struct pfr_ktable *,
|
|
struct pfr_kentryworkq *);
|
|
static void pfr_clstats_kentries(struct pfr_kentryworkq *, long,
|
|
int);
|
|
static void pfr_reset_feedback(struct pfr_addr *, int);
|
|
static void pfr_prepare_network(union sockaddr_union *, int, int);
|
|
static int pfr_route_kentry(struct pfr_ktable *,
|
|
struct pfr_kentry *);
|
|
static int pfr_unroute_kentry(struct pfr_ktable *,
|
|
struct pfr_kentry *);
|
|
static int pfr_walktree(struct radix_node *, void *);
|
|
static int pfr_validate_table(struct pfr_table *, int, int);
|
|
static int pfr_fix_anchor(char *);
|
|
static void pfr_commit_ktable(struct pfr_ktable *, long);
|
|
static void pfr_insert_ktables(struct pfr_ktableworkq *);
|
|
static void pfr_insert_ktable(struct pfr_ktable *);
|
|
static void pfr_setflags_ktables(struct pfr_ktableworkq *);
|
|
static void pfr_setflags_ktable(struct pfr_ktable *, int);
|
|
static void pfr_clstats_ktables(struct pfr_ktableworkq *, long,
|
|
int);
|
|
static void pfr_clstats_ktable(struct pfr_ktable *, long, int);
|
|
static struct pfr_ktable
|
|
*pfr_create_ktable(struct pfr_table *, long, int);
|
|
static void pfr_destroy_ktables(struct pfr_ktableworkq *, int);
|
|
static void pfr_destroy_ktable(struct pfr_ktable *, int);
|
|
static int pfr_ktable_compare(struct pfr_ktable *,
|
|
struct pfr_ktable *);
|
|
static struct pfr_ktable
|
|
*pfr_lookup_table(struct pfr_table *);
|
|
static void pfr_clean_node_mask(struct pfr_ktable *,
|
|
struct pfr_kentryworkq *);
|
|
static int pfr_table_count(struct pfr_table *, int);
|
|
static int pfr_skip_table(struct pfr_table *,
|
|
struct pfr_ktable *, int);
|
|
static struct pfr_kentry
|
|
*pfr_kentry_byidx(struct pfr_ktable *, int, int);
|
|
|
|
static RB_PROTOTYPE(pfr_ktablehead, pfr_ktable, pfrkt_tree, pfr_ktable_compare);
|
|
static RB_GENERATE(pfr_ktablehead, pfr_ktable, pfrkt_tree, pfr_ktable_compare);
|
|
|
|
struct pfr_ktablehead pfr_ktables;
|
|
struct pfr_table pfr_nulltable;
|
|
int pfr_ktable_cnt;
|
|
|
|
void
|
|
pfr_initialize(void)
|
|
{
|
|
|
|
V_pfr_kentry_z = uma_zcreate("pf table entries",
|
|
sizeof(struct pfr_kentry), NULL, NULL, NULL, NULL, UMA_ALIGN_PTR,
|
|
0);
|
|
V_pfr_kcounters_z = uma_zcreate("pf table counters",
|
|
sizeof(struct pfr_kcounters), NULL, NULL, NULL, NULL,
|
|
UMA_ALIGN_PTR, 0);
|
|
V_pf_limits[PF_LIMIT_TABLE_ENTRIES].zone = V_pfr_kentry_z;
|
|
V_pf_limits[PF_LIMIT_TABLE_ENTRIES].limit = PFR_KENTRY_HIWAT;
|
|
}
|
|
|
|
void
|
|
pfr_cleanup(void)
|
|
{
|
|
|
|
uma_zdestroy(V_pfr_kentry_z);
|
|
uma_zdestroy(V_pfr_kcounters_z);
|
|
}
|
|
|
|
int
|
|
pfr_clr_addrs(struct pfr_table *tbl, int *ndel, int flags)
|
|
{
|
|
struct pfr_ktable *kt;
|
|
struct pfr_kentryworkq workq;
|
|
|
|
PF_RULES_WASSERT();
|
|
|
|
ACCEPT_FLAGS(flags, PFR_FLAG_DUMMY);
|
|
if (pfr_validate_table(tbl, 0, flags & PFR_FLAG_USERIOCTL))
|
|
return (EINVAL);
|
|
kt = pfr_lookup_table(tbl);
|
|
if (kt == NULL || !(kt->pfrkt_flags & PFR_TFLAG_ACTIVE))
|
|
return (ESRCH);
|
|
if (kt->pfrkt_flags & PFR_TFLAG_CONST)
|
|
return (EPERM);
|
|
pfr_enqueue_addrs(kt, &workq, ndel, 0);
|
|
|
|
if (!(flags & PFR_FLAG_DUMMY)) {
|
|
pfr_remove_kentries(kt, &workq);
|
|
KASSERT(kt->pfrkt_cnt == 0, ("%s: non-null pfrkt_cnt", __func__));
|
|
}
|
|
return (0);
|
|
}
|
|
|
|
int
|
|
pfr_add_addrs(struct pfr_table *tbl, struct pfr_addr *addr, int size,
|
|
int *nadd, int flags)
|
|
{
|
|
struct pfr_ktable *kt, *tmpkt;
|
|
struct pfr_kentryworkq workq;
|
|
struct pfr_kentry *p, *q;
|
|
struct pfr_addr *ad;
|
|
int i, rv, xadd = 0;
|
|
long tzero = time_second;
|
|
|
|
PF_RULES_WASSERT();
|
|
|
|
ACCEPT_FLAGS(flags, PFR_FLAG_DUMMY | PFR_FLAG_FEEDBACK);
|
|
if (pfr_validate_table(tbl, 0, flags & PFR_FLAG_USERIOCTL))
|
|
return (EINVAL);
|
|
kt = pfr_lookup_table(tbl);
|
|
if (kt == NULL || !(kt->pfrkt_flags & PFR_TFLAG_ACTIVE))
|
|
return (ESRCH);
|
|
if (kt->pfrkt_flags & PFR_TFLAG_CONST)
|
|
return (EPERM);
|
|
tmpkt = pfr_create_ktable(&pfr_nulltable, 0, 0);
|
|
if (tmpkt == NULL)
|
|
return (ENOMEM);
|
|
SLIST_INIT(&workq);
|
|
for (i = 0, ad = addr; i < size; i++, ad++) {
|
|
if (pfr_validate_addr(ad))
|
|
senderr(EINVAL);
|
|
p = pfr_lookup_addr(kt, ad, 1);
|
|
q = pfr_lookup_addr(tmpkt, ad, 1);
|
|
if (flags & PFR_FLAG_FEEDBACK) {
|
|
if (q != NULL)
|
|
ad->pfra_fback = PFR_FB_DUPLICATE;
|
|
else if (p == NULL)
|
|
ad->pfra_fback = PFR_FB_ADDED;
|
|
else if (p->pfrke_not != ad->pfra_not)
|
|
ad->pfra_fback = PFR_FB_CONFLICT;
|
|
else
|
|
ad->pfra_fback = PFR_FB_NONE;
|
|
}
|
|
if (p == NULL && q == NULL) {
|
|
p = pfr_create_kentry(ad);
|
|
if (p == NULL)
|
|
senderr(ENOMEM);
|
|
if (pfr_route_kentry(tmpkt, p)) {
|
|
pfr_destroy_kentry(p);
|
|
ad->pfra_fback = PFR_FB_NONE;
|
|
} else {
|
|
SLIST_INSERT_HEAD(&workq, p, pfrke_workq);
|
|
xadd++;
|
|
}
|
|
}
|
|
}
|
|
pfr_clean_node_mask(tmpkt, &workq);
|
|
if (!(flags & PFR_FLAG_DUMMY))
|
|
pfr_insert_kentries(kt, &workq, tzero);
|
|
else
|
|
pfr_destroy_kentries(&workq);
|
|
if (nadd != NULL)
|
|
*nadd = xadd;
|
|
pfr_destroy_ktable(tmpkt, 0);
|
|
return (0);
|
|
_bad:
|
|
pfr_clean_node_mask(tmpkt, &workq);
|
|
pfr_destroy_kentries(&workq);
|
|
if (flags & PFR_FLAG_FEEDBACK)
|
|
pfr_reset_feedback(addr, size);
|
|
pfr_destroy_ktable(tmpkt, 0);
|
|
return (rv);
|
|
}
|
|
|
|
int
|
|
pfr_del_addrs(struct pfr_table *tbl, struct pfr_addr *addr, int size,
|
|
int *ndel, int flags)
|
|
{
|
|
struct pfr_ktable *kt;
|
|
struct pfr_kentryworkq workq;
|
|
struct pfr_kentry *p;
|
|
struct pfr_addr *ad;
|
|
int i, rv, xdel = 0, log = 1;
|
|
|
|
PF_RULES_WASSERT();
|
|
|
|
ACCEPT_FLAGS(flags, PFR_FLAG_DUMMY | PFR_FLAG_FEEDBACK);
|
|
if (pfr_validate_table(tbl, 0, flags & PFR_FLAG_USERIOCTL))
|
|
return (EINVAL);
|
|
kt = pfr_lookup_table(tbl);
|
|
if (kt == NULL || !(kt->pfrkt_flags & PFR_TFLAG_ACTIVE))
|
|
return (ESRCH);
|
|
if (kt->pfrkt_flags & PFR_TFLAG_CONST)
|
|
return (EPERM);
|
|
/*
|
|
* there are two algorithms to choose from here.
|
|
* with:
|
|
* n: number of addresses to delete
|
|
* N: number of addresses in the table
|
|
*
|
|
* one is O(N) and is better for large 'n'
|
|
* one is O(n*LOG(N)) and is better for small 'n'
|
|
*
|
|
* following code try to decide which one is best.
