commit 98815424093ca5426885218bc0afa5aa18f3e86e
Author: Jan Engelhardt <jengelh@xxxxxxxxxxxxxxx>
Date: Wed Jan 2 17:58:05 2008 +0100
[NETFILTER]: xt_hashlimit match, revision 1
Introduces the xt_hashlimit match revision 1. It adds support for
kernel-level inversion and grouping source and/or destination IP
addresses, allowing to limit on a per-subnet basis. While this would
technically obsolete xt_limit, xt_hashlimit is a more expensive due to
the hashbucketing.
Kernel-level inversion: Previously you had to do user-level inversion:
iptables -N foo
iptables -A foo -m hashlimit --hashlimit 5/s -j RETURN
iptables -A foo -j DROP
iptables -A INPUT -j foo
now it is simpler:
iptables -A INPUT -m hashlimit --hashlimit-over 5/s -j DROP
Signed-off-by: Jan Engelhardt <jengelh@xxxxxxxxxxxxxxx>
include/linux/netfilter/xt_hashlimit.h | 37 +++-
net/netfilter/xt_hashlimit.c | 311 +++++++++++++++++++++---
2 files changed, 315 insertions(+), 33 deletions(-)
diff --git a/include/linux/netfilter/xt_hashlimit.h b/include/linux/netfilter/xt_hashlimit.h
index c19972e..f15b104 100644
--- a/include/linux/netfilter/xt_hashlimit.h
+++ b/include/linux/netfilter/xt_hashlimit.h
@@ -9,13 +9,16 @@
/* details of this structure hidden by the implementation */
struct xt_hashlimit_htable;
-#define XT_HASHLIMIT_HASH_DIP 0x0001
-#define XT_HASHLIMIT_HASH_DPT 0x0002
-#define XT_HASHLIMIT_HASH_SIP 0x0004
-#define XT_HASHLIMIT_HASH_SPT 0x0008
+enum {
+ XT_HASHLIMIT_HASH_DIP = 1 << 0,
+ XT_HASHLIMIT_HASH_DPT = 1 << 1,
+ XT_HASHLIMIT_HASH_SIP = 1 << 2,
+ XT_HASHLIMIT_HASH_SPT = 1 << 3,
+ XT_HASHLIMIT_INVERT = 1 << 4,
+};
struct hashlimit_cfg {
- u_int32_t mode; /* bitmask of IPT_HASHLIMIT_HASH_* */
+ u_int32_t mode; /* bitmask of XT_HASHLIMIT_HASH_* */
u_int32_t avg; /* Average secs between packets * scale */
u_int32_t burst; /* Period multiplier for upper limit. */
@@ -37,4 +40,28 @@ struct xt_hashlimit_info {
struct xt_hashlimit_info *master;
} u;
};
+
+struct hashlimit_cfg1 {
+ u_int32_t mode; /* bitmask of XT_HASHLIMIT_HASH_* */
+ u_int32_t avg; /* Average secs between packets * scale */
+ u_int32_t burst; /* Period multiplier for upper limit. */
+
+ /* user specified */
+ u_int32_t size; /* how many buckets */
+ u_int32_t max; /* max number of entries */
+ u_int32_t gc_interval; /* gc interval */
+ u_int32_t expire; /* when do entries expire? */
+
+ u_int8_t srcmask, dstmask;
+};
+
+struct xt_hashlimit_mtinfo1 {
+ char name[IFNAMSIZ];
+ struct hashlimit_cfg1 cfg;
+
+ /* Used internally by the kernel */
+ struct xt_hashlimit_htable *hinfo __attribute__((aligned(8)));
+ struct xt_hashlimit_mtinfo1 *master;
+};
+
#endif /*_XT_HASHLIMIT_H*/