|
|
*/
|
|
for (i = kt->pfrkt_cnt; i > 0; i >>= 1)
|
|
log++;
|
|
if (size > kt->pfrkt_cnt/log) {
|
|
/* full table scan */
|
|
pfr_mark_addrs(kt);
|
|
} else {
|
|
/* iterate over addresses to delete */
|
|
for (i = 0, ad = addr; i < size; i++, ad++) {
|
|
if (pfr_validate_addr(ad))
|
|
return (EINVAL);
|
|
p = pfr_lookup_addr(kt, ad, 1);
|
|
if (p != NULL)
|
|
p->pfrke_mark = 0;
|
|
}
|
|
}
|
|
SLIST_INIT(&workq);
|
|
for (i = 0, ad = addr; i < size; i++, ad++) {
|
|
if (pfr_validate_addr(ad))
|
|
senderr(EINVAL);
|
|
p = pfr_lookup_addr(kt, ad, 1);
|
|
if (flags & PFR_FLAG_FEEDBACK) {
|
|
if (p == NULL)
|
|
ad->pfra_fback = PFR_FB_NONE;
|
|
else if (p->pfrke_not != ad->pfra_not)
|
|
ad->pfra_fback = PFR_FB_CONFLICT;
|
|
else if (p->pfrke_mark)
|
|
ad->pfra_fback = PFR_FB_DUPLICATE;
|
|
else
|
|
ad->pfra_fback = PFR_FB_DELETED;
|
|
}
|
|
if (p != NULL && p->pfrke_not == ad->pfra_not &&
|
|
!p->pfrke_mark) {
|
|
p->pfrke_mark = 1;
|
|
SLIST_INSERT_HEAD(&workq, p, pfrke_workq);
|
|
xdel++;
|
|
}
|
|
}
|
|
if (!(flags & PFR_FLAG_DUMMY))
|
|
pfr_remove_kentries(kt, &workq);
|
|
if (ndel != NULL)
|
|
*ndel = xdel;
|
|
return (0);
|
|
_bad:
|
|
if (flags & PFR_FLAG_FEEDBACK)
|
|
pfr_reset_feedback(addr, size);
|
|
return (rv);
|
|
}
|
|
|
|
int
|
|
pfr_set_addrs(struct pfr_table *tbl, struct pfr_addr *addr, int size,
|
|
int *size2, int *nadd, int *ndel, int *nchange, int flags,
|
|
u_int32_t ignore_pfrt_flags)
|
|
{
|
|
struct pfr_ktable *kt, *tmpkt;
|
|
struct pfr_kentryworkq addq, delq, changeq;
|
|
struct pfr_kentry *p, *q;
|
|
struct pfr_addr ad;
|
|
int i, rv, xadd = 0, xdel = 0, xchange = 0;
|
|
long tzero = time_second;
|
|
|
|
PF_RULES_WASSERT();
|
|
|
|
ACCEPT_FLAGS(flags, PFR_FLAG_DUMMY | PFR_FLAG_FEEDBACK);
|
|
if (pfr_validate_table(tbl, ignore_pfrt_flags, flags &
|
|
PFR_FLAG_USERIOCTL))
|
|
return (EINVAL);
|
|
kt = pfr_lookup_table(tbl);
|
|
if (kt == NULL || !(kt->pfrkt_flags & PFR_TFLAG_ACTIVE))
|
|
return (ESRCH);
|
|
if (kt->pfrkt_flags & PFR_TFLAG_CONST)
|
|
return (EPERM);
|
|
tmpkt = pfr_create_ktable(&pfr_nulltable, 0, 0);
|
|
if (tmpkt == NULL)
|
|
return (ENOMEM);
|
|
pfr_mark_addrs(kt);
|
|
SLIST_INIT(&addq);
|
|
SLIST_INIT(&delq);
|
|
SLIST_INIT(&changeq);
|
|
for (i = 0; i < size; i++) {
|
|
/*
|
|
* XXXGL: undertand pf_if usage of this function
|
|
* and make ad a moving pointer
|
|
*/
|
|
bcopy(addr + i, &ad, sizeof(ad));
|
|
if (pfr_validate_addr(&ad))
|
|
senderr(EINVAL);
|
|
ad.pfra_fback = PFR_FB_NONE;
|
|
p = pfr_lookup_addr(kt, &ad, 1);
|
|
if (p != NULL) {
|
|
if (p->pfrke_mark) {
|
|
ad.pfra_fback = PFR_FB_DUPLICATE;
|
|
goto _skip;
|
|
}
|
|
p->pfrke_mark = 1;
|
|
if (p->pfrke_not != ad.pfra_not) {
|
|
SLIST_INSERT_HEAD(&changeq, p, pfrke_workq);
|
|
ad.pfra_fback = PFR_FB_CHANGED;
|
|
xchange++;
|
|
}
|
|
} else {
|
|
q = pfr_lookup_addr(tmpkt, &ad, 1);
|
|
if (q != NULL) {
|
|
ad.pfra_fback = PFR_FB_DUPLICATE;
|
|
goto _skip;
|
|
}
|
|
p = pfr_create_kentry(&ad);
|
|
if (p == NULL)
|
|
senderr(ENOMEM);
|
|
if (pfr_route_kentry(tmpkt, p)) {
|
|
pfr_destroy_kentry(p);
|
|
ad.pfra_fback = PFR_FB_NONE;
|
|
} else {
|
|
SLIST_INSERT_HEAD(&addq, p, pfrke_workq);
|
|
ad.pfra_fback = PFR_FB_ADDED;
|
|
xadd++;
|
|
}
|
|
}
|
|
_skip:
|
|
if (flags & PFR_FLAG_FEEDBACK)
|
|
bcopy(&ad, addr + i, sizeof(ad));
|
|
}
|
|
pfr_enqueue_addrs(kt, &delq, &xdel, ENQUEUE_UNMARKED_ONLY);
|
|
if ((flags & PFR_FLAG_FEEDBACK) && *size2) {
|
|
if (*size2 < size+xdel) {
|
|
*size2 = size+xdel;
|
|
senderr(0);
|
|
}
|
|
i = 0;
|
|
SLIST_FOREACH(p, &delq, pfrke_workq) {
|
|
pfr_copyout_addr(&ad, p);
|
|
ad.pfra_fback = PFR_FB_DELETED;
|
|
bcopy(&ad, addr + size + i, sizeof(ad));
|
|
i++;
|
|
}
|
|
}
|
|
pfr_clean_node_mask(tmpkt, &addq);
|
|
if (!(flags & PFR_FLAG_DUMMY)) {
|
|
pfr_insert_kentries(kt, &addq, tzero);
|
|
pfr_remove_kentries(kt, &delq);
|
|
pfr_clstats_kentries(&changeq, tzero, INVERT_NEG_FLAG);
|
|
} else
|
|
pfr_destroy_kentries(&addq);
|
|
if (nadd != NULL)
|
|
*nadd = xadd;
|
|
if (ndel != NULL)
|
|
*ndel = xdel;
|
|
if (nchange != NULL)
|
|
*nchange = xchange;
|
|
if ((flags & PFR_FLAG_FEEDBACK) && size2)
|
|
*size2 = size+xdel;
|
|
pfr_destroy_ktable(tmpkt, 0);
|
|
return (0);
|
|
_bad:
|
|
pfr_clean_node_mask(tmpkt, &addq);
|
|
pfr_destroy_kentries(&addq);
|
|
if (flags & PFR_FLAG_FEEDBACK)
|
|
pfr_reset_feedback(addr, size);
|
|
pfr_destroy_ktable(tmpkt, 0);
|
|
return (rv);
|
|
}
|
|
|
|
int
|
|
pfr_tst_addrs(struct pfr_table *tbl, struct pfr_addr *addr, int size,
|
|
int *nmatch, int flags)
|
|
{
|
|
struct pfr_ktable *kt;
|
|
struct pfr_kentry *p;
|
|
struct pfr_addr *ad;
|
|
int i, xmatch = 0;
|
|
|
|
PF_RULES_RASSERT();
|
|
|
|
ACCEPT_FLAGS(flags, PFR_FLAG_REPLACE);
|
|
if (pfr_validate_table(tbl, 0, 0))
|
|
return (EINVAL);
|
|
kt = pfr_lookup_table(tbl);
|
|
if (kt == NULL || !(kt->pfrkt_flags & PFR_TFLAG_ACTIVE))
|
|
return (ESRCH);
|
|
|
|
for (i = 0, ad = addr; i < size; i++, ad++) {
|
|
if (pfr_validate_addr(ad))
|
|
return (EINVAL);
|
|
if (ADDR_NETWORK(ad))
|
|
return (EINVAL);
|
|
p = pfr_lookup_addr(kt, ad, 0);
|
|
if (flags & PFR_FLAG_REPLACE)
|
|
pfr_copyout_addr(ad, p);
|
|
ad->pfra_fback = (p == NULL) ? PFR_FB_NONE :
|
|
(p->pfrke_not ? PFR_FB_NOTMATCH : PFR_FB_MATCH);
|
|
if (p != NULL && !p->pfrke_not)
|
|
xmatch++;
|
|
}
|
|
if (nmatch != NULL)
|
|
*nmatch = xmatch;
|
|
return (0);
|
|
}
|
|
|
|
int
|
|
pfr_get_addrs(struct pfr_table *tbl, struct pfr_addr *addr, int *size,
|
|
int flags)
|
|
{
|
|
struct pfr_ktable *kt;
|
|
struct pfr_walktree w;
|
|
int rv;
|
|
|
|
PF_RULES_RASSERT();
|
|
|
|
ACCEPT_FLAGS(flags, 0);
|
|
if (pfr_validate_table(tbl, 0, 0))
|
|
return (EINVAL);
|
|
kt = pfr_lookup_table(tbl);
|
|
if (kt == NULL || !(kt->pfrkt_flags & PFR_TFLAG_ACTIVE))
|
|
return (ESRCH);
|
|
if (kt->pfrkt_cnt > *size) {
|
|
*size = kt->pfrkt_cnt;
|
|
return (0);
|
|
}
|
|
|
|
bzero(&w, sizeof(w));
|
|
w.pfrw_op = PFRW_GET_ADDRS;
|
|
w.pfrw_addr = addr;
|
|
w.pfrw_free = kt->pfrkt_cnt;
|
|
rv = kt->pfrkt_ip4->rnh_walktree(kt->pfrkt_ip4, pfr_walktree, &w);
|
|
if (!rv)
|
|
rv = kt->pfrkt_ip6->rnh_walktree(kt->pfrkt_ip6, pfr_walktree,
|
|
&w);
|
|
if (rv)
|
|
return (rv);
|
|
|
|
KASSERT(w.pfrw_free == 0, ("%s: corruption detected (%d)", __func__,
|
|
w.pfrw_free));
|
|
|
|
*size = kt->pfrkt_cnt;
|
|
return (0);
|
|
}
|
|
|
|
int
|
|
pfr_get_astats(struct pfr_table *tbl, struct pfr_astats *addr, int *size,
|
|
int flags)
|
|
{
|
|
struct pfr_ktable *kt;
|
|
struct pfr_walktree w;
|
|
struct pfr_kentryworkq workq;
|
|
int rv;
|
|
long tzero = time_second;
|
|
|
|
PF_RULES_RASSERT();
|
|
|
|
/* XXX PFR_FLAG_CLSTATS disabled */
|
|
ACCEPT_FLAGS(flags, 0);
|
|
if (pfr_validate_table(tbl, 0, 0))
|
|
return (EINVAL);
|
|
kt = pfr_lookup_table(tbl);
|
|
if (kt == NULL || !(kt->pfrkt_flags & PFR_TFLAG_ACTIVE))
|
|
return (ESRCH);
|
|
if (kt->pfrkt_cnt > *size) {
|
|
*size = kt->pfrkt_cnt;
|
|
return (0);
|
|
}
|
|
|
|
bzero(&w, sizeof(w));
|
|
w.pfrw_op = PFRW_GET_ASTATS;
|
|
w.pfrw_astats = addr;
|
|
w.pfrw_free = kt->pfrkt_cnt;
|
|
rv = kt->pfrkt_ip4->rnh_walktree(kt->pfrkt_ip4, pfr_walktree, &w);
|
|
if (!rv)
|
|
rv = kt->pfrkt_ip6->rnh_walktree(kt->pfrkt_ip6, pfr_walktree,
|
|
&w);
|
|
if (!rv && (flags & PFR_FLAG_CLSTATS)) {
|
|
pfr_enqueue_addrs(kt, &workq, NULL, 0);
|
|
pfr_clstats_kentries(&workq, tzero, 0);
|
|
}
|
|
if (rv)
|
|
return (rv);
|
|
|
|
if (w.pfrw_free) {
|
|
printf("pfr_get_astats: corruption detected (%d).\n",
|
|
w.pfrw_free);
|
|
return (ENOTTY);
|
|
}
|
|
*size = kt->pfrkt_cnt;
|
|
return (0);
|
|
}
|
|
|
|
int
|
|
pfr_clr_astats(struct pfr_table *tbl, struct pfr_addr *addr, int size,
|
|
int *nzero, int flags)
|
|
{
|
|
struct pfr_ktable *kt;
|
|
struct pfr_kentryworkq workq;
|
|
struct pfr_kentry *p;
|
|
struct pfr_addr *ad;
|
|
int i, rv, xzero = 0;
|
|
|
|
PF_RULES_WASSERT();
|
|
|
|
ACCEPT_FLAGS(flags, PFR_FLAG_DUMMY | PFR_FLAG_FEEDBACK);
|
|
if (pfr_validate_table(tbl, 0, 0))
|
|
return (EINVAL);
|
|
kt = pfr_lookup_table(tbl);
|
|
if (kt == NULL || !(kt->pfrkt_flags & PFR_TFLAG_ACTIVE))
|
|
return (ESRCH);
|
|
SLIST_INIT(&workq);
|
|
for (i = 0, ad = addr; i < size; i++, ad++) {
|
|
if (pfr_validate_addr(ad))
|
|
senderr(EINVAL);
|
|
p = pfr_lookup_addr(kt, ad, 1);
|
|
if (flags & PFR_FLAG_FEEDBACK) {
|
|
ad->pfra_fback = (p != NULL) ?
|
|
PFR_FB_CLEARED : PFR_FB_NONE;
|
|
}
|
|
if (p != NULL) {
|
|
SLIST_INSERT_HEAD(&workq, p, pfrke_workq);
|
|
xzero++;
|
|
}
|
|
}
|
|
|
|
if (!(flags & PFR_FLAG_DUMMY))
|
|
pfr_clstats_kentries(&workq, 0, 0);
|
|
if (nzero != NULL)
|
|
*nzero = xzero;
|
|
return (0);
|
|
_bad:
|
|
if (flags & PFR_FLAG_FEEDBACK)
|
|
pfr_reset_feedback(addr, size);
|
|
return (rv);
|
|
}
|
|
|
|
static int
|
|
pfr_validate_addr(struct pfr_addr *ad)
|
|
{
|
|
int i;
|
|
|
|
switch (ad->pfra_af) {
|
|
#ifdef INET
|
|
case AF_INET:
|
|
if (ad->pfra_net > 32)
|
|
return (-1);
|
|
break;
|
|
#endif /* INET */
|
|
#ifdef INET6
|
|
case AF_INET6:
|
|
if (ad->pfra_net > 128)
|
|
return (-1);
|
|
break;
|
|
#endif /* INET6 */
|
|
default:
|
|
return (-1);
|
|
}
|
|
if (ad->pfra_net < 128 &&
|
|
(((caddr_t)ad)[ad->pfra_net/8] & (0xFF >> (ad->pfra_net%8))))
|
|
return (-1);
|
|
for (i = (ad->pfra_net+7)/8; i < sizeof(ad->pfra_u); i++)
|
|
if (((caddr_t)ad)[i])
|
|
return (-1);
|
|
if (ad->pfra_not && ad->pfra_not != 1)
|
|
return (-1);
|
|
if (ad->pfra_fback)
|
|
return (-1);
|
|
return (0);
|
|
}
|
|
|
|
static void
|
|
pfr_enqueue_addrs(struct pfr_ktable *kt, struct pfr_kentryworkq *workq,
|
|
int *naddr, int sweep)
|
|
{
|
|
struct pfr_walktree w;
|
|
|
|
SLIST_INIT(workq);
|
|
bzero(&w, sizeof(w));
|
|
w.pfrw_op = sweep ? PFRW_SWEEP : PFRW_ENQUEUE;
|
|
w.pfrw_workq = workq;
|
|
if (kt->pfrkt_ip4 != NULL)
|
|
if (kt->pfrkt_ip4->rnh_walktree(kt->pfrkt_ip4, pfr_walktree,
|
|
&w))
|
|
printf("pfr_enqueue_addrs: IPv4 walktree failed.\n");
|
|
if (kt->pfrkt_ip6 != NULL)
|
|
if (kt->pfrkt_ip6->rnh_walktree(kt->pfrkt_ip6, pfr_walktree,
|
|
&w))
|
|
printf("pfr_enqueue_addrs: IPv6 walktree failed.\n");
|
|
if (naddr != NULL)
|
|
*naddr = w.pfrw_cnt;
|
|
}
|
|
|
|
static void
|
|
pfr_mark_addrs(struct pfr_ktable *kt)
|
|
{
|
|
struct pfr_walktree w;
|
|
|
|
bzero(&w, sizeof(w));
|
|
w.pfrw_op = PFRW_MARK;
|
|
if (kt->pfrkt_ip4->rnh_walktree(kt->pfrkt_ip4, pfr_walktree, &w))
|
|
printf("pfr_mark_addrs: IPv4 walktree failed.\n");
|
|
if (kt->pfrkt_ip6->rnh_walktree(kt->pfrkt_ip6, pfr_walktree, &w))
|
|
printf("pfr_mark_addrs: IPv6 walktree failed.\n");
|
|
}
|
|
|
|
|
|
static struct pfr_kentry *
|
|
pfr_lookup_addr(struct pfr_ktable *kt, struct pfr_addr *ad, int exact)
|
|
{
|
|
union sockaddr_union sa, mask;
|
|
struct radix_node_head *head = NULL;
|
|
struct pfr_kentry *ke;
|
|
|
|
PF_RULES_ASSERT();
|
|
|
|
bzero(&sa, sizeof(sa));
|
|
if (ad->pfra_af == AF_INET) {
|
|
FILLIN_SIN(sa.sin, ad->pfra_ip4addr);
|
|
head = kt->pfrkt_ip4;
|
|
} else if ( ad->pfra_af == AF_INET6 ) {
|
|
FILLIN_SIN6(sa.sin6, ad->pfra_ip6addr);
|
|
head = kt->pfrkt_ip6;
|
|
}
|
|
if (ADDR_NETWORK(ad)) {
|
|
pfr_prepare_network(&mask, ad->pfra_af, ad->pfra_net);
|
|
ke = (struct pfr_kentry *)rn_lookup(&sa, &mask, head);
|
|
if (ke && KENTRY_RNF_ROOT(ke))
|
|
ke = NULL;
|
|
} else {
|
|
ke = (struct pfr_kentry *)rn_match(&sa, head);
|
|
if (ke && KENTRY_RNF_ROOT(ke))
|
|
ke = NULL;
|
|
if (exact && ke && KENTRY_NETWORK(ke))
|
|
ke = NULL;
|
|
}
|
|
return (ke);
|
|
}
|
|
|
|
static struct pfr_kentry *
|
|
pfr_create_kentry(struct pfr_addr *ad)
|
|
{
|
|
struct pfr_kentry *ke;
|
|
|
|
ke = uma_zalloc(V_pfr_kentry_z, M_NOWAIT | M_ZERO);
|
|
if (ke == NULL)
|
|
return (NULL);
|
|
|
|
if (ad->pfra_af == AF_INET)
|
|
FILLIN_SIN(ke->pfrke_sa.sin, ad->pfra_ip4addr);
|
|
else if (ad->pfra_af == AF_INET6)
|
|
FILLIN_SIN6(ke->pfrke_sa.sin6, ad->pfra_ip6addr);
|
|
ke->pfrke_af = ad->pfra_af;
|
|
ke->pfrke_net = ad->pfra_net;
|
|
ke->pfrke_not = ad->pfra_not;
|
|
return (ke);
|
|
}
|
|
|
|
static void
|
|
pfr_destroy_kentries(struct pfr_kentryworkq *workq)
|
|
{
|
|
struct pfr_kentry *p, *q;
|
|
|
|
for (p = SLIST_FIRST(workq); p != NULL; p = q) {
|
|
q = SLIST_NEXT(p, pfrke_workq);
|
|
pfr_destroy_kentry(p);
|
|
}
|
|
}
|
|
|
|
static void
|
|
pfr_destroy_kentry(struct pfr_kentry *ke)
|
|
{
|
|
if (ke->pfrke_counters)
|
|
uma_zfree(V_pfr_kcounters_z, ke->pfrke_counters);
|
|
uma_zfree(V_pfr_kentry_z, ke);
|
|
}
|
|
|
|
static void
|
|
pfr_insert_kentries(struct pfr_ktable *kt,
|
|
struct pfr_kentryworkq *workq, long tzero)
|
|
{
|
|
struct pfr_kentry *p;
|
|
int rv, n = 0;
|
|
|
|
SLIST_FOREACH(p, workq, pfrke_workq) {
|
|
rv = pfr_route_kentry(kt, p);
|
|
if (rv) {
|
|
printf("pfr_insert_kentries: cannot route entry "
|
|
"(code=%d).\n", rv);
|
|
break;
|
|
}
|
|
p->pfrke_tzero = tzero;
|
|
n++;
|
|
}
|
|
kt->pfrkt_cnt += n;
|
|
}
|
|
|
|
int
|
|
pfr_insert_kentry(struct pfr_ktable *kt, struct pfr_addr *ad, long tzero)
|
|
{
|
|
struct pfr_kentry *p;
|
|
int rv;
|
|
|
|
p = pfr_lookup_addr(kt, ad, 1);
|
|
if (p != NULL)
|
|
return (0);
|
|
p = pfr_create_kentry(ad);
|
|
if (p == NULL)
|
|
return (ENOMEM);
|
|
|
|
rv = pfr_route_kentry(kt, p);
|
|
if (rv)
|
|
return (rv);
|
|
|
|
p->pfrke_tzero = tzero;
|
|
kt->pfrkt_cnt++;
|
|
|
|
return (0);
|
|
}
|
|
|
|
static void
|
|
pfr_remove_kentries(struct pfr_ktable *kt,
|
|
struct pfr_kentryworkq *workq)
|
|
{
|
|
struct pfr_kentry *p;
|
|
int n = 0;
|
|
|
|
SLIST_FOREACH(p, workq, pfrke_workq) {
|
|
pfr_unroute_kentry(kt, p);
|
|
n++;
|
|
}
|
|
kt->pfrkt_cnt -= n;
|
|
pfr_destroy_kentries(workq);
|
|
}
|
|
|
|
static void
|
|
pfr_clean_node_mask(struct pfr_ktable *kt,
|
|
struct pfr_kentryworkq *workq)
|
|
{
|
|
struct pfr_kentry *p;
|
|
|
|
SLIST_FOREACH(p, workq, pfrke_workq)
|
|
pfr_unroute_kentry(kt, p);
|
|
}
|
|
|
|
static void
|
|
pfr_clstats_kentries(struct pfr_kentryworkq *workq, long tzero, int negchange)
|
|
{
|
|
struct pfr_kentry *p;
|
|
|
|
SLIST_FOREACH(p, workq, pfrke_workq) {
|
|
if (negchange)
|
|
p->pfrke_not = !p->pfrke_not;
|
|
if (p->pfrke_counters) {
|
|
uma_zfree(V_pfr_kcounters_z, p->pfrke_counters);
|
|
p->pfrke_counters = NULL;
|
|
}
|
|
p->pfrke_tzero = tzero;
|
|
}
|
|
}
|
|
|
|
static void
|
|
pfr_reset_feedback(struct pfr_addr *addr, int size)
|
|
{
|
|
struct pfr_addr *ad;
|
|
int i;
|
|
|
|
for (i = 0, ad = addr; i < size; i++, ad++)
|
|
ad->pfra_fback = PFR_FB_NONE;
|
|
}
|
|
|
|
static void
|
|
pfr_prepare_network(union sockaddr_union *sa, int af, int net)
|
|
{
|
|
int i;
|
|
|
|
bzero(sa, sizeof(*sa));
|
|
if (af == AF_INET) {
|
|
sa->sin.sin_len = sizeof(sa->sin);
|
|
sa->sin.sin_family = AF_INET;
|
|
sa->sin.sin_addr.s_addr = net ? htonl(-1 << (32-net)) : 0;
|
|
} else if (af == AF_INET6) {
|
|
sa->sin6.sin6_len = sizeof(sa->sin6);
|
|
sa->sin6.sin6_family = AF_INET6;
|
|
for (i = 0; i < 4; i++) {
|
|
if (net <= 32) {
|
|
sa->sin6.sin6_addr.s6_addr32[i] =
|
|
net ? htonl(-1 << (32-net)) : 0;
|
|
break;
|
|
}
|
|
sa->sin6.sin6_addr.s6_addr32[i] = 0xFFFFFFFF;
|
|
net -= 32;
|
|
}
|
|
}
|
|
}
|
|
|
|
static int
|
|
pfr_route_kentry(struct pfr_ktable *kt, struct pfr_kentry *ke)
|
|
{
|
|
union sockaddr_union mask;
|
|
struct radix_node *rn;
|
|
struct radix_node_head *head = NULL;
|
|
|
|
PF_RULES_WASSERT();
|
|
|
|
bzero(ke->pfrke_node, sizeof(ke->pfrke_node));
|
|
if (ke->pfrke_af == AF_INET)
|
|
head = kt->pfrkt_ip4;
|
|
else if (ke->pfrke_af == AF_INET6)
|
|
head = kt->pfrkt_ip6;
|
|
|
|
if (KENTRY_NETWORK(ke)) {
|
|
pfr_prepare_network(&mask, ke->pfrke_af, ke->pfrke_net);
|
|
rn = rn_addroute(&ke->pfrke_sa, &mask, head, ke->pfrke_node);