diff --git a/net/netfilter/xt_hashlimit.c b/net/netfilter/xt_hashlimit.c
index 8439ee4..b747a74 100644
--- a/net/netfilter/xt_hashlimit.c
+++ b/net/netfilter/xt_hashlimit.c
@@ -1,9 +1,10 @@
-/* iptables match extension to limit the number of packets per second
- * seperately for each hashbucket (sourceip/sourceport/dstip/dstport)
+/*
+ * xt_hashlimit - Netfilter module to limit the number of packets per time
+ * seperately for each hashbucket (sourceip/sourceport/dstip/dstport)
*
- * (C) 2003-2004 by Harald Welte <laforge@xxxxxxxxxxxxx>
- *
- * $Id: ipt_hashlimit.c 3244 2004-10-20 16:24:29Z laforge@xxxxxxxxxxxxx $
+ * (C) 2003-2004 by Harald Welte <laforge@xxxxxxxxxxxxx>
+ * Copyright © CC Computer Consultants GmbH, 2007 - 2008
+ * Jan Engelhardt <jengelh@xxxxxxxxxxxxxxx>
*
* Development of this code was funded by Astaro AG, http://www.astaro.com/
*/
@@ -71,7 +72,7 @@ struct xt_hashlimit_htable {
atomic_t use;
int family;
- struct hashlimit_cfg cfg; /* config */
+ struct hashlimit_cfg1 cfg; /* config */
/* used internally */
spinlock_t lock; /* lock for list_head */
@@ -174,7 +175,7 @@ dsthash_free(struct xt_hashlimit_htable *ht, struct dsthash_ent *ent)
}
static void htable_gc(unsigned long htlong);
-static int htable_create(struct xt_hashlimit_info *minfo, int family)
+static int htable_create_v0(struct xt_hashlimit_info *minfo, int family)
{
struct xt_hashlimit_htable *hinfo;
unsigned int size;
@@ -200,7 +201,18 @@ static int htable_create(struct xt_hashlimit_info *minfo, int family)
minfo->hinfo = hinfo;
/* copy match config into hashtable config */
- memcpy(&hinfo->cfg, &minfo->cfg, sizeof(hinfo->cfg));
+ hinfo->cfg.mode = minfo->cfg.mode;
+ hinfo->cfg.avg = minfo->cfg.avg;
+ hinfo->cfg.burst = hinfo->cfg.burst;
+ hinfo->cfg.max = hinfo->cfg.max;
+ hinfo->cfg.gc_interval = hinfo->cfg.gc_interval;
+ hinfo->cfg.expire = hinfo->cfg.expire;
+
+ if (family == AF_INET)
+ hinfo->cfg.srcmask = hinfo->cfg.dstmask = 32;
+ else
+ hinfo->cfg.srcmask = hinfo->cfg.dstmask = 128;
+
hinfo->cfg.size = size;
if (!hinfo->cfg.max)
hinfo->cfg.max = 8 * hinfo->cfg.size;
@@ -236,6 +248,72 @@ static int htable_create(struct xt_hashlimit_info *minfo, int family)
return 0;
}
+static int htable_create(struct xt_hashlimit_mtinfo1 *minfo,
+ unsigned int family)
+{
+ struct xt_hashlimit_htable *hinfo;
+ unsigned int size;
+ unsigned int i;
+
+ if (minfo->cfg.size) {
+ size = minfo->cfg.size;
+ } else {
+ size = (num_physpages << PAGE_SHIFT) / 16384 /
+ sizeof(struct list_head);
+ if (num_physpages > 1024 * 1024 * 1024 / PAGE_SIZE)
+ size = 8192;
+ if (size < 16)
+ size = 16;
+ }
+ /* FIXME: don't use vmalloc() here or anywhere else -HW */