|
|
} else
|
|
rn = rn_addroute(&ke->pfrke_sa, NULL, head, ke->pfrke_node);
|
|
|
|
return (rn == NULL ? -1 : 0);
|
|
}
|
|
|
|
static int
|
|
pfr_unroute_kentry(struct pfr_ktable *kt, struct pfr_kentry *ke)
|
|
{
|
|
union sockaddr_union mask;
|
|
struct radix_node *rn;
|
|
struct radix_node_head *head = NULL;
|
|
|
|
if (ke->pfrke_af == AF_INET)
|
|
head = kt->pfrkt_ip4;
|
|
else if (ke->pfrke_af == AF_INET6)
|
|
head = kt->pfrkt_ip6;
|
|
|
|
if (KENTRY_NETWORK(ke)) {
|
|
pfr_prepare_network(&mask, ke->pfrke_af, ke->pfrke_net);
|
|
rn = rn_delete(&ke->pfrke_sa, &mask, head);
|
|
} else
|
|
rn = rn_delete(&ke->pfrke_sa, NULL, head);
|
|
|
|
if (rn == NULL) {
|
|
printf("pfr_unroute_kentry: delete failed.\n");
|
|
return (-1);
|
|
}
|
|
return (0);
|
|
}
|
|
|
|
static void
|
|
pfr_copyout_addr(struct pfr_addr *ad, struct pfr_kentry *ke)
|
|
{
|
|
bzero(ad, sizeof(*ad));
|
|
if (ke == NULL)
|
|
return;
|
|
ad->pfra_af = ke->pfrke_af;
|
|
ad->pfra_net = ke->pfrke_net;
|
|
ad->pfra_not = ke->pfrke_not;
|
|
if (ad->pfra_af == AF_INET)
|
|
ad->pfra_ip4addr = ke->pfrke_sa.sin.sin_addr;
|
|
else if (ad->pfra_af == AF_INET6)
|
|
ad->pfra_ip6addr = ke->pfrke_sa.sin6.sin6_addr;
|
|
}
|
|
|
|
static int
|
|
pfr_walktree(struct radix_node *rn, void *arg)
|
|
{
|
|
struct pfr_kentry *ke = (struct pfr_kentry *)rn;
|
|
struct pfr_walktree *w = arg;
|
|
|
|
switch (w->pfrw_op) {
|
|
case PFRW_MARK:
|
|
ke->pfrke_mark = 0;
|
|
break;
|
|
case PFRW_SWEEP:
|
|
if (ke->pfrke_mark)
|
|
break;
|
|
/* FALLTHROUGH */
|
|
case PFRW_ENQUEUE:
|
|
SLIST_INSERT_HEAD(w->pfrw_workq, ke, pfrke_workq);
|
|
w->pfrw_cnt++;
|
|
break;
|
|
case PFRW_GET_ADDRS:
|
|
if (w->pfrw_free-- > 0) {
|
|
pfr_copyout_addr(w->pfrw_addr, ke);
|
|
w->pfrw_addr++;
|
|
}
|
|
break;
|
|
case PFRW_GET_ASTATS:
|
|
if (w->pfrw_free-- > 0) {
|
|
struct pfr_astats as;
|
|
|
|
pfr_copyout_addr(&as.pfras_a, ke);
|
|
|
|
if (ke->pfrke_counters) {
|
|
bcopy(ke->pfrke_counters->pfrkc_packets,
|
|
as.pfras_packets, sizeof(as.pfras_packets));
|
|
bcopy(ke->pfrke_counters->pfrkc_bytes,
|
|
as.pfras_bytes, sizeof(as.pfras_bytes));
|
|
} else {
|
|
bzero(as.pfras_packets, sizeof(as.pfras_packets));
|
|
bzero(as.pfras_bytes, sizeof(as.pfras_bytes));
|
|
as.pfras_a.pfra_fback = PFR_FB_NOCOUNT;
|
|
}
|
|
as.pfras_tzero = ke->pfrke_tzero;
|
|
|
|
bcopy(&as, w->pfrw_astats, sizeof(as));
|
|
w->pfrw_astats++;
|
|
}
|
|
break;
|
|
case PFRW_POOL_GET:
|
|
if (ke->pfrke_not)
|
|
break; /* negative entries are ignored */
|
|
if (!w->pfrw_cnt--) {
|
|
w->pfrw_kentry = ke;
|
|
return (1); /* finish search */
|
|
}
|
|
break;
|
|
case PFRW_DYNADDR_UPDATE:
|
|
{
|
|
union sockaddr_union pfr_mask;
|
|
|
|
if (ke->pfrke_af == AF_INET) {
|
|
if (w->pfrw_dyn->pfid_acnt4++ > 0)
|
|
break;
|
|
pfr_prepare_network(&pfr_mask, AF_INET, ke->pfrke_net);
|
|
w->pfrw_dyn->pfid_addr4 = *SUNION2PF(&ke->pfrke_sa,
|
|
AF_INET);
|
|
w->pfrw_dyn->pfid_mask4 = *SUNION2PF(&pfr_mask,
|
|
AF_INET);
|
|
} else if (ke->pfrke_af == AF_INET6){
|
|
if (w->pfrw_dyn->pfid_acnt6++ > 0)
|
|
break;
|
|
pfr_prepare_network(&pfr_mask, AF_INET6, ke->pfrke_net);
|
|
w->pfrw_dyn->pfid_addr6 = *SUNION2PF(&ke->pfrke_sa,
|
|
AF_INET6);
|
|
w->pfrw_dyn->pfid_mask6 = *SUNION2PF(&pfr_mask,
|
|
AF_INET6);
|
|
}
|
|
break;
|
|
}
|
|
}
|
|
return (0);
|
|
}
|
|
|
|
int
|
|
pfr_clr_tables(struct pfr_table *filter, int *ndel, int flags)
|
|
{
|
|
struct pfr_ktableworkq workq;
|
|
struct pfr_ktable *p;
|
|
int xdel = 0;
|
|
|
|
ACCEPT_FLAGS(flags, PFR_FLAG_DUMMY | PFR_FLAG_ALLRSETS);
|
|
if (pfr_fix_anchor(filter->pfrt_anchor))
|
|
return (EINVAL);
|
|
if (pfr_table_count(filter, flags) < 0)
|
|
return (ENOENT);
|
|
|
|
SLIST_INIT(&workq);
|
|
RB_FOREACH(p, pfr_ktablehead, &pfr_ktables) {
|
|
if (pfr_skip_table(filter, p, flags))
|
|
continue;
|
|
if (!strcmp(p->pfrkt_anchor, PF_RESERVED_ANCHOR))
|
|
continue;
|
|
if (!(p->pfrkt_flags & PFR_TFLAG_ACTIVE))
|
|
continue;
|
|
p->pfrkt_nflags = p->pfrkt_flags & ~PFR_TFLAG_ACTIVE;
|
|
SLIST_INSERT_HEAD(&workq, p, pfrkt_workq);
|
|
xdel++;
|
|
}
|
|
if (!(flags & PFR_FLAG_DUMMY))
|
|
pfr_setflags_ktables(&workq);
|
|
if (ndel != NULL)
|
|
*ndel = xdel;
|
|
return (0);
|
|
}
|
|
|
|
int
|
|
pfr_add_tables(struct pfr_table *tbl, int size, int *nadd, int flags)
|
|
{
|
|
struct pfr_ktableworkq addq, changeq;
|
|
struct pfr_ktable *p, *q, *r, key;
|
|
int i, rv, xadd = 0;
|
|
long tzero = time_second;
|
|
|
|
ACCEPT_FLAGS(flags, PFR_FLAG_DUMMY);
|
|
SLIST_INIT(&addq);
|
|
SLIST_INIT(&changeq);
|
|
for (i = 0; i < size; i++) {
|
|
bcopy(tbl+i, &key.pfrkt_t, sizeof(key.pfrkt_t));
|
|
if (pfr_validate_table(&key.pfrkt_t, PFR_TFLAG_USRMASK,
|
|
flags & PFR_FLAG_USERIOCTL))
|
|
senderr(EINVAL);
|
|
key.pfrkt_flags |= PFR_TFLAG_ACTIVE;
|
|
p = RB_FIND(pfr_ktablehead, &pfr_ktables, &key);
|
|
if (p == NULL) {
|
|
p = pfr_create_ktable(&key.pfrkt_t, tzero, 1);
|
|
if (p == NULL)
|
|
senderr(ENOMEM);
|
|
SLIST_FOREACH(q, &addq, pfrkt_workq) {
|
|
if (!pfr_ktable_compare(p, q))
|
|
goto _skip;
|
|
}
|
|
SLIST_INSERT_HEAD(&addq, p, pfrkt_workq);
|
|
xadd++;
|
|
if (!key.pfrkt_anchor[0])
|
|
goto _skip;
|
|
|
|
/* find or create root table */
|
|
bzero(key.pfrkt_anchor, sizeof(key.pfrkt_anchor));
|
|
r = RB_FIND(pfr_ktablehead, &pfr_ktables, &key);
|
|
if (r != NULL) {
|
|
p->pfrkt_root = r;
|
|
goto _skip;
|
|
}
|
|
SLIST_FOREACH(q, &addq, pfrkt_workq) {
|
|
if (!pfr_ktable_compare(&key, q)) {
|
|
p->pfrkt_root = q;
|
|
goto _skip;
|
|
}
|
|
}
|
|
key.pfrkt_flags = 0;
|
|
r = pfr_create_ktable(&key.pfrkt_t, 0, 1);
|
|
if (r == NULL)
|
|
senderr(ENOMEM);
|
|
SLIST_INSERT_HEAD(&addq, r, pfrkt_workq);
|
|
p->pfrkt_root = r;
|
|
} else if (!(p->pfrkt_flags & PFR_TFLAG_ACTIVE)) {
|
|
SLIST_FOREACH(q, &changeq, pfrkt_workq)
|
|
if (!pfr_ktable_compare(&key, q))
|
|
goto _skip;
|
|
p->pfrkt_nflags = (p->pfrkt_flags &
|
|
~PFR_TFLAG_USRMASK) | key.pfrkt_flags;
|
|
SLIST_INSERT_HEAD(&changeq, p, pfrkt_workq);
|
|
xadd++;
|
|
}
|
|
_skip:
|
|
;
|
|
}
|
|
if (!(flags & PFR_FLAG_DUMMY)) {
|
|
pfr_insert_ktables(&addq);
|
|
pfr_setflags_ktables(&changeq);
|
|
} else
|
|
pfr_destroy_ktables(&addq, 0);
|
|
if (nadd != NULL)
|
|
*nadd = xadd;
|
|
return (0);
|
|
_bad:
|
|
pfr_destroy_ktables(&addq, 0);
|
|
return (rv);
|
|
}
|
|
|
|
int
|
|
pfr_del_tables(struct pfr_table *tbl, int size, int *ndel, int flags)
|
|
{
|
|
struct pfr_ktableworkq workq;
|
|
struct pfr_ktable *p, *q, key;
|
|
int i, xdel = 0;
|
|
|
|
ACCEPT_FLAGS(flags, PFR_FLAG_DUMMY);
|
|
SLIST_INIT(&workq);
|
|
for (i = 0; i < size; i++) {
|
|
bcopy(tbl+i, &key.pfrkt_t, sizeof(key.pfrkt_t));
|
|
if (pfr_validate_table(&key.pfrkt_t, 0,
|
|
flags & PFR_FLAG_USERIOCTL))
|
|
return (EINVAL);
|
|
p = RB_FIND(pfr_ktablehead, &pfr_ktables, &key);
|
|
if (p != NULL && (p->pfrkt_flags & PFR_TFLAG_ACTIVE)) {
|
|
SLIST_FOREACH(q, &workq, pfrkt_workq)
|
|
if (!pfr_ktable_compare(p, q))
|
|
goto _skip;
|
|
p->pfrkt_nflags = p->pfrkt_flags & ~PFR_TFLAG_ACTIVE;
|
|
SLIST_INSERT_HEAD(&workq, p, pfrkt_workq);
|
|
xdel++;
|
|
}
|
|
_skip:
|
|
;
|
|
}
|
|
|
|
if (!(flags & PFR_FLAG_DUMMY))
|
|
pfr_setflags_ktables(&workq);
|
|
if (ndel != NULL)
|
|
*ndel = xdel;
|
|
return (0);
|
|
}
|
|
|
|
int
|
|
pfr_get_tables(struct pfr_table *filter, struct pfr_table *tbl, int *size,
|
|
int flags)
|
|
{
|
|
struct pfr_ktable *p;
|
|
int n, nn;
|
|
|
|
PF_RULES_RASSERT();
|
|
|
|
ACCEPT_FLAGS(flags, PFR_FLAG_ALLRSETS);
|
|
if (pfr_fix_anchor(filter->pfrt_anchor))
|
|
return (EINVAL);
|
|
n = nn = pfr_table_count(filter, flags);
|
|
if (n < 0)
|
|
return (ENOENT);
|
|
if (n > *size) {
|
|
*size = n;
|
|
return (0);
|
|
}
|
|
RB_FOREACH(p, pfr_ktablehead, &pfr_ktables) {
|
|
if (pfr_skip_table(filter, p, flags))
|
|
continue;
|
|
if (n-- <= 0)
|
|
continue;
|
|
bcopy(&p->pfrkt_t, tbl++, sizeof(*tbl));
|
|
}
|
|
|
|
KASSERT(n == 0, ("%s: corruption detected (%d)", __func__, n));
|
|
|
|
*size = nn;
|
|
return (0);
|
|
}
|
|
|
|
int
|
|
pfr_get_tstats(struct pfr_table *filter, struct pfr_tstats *tbl, int *size,
|
|
int flags)
|
|
{
|
|
struct pfr_ktable *p;
|
|
struct pfr_ktableworkq workq;
|
|
int n, nn;
|
|
long tzero = time_second;
|
|
|
|
/* XXX PFR_FLAG_CLSTATS disabled */
|
|
ACCEPT_FLAGS(flags, PFR_FLAG_ALLRSETS);
|
|
if (pfr_fix_anchor(filter->pfrt_anchor))
|
|
return (EINVAL);
|
|
n = nn = pfr_table_count(filter, flags);
|
|
if (n < 0)
|
|
return (ENOENT);
|
|
if (n > *size) {
|
|
*size = n;
|
|
return (0);
|
|
}
|
|
SLIST_INIT(&workq);
|
|
RB_FOREACH(p, pfr_ktablehead, &pfr_ktables) {
|
|
if (pfr_skip_table(filter, p, flags))
|
|
continue;
|
|
if (n-- <= 0)
|
|
continue;
|
|
bcopy(&p->pfrkt_ts, tbl++, sizeof(*tbl));
|
|
SLIST_INSERT_HEAD(&workq, p, pfrkt_workq);
|
|
}
|
|
if (flags & PFR_FLAG_CLSTATS)
|
|
pfr_clstats_ktables(&workq, tzero,
|
|
flags & PFR_FLAG_ADDRSTOO);
|
|
|
|
KASSERT(n == 0, ("%s: corruption detected (%d)", __func__, n));
|
|
|
|
*size = nn;
|
|
return (0);
|
|
}
|
|
|
|
int
|
|
pfr_clr_tstats(struct pfr_table *tbl, int size, int *nzero, int flags)
|
|
{
|
|
struct pfr_ktableworkq workq;
|
|
struct pfr_ktable *p, key;
|
|
int i, xzero = 0;
|
|
long tzero = time_second;
|
|
|
|
ACCEPT_FLAGS(flags, PFR_FLAG_DUMMY | PFR_FLAG_ADDRSTOO);
|
|
SLIST_INIT(&workq);
|
|
for (i = 0; i < size; i++) {
|
|
bcopy(tbl + i, &key.pfrkt_t, sizeof(key.pfrkt_t));
|
|
if (pfr_validate_table(&key.pfrkt_t, 0, 0))
|
|
return (EINVAL);
|
|
p = RB_FIND(pfr_ktablehead, &pfr_ktables, &key);
|
|
if (p != NULL) {
|
|
SLIST_INSERT_HEAD(&workq, p, pfrkt_workq);
|
|
xzero++;
|
|
}
|
|
}
|
|
if (!(flags & PFR_FLAG_DUMMY))
|
|
pfr_clstats_ktables(&workq, tzero, flags & PFR_FLAG_ADDRSTOO);
|
|
if (nzero != NULL)
|
|
*nzero = xzero;
|
|
return (0);
|
|
}
|
|
|
|
int
|
|
pfr_set_tflags(struct pfr_table *tbl, int size, int setflag, int clrflag,
|
|
int *nchange, int *ndel, int flags)
|
|
{
|
|
struct pfr_ktableworkq workq;
|
|
struct pfr_ktable *p, *q, key;
|
|
int i, xchange = 0, xdel = 0;
|
|
|
|
ACCEPT_FLAGS(flags, PFR_FLAG_DUMMY);
|
|
if ((setflag & ~PFR_TFLAG_USRMASK) ||
|
|
(clrflag & ~PFR_TFLAG_USRMASK) ||
|
|
(setflag & clrflag))
|
|
return (EINVAL);
|
|
SLIST_INIT(&workq);
|
|
for (i = 0; i < size; i++) {
|
|
bcopy(tbl + i, &key.pfrkt_t, sizeof(key.pfrkt_t));
|
|
if (pfr_validate_table(&key.pfrkt_t, 0,
|
|
flags & PFR_FLAG_USERIOCTL))
|
|
return (EINVAL);
|
|
p = RB_FIND(pfr_ktablehead, &pfr_ktables, &key);
|
|
if (p != NULL && (p->pfrkt_flags & PFR_TFLAG_ACTIVE)) {
|
|
p->pfrkt_nflags = (p->pfrkt_flags | setflag) &
|
|
~clrflag;
|
|
if (p->pfrkt_nflags == p->pfrkt_flags)
|
|
goto _skip;
|
|
SLIST_FOREACH(q, &workq, pfrkt_workq)
|
|
if (!pfr_ktable_compare(p, q))
|
|
goto _skip;
|
|
SLIST_INSERT_HEAD(&workq, p, pfrkt_workq);
|
|
if ((p->pfrkt_flags & PFR_TFLAG_PERSIST) &&
|
|
(clrflag & PFR_TFLAG_PERSIST) &&
|
|
!(p->pfrkt_flags & PFR_TFLAG_REFERENCED))
|
|
xdel++;
|
|
else
|
|
xchange++;
|
|
}
|
|
_skip:
|
|
;
|
|
}
|
|
if (!(flags & PFR_FLAG_DUMMY))
|
|
pfr_setflags_ktables(&workq);
|
|
if (nchange != NULL)
|
|
*nchange = xchange;
|
|
if (ndel != NULL)
|
|
*ndel = xdel;
|
|
return (0);
|
|
}
|
|
|
|
int
|
|
pfr_ina_begin(struct pfr_table *trs, u_int32_t *ticket, int *ndel, int flags)
|
|
{
|
|
struct pfr_ktableworkq workq;
|
|
struct pfr_ktable *p;
|
|
struct pf_ruleset *rs;
|
|
int xdel = 0;
|
|
|
|
ACCEPT_FLAGS(flags, PFR_FLAG_DUMMY);
|
|
rs = pf_find_or_create_ruleset(trs->pfrt_anchor);
|
|
if (rs == NULL)
|
|
return (ENOMEM);
|
|
SLIST_INIT(&workq);
|
|
RB_FOREACH(p, pfr_ktablehead, &pfr_ktables) {
|
|
if (!(p->pfrkt_flags & PFR_TFLAG_INACTIVE) ||
|
|
pfr_skip_table(trs, p, 0))
|
|
continue;
|
|
p->pfrkt_nflags = p->pfrkt_flags & ~PFR_TFLAG_INACTIVE;
|
|
SLIST_INSERT_HEAD(&workq, p, pfrkt_workq);
|
|
xdel++;
|
|
}
|
|
if (!(flags & PFR_FLAG_DUMMY)) {
|
|
pfr_setflags_ktables(&workq);
|
|
if (ticket != NULL)
|
|
*ticket = ++rs->tticket;
|
|
rs->topen = 1;
|
|
} else
|
|
pf_remove_if_empty_ruleset(rs);
|
|
if (ndel != NULL)
|
|
*ndel = xdel;
|
|
return (0);
|
|
}
|
|
|
|
int
|
|
pfr_ina_define(struct pfr_table *tbl, struct pfr_addr *addr, int size,
|
|
int *nadd, int *naddr, u_int32_t ticket, int flags)
|
|
{
|
|
struct pfr_ktableworkq tableq;
|
|
struct pfr_kentryworkq addrq;
|
|
struct pfr_ktable *kt, *rt, *shadow, key;
|
|
struct pfr_kentry *p;
|
|
struct pfr_addr *ad;
|
|
struct pf_ruleset *rs;
|
|
int i, rv, xadd = 0, xaddr = 0;
|
|
|
|
PF_RULES_WASSERT();
|
|
|
|
ACCEPT_FLAGS(flags, PFR_FLAG_DUMMY | PFR_FLAG_ADDRSTOO);
|
|
if (size && !(flags & PFR_FLAG_ADDRSTOO))
|
|
return (EINVAL);
|
|
if (pfr_validate_table(tbl, PFR_TFLAG_USRMASK,
|
|
flags & PFR_FLAG_USERIOCTL))
|
|
return (EINVAL);
|
|
rs = pf_find_ruleset(tbl->pfrt_anchor);
|
|
if (rs == NULL || !rs->topen || ticket != rs->tticket)
|
|
return (EBUSY);
|
|
tbl->pfrt_flags |= PFR_TFLAG_INACTIVE;
|
|
SLIST_INIT(&tableq);
|
|
kt = RB_FIND(pfr_ktablehead, &pfr_ktables, (struct pfr_ktable *)tbl);
|
|
if (kt == NULL) {
|
|
kt = pfr_create_ktable(tbl, 0, 1);
|
|
if (kt == NULL)
|
|
return (ENOMEM);
|
|
SLIST_INSERT_HEAD(&tableq, kt, pfrkt_workq);
|
|
xadd++;
|
|
if (!tbl->pfrt_anchor[0])
|
|
goto _skip;
|
|
|
|
/* find or create root table */
|
|
bzero(&key, sizeof(key));
|
|
strlcpy(key.pfrkt_name, tbl->pfrt_name, sizeof(key.pfrkt_name));
|
|
rt = RB_FIND(pfr_ktablehead, &pfr_ktables, &key);
|
|
if (rt != NULL) {
|
|
kt->pfrkt_root = rt;
|
|
goto _skip;
|
|
}
|
|
rt = pfr_create_ktable(&key.pfrkt_t, 0, 1);
|
|
if (rt == NULL) {
|
|
pfr_destroy_ktables(&tableq, 0);
|
|
return (ENOMEM);
|
|
}
|
|
SLIST_INSERT_HEAD(&tableq, rt, pfrkt_workq);
|
|
kt->pfrkt_root = rt;
|
|
} else if (!(kt->pfrkt_flags & PFR_TFLAG_INACTIVE))
|
|
xadd++;
|
|
_skip:
|
|
shadow = pfr_create_ktable(tbl, 0, 0);
|
|
if (shadow == NULL) {
|
|
pfr_destroy_ktables(&tableq, 0);
|
|
return (ENOMEM);
|
|
}
|
|
SLIST_INIT(&addrq);
|
|
for (i = 0, ad = addr; i < size; i++, ad++) {
|
|
if (pfr_validate_addr(ad))
|
|
senderr(EINVAL);
|
|
if (pfr_lookup_addr(shadow, ad, 1) != NULL)
|
|
continue;
|
|
p = pfr_create_kentry(ad);
|
|
if (p == NULL)
|
|
senderr(ENOMEM);
|
|
if (pfr_route_kentry(shadow, p)) {
|
|
pfr_destroy_kentry(p);
|
|
continue;
|
|
}
|
|
SLIST_INSERT_HEAD(&addrq, p, pfrke_workq);
|
|
xaddr++;
|
|
}
|
|
if (!(flags & PFR_FLAG_DUMMY)) {
|
|
if (kt->pfrkt_shadow != NULL)
|
|
pfr_destroy_ktable(kt->pfrkt_shadow, 1);
|
|
kt->pfrkt_flags |= PFR_TFLAG_INACTIVE;
|
|
pfr_insert_ktables(&tableq);
|
|
shadow->pfrkt_cnt = (flags & PFR_FLAG_ADDRSTOO) ?
|
|
xaddr : NO_ADDRESSES;
|
|
kt->pfrkt_shadow = shadow;
|
|
} else {
|
|
pfr_clean_node_mask(shadow, &addrq);
|
|
pfr_destroy_ktable(shadow, 0);
|
|
pfr_destroy_ktables(&tableq, 0);
|
|
pfr_destroy_kentries(&addrq);
|
|
}
|
|
if (nadd != NULL)
|
|
*nadd = xadd;
|
|
if (naddr != NULL)
|
|
*naddr = xaddr;
|
|
return (0);
|
|
_bad:
|
|
pfr_destroy_ktable(shadow, 0);
|
|
pfr_destroy_ktables(&tableq, 0);
|
|
pfr_destroy_kentries(&addrq);
|
|
return (rv);
|
|
}
|
|
|
|
int
|
|
pfr_ina_rollback(struct pfr_table *trs, u_int32_t ticket, int *ndel, int flags)
|
|
{
|
|
struct pfr_ktableworkq workq;
|
|
struct pfr_ktable *p;
|
|
struct pf_ruleset *rs;
|
|
int xdel = 0;
|
|
|
|
PF_RULES_WASSERT();
|
|
|
|
ACCEPT_FLAGS(flags, PFR_FLAG_DUMMY);
|
|
rs = pf_find_ruleset(trs->pfrt_anchor);
|
|
if (rs == NULL || !rs->topen || ticket != rs->tticket)
|
|
return (0);
|
|
SLIST_INIT(&workq);
|
|
RB_FOREACH(p, pfr_ktablehead, &pfr_ktables) {
|
|
if (!(p->pfrkt_flags & PFR_TFLAG_INACTIVE) ||
|
|
pfr_skip_table(trs, p, 0))
|
|
continue;
|
|
p->pfrkt_nflags = p->pfrkt_flags & ~PFR_TFLAG_INACTIVE;
|
|
SLIST_INSERT_HEAD(&workq, p, pfrkt_workq);
|
|
xdel++;
|
|
}
|
|
if (!(flags & PFR_FLAG_DUMMY)) {
|
|
pfr_setflags_ktables(&workq);
|
|
rs->topen = 0;
|
|
pf_remove_if_empty_ruleset(rs);
|
|
}
|
|
if (ndel != NULL)
|
|
*ndel = xdel;
|
|
return (0);
|
|
}
|
|
|
|
int
|
|
pfr_ina_commit(struct pfr_table *trs, u_int32_t ticket, int *nadd,
|
|
int *nchange, int flags)
|
|
{
|
|
struct pfr_ktable *p, *q;
|
|
struct pfr_ktableworkq workq;
|
|
struct pf_ruleset *rs;
|
|
int xadd = 0, xchange = 0;
|
|
long tzero = time_second;
|
|
|
|
PF_RULES_WASSERT();
|
|
|
|
ACCEPT_FLAGS(flags, PFR_FLAG_DUMMY);
|
|
rs = pf_find_ruleset(trs->pfrt_anchor);
|
|
if (rs == NULL || !rs->topen || ticket != rs->tticket)
|
|
return (EBUSY);
|
|
|
|
SLIST_INIT(&workq);
|
|
RB_FOREACH(p, pfr_ktablehead, &pfr_ktables) {
|
|
if (!(p->pfrkt_flags & PFR_TFLAG_INACTIVE) ||
|
|
pfr_skip_table(trs, p, 0))
|
|
continue;
|
|
SLIST_INSERT_HEAD(&workq, p, pfrkt_workq);
|
|
if (p->pfrkt_flags & PFR_TFLAG_ACTIVE)
|
|
xchange++;
|
|
else
|
|
xadd++;
|
|
}
|
|
|
|
if (!(flags & PFR_FLAG_DUMMY)) {
|
|
for (p = SLIST_FIRST(&workq); p != NULL; p = q) {
|
|
q = SLIST_NEXT(p, pfrkt_workq);
|
|
pfr_commit_ktable(p, tzero);
|
|
}
|
|
rs->topen = 0;
|
|
pf_remove_if_empty_ruleset(rs);
|
|
}
|
|
if (nadd != NULL)
|
|
*nadd = xadd;
|
|
if (nchange != NULL)
|
|
*nchange = xchange;
|
|
|
|
return (0);
|
|
}
|
|
|
|
static void
|
|
pfr_commit_ktable(struct pfr_ktable *kt, long tzero)
|
|
{
|
|
struct pfr_ktable *shadow = kt->pfrkt_shadow;
|
|
int nflags;
|
|
|
|
PF_RULES_WASSERT();
|
|
|
|
if (shadow->pfrkt_cnt == NO_ADDRESSES) {
|
|
if (!(kt->pfrkt_flags & PFR_TFLAG_ACTIVE))
|
|
pfr_clstats_ktable(kt, tzero, 1);
|
|
} else if (kt->pfrkt_flags & PFR_TFLAG_ACTIVE) {
|
|
/* kt might contain addresses */
|
|
struct pfr_kentryworkq addrq, addq, changeq, delq, garbageq;
|
|
struct pfr_kentry *p, *q, *next;
|
|
struct pfr_addr ad;
|
|
|
|
pfr_enqueue_addrs(shadow, &addrq, NULL, 0);
|
|
pfr_mark_addrs(kt);
|
|
SLIST_INIT(&addq);
|
|
SLIST_INIT(&changeq);
|
|
SLIST_INIT(&delq);
|
|
SLIST_INIT(&garbageq);
|
|
pfr_clean_node_mask(shadow, &addrq);
|
|
for (p = SLIST_FIRST(&addrq); p != NULL; p = next) {
|
|
next = SLIST_NEXT(p, pfrke_workq); /* XXX */
|
|
pfr_copyout_addr(&ad, p);
|
|
q = pfr_lookup_addr(kt, &ad, 1);
|
|
if (q != NULL) {
|
|
if (q->pfrke_not != p->pfrke_not)
|
|
SLIST_INSERT_HEAD(&changeq, q,
|
|
pfrke_workq);
|
|
q->pfrke_mark = 1;
|
|
SLIST_INSERT_HEAD(&garbageq, p, pfrke_workq);
|
|
} else {
|
|
p->pfrke_tzero = tzero;
|
|
SLIST_INSERT_HEAD(&addq, p, pfrke_workq);
|
|
}
|
|
}
|
|
pfr_enqueue_addrs(kt, &delq, NULL, ENQUEUE_UNMARKED_ONLY);
|
|
pfr_insert_kentries(kt, &addq, tzero);
|
|
pfr_remove_kentries(kt, &delq);
|
|
pfr_clstats_kentries(&changeq, tzero, INVERT_NEG_FLAG);
|
|
pfr_destroy_kentries(&garbageq);
|
|
} else {
|
|
/* kt cannot contain addresses */
|
|
SWAP(struct radix_node_head *, kt->pfrkt_ip4,
|
|
shadow->pfrkt_ip4);
|
|
SWAP(struct radix_node_head *, kt->pfrkt_ip6,
|
|
shadow->pfrkt_ip6);
|
|
SWAP(int, kt->pfrkt_cnt, shadow->pfrkt_cnt);
|
|
pfr_clstats_ktable(kt, tzero, 1);
|
|
}
|
|
nflags = ((shadow->pfrkt_flags & PFR_TFLAG_USRMASK) |
|
|
(kt->pfrkt_flags & PFR_TFLAG_SETMASK) | PFR_TFLAG_ACTIVE)
|
|
& ~PFR_TFLAG_INACTIVE;
|
|
pfr_destroy_ktable(shadow, 0);
|
|
kt->pfrkt_shadow = NULL;
|
|
pfr_setflags_ktable(kt, nflags);
|
|
}
|
|
|
|
static int
|
|
pfr_validate_table(struct pfr_table *tbl, int allowedflags, int no_reserved)
|
|
{
|
|
int i;
|
|
|
|
if (!tbl->pfrt_name[0])
|
|
return (-1);
|
|
if (no_reserved && !strcmp(tbl->pfrt_anchor, PF_RESERVED_ANCHOR))
|
|
return (-1);
|
|
if (tbl->pfrt_name[PF_TABLE_NAME_SIZE-1])
|
|
return (-1);
|
|
for (i = strlen(tbl->pfrt_name); i < PF_TABLE_NAME_SIZE; i++)
|
|
if (tbl->pfrt_name[i])
|
|
return (-1);
|
|
if (pfr_fix_anchor(tbl->pfrt_anchor))
|
|
return (-1);
|
|
if (tbl->pfrt_flags & ~allowedflags)
|
|
return (-1);
|
|
return (0);
|
|
}
|
|
|
|
/*
|
|
* Rewrite anchors referenced by tables to remove slashes
|
|
* and check for validity.
|
|
*/
|
|
static int
|
|
pfr_fix_anchor(char *anchor)
|
|
{
|
|
size_t siz = MAXPATHLEN;
|
|
int i;
|
|
|
|
if (anchor[0] == '/') {
|
|
char *path;
|
|
int off;
|
|
|
|
path = anchor;
|
|
off = 1;
|
|
while (*++path == '/')
|
|
off++;
|
|
bcopy(path, anchor, siz - off);
|
|
memset(anchor + siz - off, 0, off);
|
|
}
|
|
if (anchor[siz - 1])
|
|
return (-1);
|
|
for (i = strlen(anchor); i < siz; i++)
|
|
if (anchor[i])
|
|
return (-1);
|
|
return (0);
|
|
}
|
|
|
|
static int
|
|
pfr_table_count(struct pfr_table *filter, int flags)
|
|
{
|
|
struct pf_ruleset *rs;
|
|
|
|
PF_RULES_ASSERT();
|
|
|
|
if (flags & PFR_FLAG_ALLRSETS)
|
|
return (pfr_ktable_cnt);
|
|
if (filter->pfrt_anchor[0]) {
|
|
rs = pf_find_ruleset(filter->pfrt_anchor);
|
|
return ((rs != NULL) ? rs->tables : -1);
|
|
}
|
|
return (pf_main_ruleset.tables);
|
|
}
|
|
|
|
static int
|
|
pfr_skip_table(struct pfr_table *filter, struct pfr_ktable *kt, int flags)
|
|
{
|
|
if (flags & PFR_FLAG_ALLRSETS)
|
|
return (0);
|
|
if (strcmp(filter->pfrt_anchor, kt->pfrkt_anchor))
|
|
return (1);
|
|
return (0);
|
|
}
|
|
|
|
static void
|
|
pfr_insert_ktables(struct pfr_ktableworkq *workq)
|
|
{
|
|
struct pfr_ktable *p;
|
|
|
|
SLIST_FOREACH(p, workq, pfrkt_workq)
|
|
pfr_insert_ktable(p);
|
|
}
|
|
|
|
static void
|
|
pfr_insert_ktable(struct pfr_ktable *kt)
|
|
{
|
|
|
|
PF_RULES_WASSERT();
|
|
|
|
RB_INSERT(pfr_ktablehead, &pfr_ktables, kt);
|
|
pfr_ktable_cnt++;
|
|
if (kt->pfrkt_root != NULL)
|
|
if (!kt->pfrkt_root->pfrkt_refcnt[PFR_REFCNT_ANCHOR]++)
|
|
pfr_setflags_ktable(kt->pfrkt_root,
|
|
kt->pfrkt_root->pfrkt_flags|PFR_TFLAG_REFDANCHOR);
|
|
}
|
|
|
|
static void
|
|
pfr_setflags_ktables(struct pfr_ktableworkq *workq)
|
|
{
|
|
struct pfr_ktable *p, *q;
|
|
|
|
for (p = SLIST_FIRST(workq); p; p = q) {
|
|
q = SLIST_NEXT(p, pfrkt_workq);
|
|
pfr_setflags_ktable(p, p->pfrkt_nflags);
|
|
}
|
|
}
|
|
|
|
static void
|
|
pfr_setflags_ktable(struct pfr_ktable *kt, int newf)
|
|
{
|
|
struct pfr_kentryworkq addrq;
|
|
|
|
PF_RULES_WASSERT();
|
|
|
|
if (!(newf & PFR_TFLAG_REFERENCED) &&
|
|
!(newf & PFR_TFLAG_PERSIST))
|
|
newf &= ~PFR_TFLAG_ACTIVE;
|
|
if (!(newf & PFR_TFLAG_ACTIVE))
|
|
newf &= ~PFR_TFLAG_USRMASK;
|
|
if (!(newf & PFR_TFLAG_SETMASK)) {
|
|
RB_REMOVE(pfr_ktablehead, &pfr_ktables, kt);
|
|
if (kt->pfrkt_root != NULL)
|
|
if (!--kt->pfrkt_root->pfrkt_refcnt[PFR_REFCNT_ANCHOR])
|
|
pfr_setflags_ktable(kt->pfrkt_root,
|
|
kt->pfrkt_root->pfrkt_flags &
|
|
~PFR_TFLAG_REFDANCHOR);
|
|
pfr_destroy_ktable(kt, 1);
|
|
pfr_ktable_cnt--;
|
|
return;
|
|
}
|
|
if (!(newf & PFR_TFLAG_ACTIVE) && kt->pfrkt_cnt) {
|
|
pfr_enqueue_addrs(kt, &addrq, NULL, 0);
|
|
pfr_remove_kentries(kt, &addrq);
|
|
}
|
|
if (!(newf & PFR_TFLAG_INACTIVE) && kt->pfrkt_shadow != NULL) {
|
|
pfr_destroy_ktable(kt->pfrkt_shadow, 1);
|
|
kt->pfrkt_shadow = NULL;
|
|
}
|
|
kt->pfrkt_flags = newf;
|
|
}
|
|
|
|
static void
|
|
pfr_clstats_ktables(struct pfr_ktableworkq *workq, long tzero, int recurse)
|
|
{
|
|
struct pfr_ktable *p;
|
|
|
|
SLIST_FOREACH(p, workq, pfrkt_workq)
|
|
pfr_clstats_ktable(p, tzero, recurse);
|
|
}
|
|
|
|
static void
|
|
pfr_clstats_ktable(struct pfr_ktable *kt, long tzero, int recurse)
|
|
{
|
|
struct pfr_kentryworkq addrq;
|
|
|
|
if (recurse) {
|
|
pfr_enqueue_addrs(kt, &addrq, NULL, 0);
|
|
pfr_clstats_kentries(&addrq, tzero, 0);
|
|
}
|
|
bzero(kt->pfrkt_packets, sizeof(kt->pfrkt_packets));
|
|
bzero(kt->pfrkt_bytes, sizeof(kt->pfrkt_bytes));
|
|
kt->pfrkt_match = kt->pfrkt_nomatch = 0;
|
|
kt->pfrkt_tzero = tzero;
|
|
}
|
|
|
|
static struct pfr_ktable *
|
|
pfr_create_ktable(struct pfr_table *tbl, long tzero, int attachruleset)
|
|
{
|
|
struct pfr_ktable *kt;
|
|
struct pf_ruleset *rs;
|
|
|
|
PF_RULES_WASSERT();
|
|
|
|
kt = malloc(sizeof(*kt), M_PFTABLE, M_NOWAIT|M_ZERO);
|
|
if (kt == NULL)
|
|
return (NULL);
|
|
kt->pfrkt_t = *tbl;
|
|
|
|
if (attachruleset) {
|
|
rs = pf_find_or_create_ruleset(tbl->pfrt_anchor);
|
|
if (!rs) {
|
|
pfr_destroy_ktable(kt, 0);
|
|
return (NULL);
|
|
}
|
|
kt->pfrkt_rs = rs;
|
|
rs->tables++;
|
|
}
|
|
|
|
if (!rn_inithead((void **)&kt->pfrkt_ip4,
|
|
offsetof(struct sockaddr_in, sin_addr) * 8) ||
|
|
!rn_inithead((void **)&kt->pfrkt_ip6,
|
|
offsetof(struct sockaddr_in6, sin6_addr) * 8)) {
|
|
pfr_destroy_ktable(kt, 0);
|
|
return (NULL);
|
|
}
|
|
kt->pfrkt_tzero = tzero;
|
|
|
|
return (kt);
|
|
}
|
|
|
|
static void
|
|
pfr_destroy_ktables(struct pfr_ktableworkq *workq, int flushaddr)
|
|
{
|
|
struct pfr_ktable *p, *q;
|
|
|
|
for (p = SLIST_FIRST(workq); p; p = q) {
|
|
q = SLIST_NEXT(p, pfrkt_workq);
|
|
pfr_destroy_ktable(p, flushaddr);
|
|
}
|
|
}
|
|
|
|
static void
|
|
pfr_destroy_ktable(struct pfr_ktable *kt, int flushaddr)
|
|
{
|
|
struct pfr_kentryworkq addrq;
|
|
|
|
if (flushaddr) {
|
|
pfr_enqueue_addrs(kt, &addrq, NULL, 0);
|
|
pfr_clean_node_mask(kt, &addrq);
|
|
pfr_destroy_kentries(&addrq);
|
|
}
|
|
if (kt->pfrkt_ip4 != NULL) {
|
|
RADIX_NODE_HEAD_DESTROY(kt->pfrkt_ip4);
|
|
free((caddr_t)kt->pfrkt_ip4, M_RTABLE);
|
|
}
|
|
if (kt->pfrkt_ip6 != NULL) {
|
|
RADIX_NODE_HEAD_DESTROY(kt->pfrkt_ip6);
|
|
free((caddr_t)kt->pfrkt_ip6, M_RTABLE);
|
|
}
|
|
if (kt->pfrkt_shadow != NULL)
|
|
pfr_destroy_ktable(kt->pfrkt_shadow, flushaddr);
|
|
if (kt->pfrkt_rs != NULL) {
|
|
kt->pfrkt_rs->tables--;
|
|
pf_remove_if_empty_ruleset(kt->pfrkt_rs);
|
|
}
|
|
free(kt, M_PFTABLE);
|
|
}
|
|
|
|
static int
|
|
pfr_ktable_compare(struct pfr_ktable *p, struct pfr_ktable *q)
|
|
{
|
|
int d;
|
|
|
|
if ((d = strncmp(p->pfrkt_name, q->pfrkt_name, PF_TABLE_NAME_SIZE)))
|
|
return (d);
|
|
return (strcmp(p->pfrkt_anchor, q->pfrkt_anchor));
|
|
}
|
|
|
|
static struct pfr_ktable *
|
|
pfr_lookup_table(struct pfr_table *tbl)
|
|
{
|
|
/* struct pfr_ktable start like a struct pfr_table */
|
|
return (RB_FIND(pfr_ktablehead, &pfr_ktables,
|
|
(struct pfr_ktable *)tbl));
|
|
}
|
|
|
|
int
|
|
pfr_match_addr(struct pfr_ktable *kt, struct pf_addr *a, sa_family_t af)
|
|
{
|
|
struct pfr_kentry *ke = NULL;
|
|
int match;
|
|
|
|
PF_RULES_RASSERT();
|
|
|
|
if (!(kt->pfrkt_flags & PFR_TFLAG_ACTIVE) && kt->pfrkt_root != NULL)
|
|
kt = kt->pfrkt_root;
|
|
if (!(kt->pfrkt_flags & PFR_TFLAG_ACTIVE))
|
|
return (0);
|
|
|
|
switch (af) {
|
|
#ifdef INET
|
|
case AF_INET:
|
|
{
|
|
struct sockaddr_in sin;
|
|
|
|
bzero(&sin, sizeof(sin));
|
|
sin.sin_len = sizeof(sin);
|
|
sin.sin_family = AF_INET;
|
|
sin.sin_addr.s_addr = a->addr32[0];
|
|
ke = (struct pfr_kentry *)rn_match(&sin, kt->pfrkt_ip4);
|
|
if (ke && KENTRY_RNF_ROOT(ke))
|
|
ke = NULL;
|
|
break;
|
|
}
|
|
#endif /* INET */
|
|
#ifdef INET6
|
|
case AF_INET6:
|
|
{
|
|
struct sockaddr_in6 sin6;
|
|
|
|
bzero(&sin6, sizeof(sin6));
|
|
sin6.sin6_len = sizeof(sin6);
|
|
sin6.sin6_family = AF_INET6;
|
|
bcopy(a, &sin6.sin6_addr, sizeof(sin6.sin6_addr));
|
|
ke = (struct pfr_kentry *)rn_match(&sin6, kt->pfrkt_ip6);
|
|
if (ke && KENTRY_RNF_ROOT(ke))
|
|
ke = NULL;
|
|
break;
|
|
}
|
|
#endif /* INET6 */
|
|
}
|
|
match = (ke && !ke->pfrke_not);
|
|
if (match)
|
|
kt->pfrkt_match++;
|
|
else
|
|
kt->pfrkt_nomatch++;
|
|
return (match);
|
|
}
|
|
|
|
void
|
|
pfr_update_stats(struct pfr_ktable *kt, struct pf_addr *a, sa_family_t af,
|
|
u_int64_t len, int dir_out, int op_pass, int notrule)
|
|
{
|
|
struct pfr_kentry *ke = NULL;
|
|
|
|
if (!(kt->pfrkt_flags & PFR_TFLAG_ACTIVE) && kt->pfrkt_root != NULL)
|
|
kt = kt->pfrkt_root;
|
|
if (!(kt->pfrkt_flags & PFR_TFLAG_ACTIVE))
|
|
return;
|
|
|
|
switch (af) {
|
|
#ifdef INET
|
|
case AF_INET:
|
|
{
|
|
struct sockaddr_in sin;
|
|
|
|
bzero(&sin, sizeof(sin));
|
|
sin.sin_len = sizeof(sin);
|
|
sin.sin_family = AF_INET;
|
|
sin.sin_addr.s_addr = a->addr32[0];
|
|
ke = (struct pfr_kentry *)rn_match(&sin, kt->pfrkt_ip4);
|
|
if (ke && KENTRY_RNF_ROOT(ke))
|
|
ke = NULL;
|
|
break;
|
|
}
|
|
#endif /* INET */
|
|
#ifdef INET6
|
|
case AF_INET6:
|
|
{
|
|
struct sockaddr_in6 sin6;
|
|
|
|
bzero(&sin6, sizeof(sin6));
|
|
sin6.sin6_len = sizeof(sin6);
|
|
sin6.sin6_family = AF_INET6;
|
|
bcopy(a, &sin6.sin6_addr, sizeof(sin6.sin6_addr));
|
|
ke = (struct pfr_kentry *)rn_match(&sin6, kt->pfrkt_ip6);
|
|
if (ke && KENTRY_RNF_ROOT(ke))
|
|
ke = NULL;
|
|
break;
|
|
}
|
|
#endif /* INET6 */
|
|
default:
|
|
panic("%s: unknown address family %u", __func__, af);
|
|
}
|
|
if ((ke == NULL || ke->pfrke_not) != notrule) {
|
|
if (op_pass != PFR_OP_PASS)
|
|
printf("pfr_update_stats: assertion failed.\n");
|
|
op_pass = PFR_OP_XPASS;
|
|
}
|
|
kt->pfrkt_packets[dir_out][op_pass]++;
|
|
kt->pfrkt_bytes[dir_out][op_pass] += len;
|
|
if (ke != NULL && op_pass != PFR_OP_XPASS &&
|
|
(kt->pfrkt_flags & PFR_TFLAG_COUNTERS)) {
|
|
if (ke->pfrke_counters == NULL)
|
|
ke->pfrke_counters = uma_zalloc(V_pfr_kcounters_z,
|
|
M_NOWAIT | M_ZERO);
|
|
if (ke->pfrke_counters != NULL) {
|
|
ke->pfrke_counters->pfrkc_packets[dir_out][op_pass]++;
|
|
ke->pfrke_counters->pfrkc_bytes[dir_out][op_pass] += len;
|
|
}
|
|
}
|
|
}
|
|
|
|
struct pfr_ktable *
|
|
pfr_attach_table(struct pf_ruleset *rs, char *name)
|
|
{
|
|
struct pfr_ktable *kt, *rt;
|
|
struct pfr_table tbl;
|
|
struct pf_anchor *ac = rs->anchor;
|
|
|
|
PF_RULES_WASSERT();
|
|
|
|
bzero(&tbl, sizeof(tbl));
|
|
strlcpy(tbl.pfrt_name, name, sizeof(tbl.pfrt_name));
|
|
if (ac != NULL)
|
|
strlcpy(tbl.pfrt_anchor, ac->path, sizeof(tbl.pfrt_anchor));
|
|
kt = pfr_lookup_table(&tbl);
|
|
if (kt == NULL) {
|
|
kt = pfr_create_ktable(&tbl, time_second, 1);
|
|
if (kt == NULL)
|
|
return (NULL);
|
|
if (ac != NULL) {
|
|
bzero(tbl.pfrt_anchor, sizeof(tbl.pfrt_anchor));
|
|
rt = pfr_lookup_table(&tbl);
|
|
if (rt == NULL) {
|
|
rt = pfr_create_ktable(&tbl, 0, 1);
|
|
if (rt == NULL) {
|
|
pfr_destroy_ktable(kt, 0);
|
|
return (NULL);
|
|
}
|
|
pfr_insert_ktable(rt);
|
|
}
|
|
kt->pfrkt_root = rt;
|
|
}
|
|
pfr_insert_ktable(kt);
|
|
}
|
|
if (!kt->pfrkt_refcnt[PFR_REFCNT_RULE]++)
|
|
pfr_setflags_ktable(kt, kt->pfrkt_flags|PFR_TFLAG_REFERENCED);
|
|
return (kt);
|
|
}
|
|
|
|
void
|
|
pfr_detach_table(struct pfr_ktable *kt)
|
|
{
|
|
|
|
PF_RULES_WASSERT();
|
|
KASSERT(kt->pfrkt_refcnt[PFR_REFCNT_RULE] > 0, ("%s: refcount %d\n",
|
|
__func__, kt->pfrkt_refcnt[PFR_REFCNT_RULE]));
|
|
|
|
if (!--kt->pfrkt_refcnt[PFR_REFCNT_RULE])
|
|
pfr_setflags_ktable(kt, kt->pfrkt_flags&~PFR_TFLAG_REFERENCED);
|
|
}
|
|
|
|
int
|
|
pfr_pool_get(struct pfr_ktable *kt, int *pidx, struct pf_addr *counter,
|
|
sa_family_t af)
|
|
{
|
|
struct pf_addr *addr, *cur, *mask;
|
|
union sockaddr_union uaddr, umask;
|
|
struct pfr_kentry *ke, *ke2 = NULL;
|
|
int idx = -1, use_counter = 0;
|
|
|
|
switch (af) {
|
|
case AF_INET:
|
|
uaddr.sin.sin_len = sizeof(struct sockaddr_in);
|
|
uaddr.sin.sin_family = AF_INET;
|
|
break;
|
|
case AF_INET6:
|
|
uaddr.sin6.sin6_len = sizeof(struct sockaddr_in6);
|
|
uaddr.sin6.sin6_family = AF_INET6;
|
|
break;
|
|
}
|
|
addr = SUNION2PF(&uaddr, af);
|
|
|
|
if (!(kt->pfrkt_flags & PFR_TFLAG_ACTIVE) && kt->pfrkt_root != NULL)
|
|
kt = kt->pfrkt_root;
|
|
if (!(kt->pfrkt_flags & PFR_TFLAG_ACTIVE))
|
|
return (-1);
|
|
|
|
if (pidx != NULL)
|
|
idx = *pidx;
|
|
if (counter != NULL && idx >= 0)
|
|
use_counter = 1;
|
|
if (idx < 0)
|
|
idx = 0;
|
|
|
|
_next_block:
|
|
ke = pfr_kentry_byidx(kt, idx, af);
|
|
if (ke == NULL) {
|
|
kt->pfrkt_nomatch++;
|
|
return (1);
|
|
}
|
|
pfr_prepare_network(&umask, af, ke->pfrke_net);
|
|
cur = SUNION2PF(&ke->pfrke_sa, af);
|
|
mask = SUNION2PF(&umask, af);
|
|
|
|
if (use_counter) {
|
|
/* is supplied address within block? */
|
|
if (!PF_MATCHA(0, cur, mask, counter, af)) {
|
|
/* no, go to next block in table */
|
|
idx++;
|
|
use_counter = 0;
|
|
goto _next_block;
|
|
}
|
|
PF_ACPY(addr, counter, af);
|
|
} else {
|
|
/* use first address of block */
|
|
PF_ACPY(addr, cur, af);
|
|
}
|
|
|
|
if (!KENTRY_NETWORK(ke)) {
|
|
/* this is a single IP address - no possible nested block */
|
|
PF_ACPY(counter, addr, af);
|
|
*pidx = idx;
|
|
kt->pfrkt_match++;
|
|
return (0);
|
|
}
|
|
for (;;) {
|
|
/* we don't want to use a nested block */
|
|
switch (af) {
|
|
case AF_INET:
|
|
ke2 = (struct pfr_kentry *)rn_match(&uaddr,
|
|
kt->pfrkt_ip4);
|
|
break;
|
|
case AF_INET6:
|
|
ke2 = (struct pfr_kentry *)rn_match(&uaddr,
|
|
kt->pfrkt_ip6);
|
|
break;
|
|
}
|
|
/* no need to check KENTRY_RNF_ROOT() here */
|
|
if (ke2 == ke) {
|
|
/* lookup return the same block - perfect */
|
|
PF_ACPY(counter, addr, af);
|
|
*pidx = idx;
|
|
kt->pfrkt_match++;
|
|
return (0);
|
|
}
|
|
|
|
/* we need to increase the counter past the nested block */
|
|
pfr_prepare_network(&umask, AF_INET, ke2->pfrke_net);
|
|
PF_POOLMASK(addr, addr, SUNION2PF(&umask, af), &pfr_ffaddr, af);
|
|
PF_AINC(addr, af);
|
|
if (!PF_MATCHA(0, cur, mask, addr, af)) {
|
|
/* ok, we reached the end of our main block */
|
|
/* go to next block in table */
|
|
idx++;
|
|
use_counter = 0;
|
|
goto _next_block;
|
|
}
|
|
}
|
|
}
|
|
|
|
static struct pfr_kentry *
|
|
pfr_kentry_byidx(struct pfr_ktable *kt, int idx, int af)
|
|
{
|
|
struct pfr_walktree w;
|
|
|
|
bzero(&w, sizeof(w));
|
|
w.pfrw_op = PFRW_POOL_GET;
|
|
w.pfrw_cnt = idx;
|
|
|
|
switch (af) {
|
|
#ifdef INET
|
|
case AF_INET:
|
|
kt->pfrkt_ip4->rnh_walktree(kt->pfrkt_ip4, pfr_walktree, &w);
|
|
return (w.pfrw_kentry);
|
|
#endif /* INET */
|
|
#ifdef INET6
|
|
case AF_INET6:
|
|
kt->pfrkt_ip6->rnh_walktree(kt->pfrkt_ip6, pfr_walktree, &w);
|
|
return (w.pfrw_kentry);
|
|
#endif /* INET6 */
|
|
default:
|
|
return (NULL);
|
|
}
|
|
}
|
|
|
|
void
|
|
pfr_dynaddr_update(struct pfr_ktable *kt, struct pfi_dynaddr *dyn)
|
|
{
|
|
struct pfr_walktree w;
|
|
|
|
bzero(&w, sizeof(w));
|
|
w.pfrw_op = PFRW_DYNADDR_UPDATE;
|
|
w.pfrw_dyn = dyn;
|
|
|
|
dyn->pfid_acnt4 = 0;
|
|
dyn->pfid_acnt6 = 0;
|
|
if (!dyn->pfid_af || dyn->pfid_af == AF_INET)
|
|
kt->pfrkt_ip4->rnh_walktree(kt->pfrkt_ip4, pfr_walktree, &w);
|
|
if (!dyn->pfid_af || dyn->pfid_af == AF_INET6)
|
|
kt->pfrkt_ip6->rnh_walktree(kt->pfrkt_ip6, pfr_walktree, &w);
|
|
}
|