+ hinfo = vmalloc(sizeof(struct xt_hashlimit_htable) +
+ sizeof(struct list_head) * size);
+ if (hinfo == NULL) {
+ printk(KERN_ERR "xt_hashlimit: unable to create hashtable\n");
+ return -1;
+ }
+ minfo->hinfo = hinfo;
+
+ /* copy match config into hashtable config */
+ memcpy(&hinfo->cfg, &minfo->cfg, sizeof(hinfo->cfg));
+ hinfo->cfg.size = size;
+ if (hinfo->cfg.max == 0)
+ hinfo->cfg.max = 8 * hinfo->cfg.size;
+ else if (hinfo->cfg.max < hinfo->cfg.size)
+ hinfo->cfg.max = hinfo->cfg.size;
+
+ for (i = 0; i < hinfo->cfg.size; i++)
+ INIT_HLIST_HEAD(&hinfo->hash[i]);
+
+ atomic_set(&hinfo->use, 1);
+ hinfo->count = 0;
+ hinfo->family = family;
+ hinfo->rnd_initialized = 0;
+ spin_lock_init(&hinfo->lock);
+ if (family == AF_INET)
+ hinfo->pde = create_proc_entry(minfo->name,
+ 0, hashlimit_procdir4);
+ else
+ hinfo->pde = create_proc_entry(minfo->name,
+ 0, hashlimit_procdir6);
+ if (hinfo->pde == NULL) {
+ vfree(hinfo);
+ return -1;
+ }
+ hinfo->pde->proc_fops = &dl_file_ops;
+ hinfo->pde->data = hinfo;
+
+ setup_timer(&hinfo->timer, htable_gc, (unsigned long)hinfo);
+ hinfo->timer.expires = jiffies + msecs_to_jiffies(hinfo->cfg.gc_interval);
+ add_timer(&hinfo->timer);
+
+ spin_lock_bh(&hashlimit_lock);
+ hlist_add_head(&hinfo->node, &hashlimit_htables);
+ spin_unlock_bh(&hashlimit_lock);
+
+ return 0;
+}
+
static bool select_all(const struct xt_hashlimit_htable *ht,
const struct dsthash_ent *he)
{
@@ -378,6 +456,46 @@ static inline void rateinfo_recalc(struct dsthash_ent *dh, unsigned long now)
dh->rateinfo.prev = now;
}
+static inline u_int32_t maskl(u_int32_t a, unsigned int l)
+{
+ return htonl(ntohl(a) & ~(~(u_int32_t)0 >> l));
+}
+
+static void hashlimit_ipv6_mask(__be32 *i, unsigned int p)
+{
+ switch (p) {
+ case 0:
+ i[0] = i[1] = 0;
+ i[2] = i[3] = 0;
+ break;
+ case 1 ... 31:
+ i[0] = maskl(i[0], p);
+ i[1] = i[2] = i[3] = 0;
+ break;
+ case 32:
+ i[1] = i[2] = i[3] = 0;
+ break;
+ case 33 ... 63:
+ i[1] = maskl(i[1], p - 32);
+ i[2] = i[3] = 0;
+ break;
+ case 64:
+ i[2] = i[3] = 0;
+ break;
+ case 65 ... 95:
+ i[2] = maskl(i[2], p - 64);
+ i[3] = 0;
+ case 96:
+ i[3] = 0;
+ break;
+ case 97 ... 127:
+ i[3] = maskl(i[3], p - 96);
+ break;
+ case 128:
+ break;
+ }
+}
+
static int
hashlimit_init_dst(const struct xt_hashlimit_htable *hinfo,
struct dsthash_dst *dst,
@@ -390,10 +508,12 @@ hashlimit_init_dst(const struct xt_hashlimit_htable *hinfo,
switch (hinfo->family) {
case AF_INET:
+ if (hinfo->cfg.mode & XT_HASHLIMIT_HASH_SIP)
+ dst->src.ip = maskl(ip_hdr(skb)->saddr,
+ hinfo->cfg.srcmask);
if (hinfo->cfg.mode & XT_HASHLIMIT_HASH_DIP)
- dst->dst.ip = ip_hdr(skb)->daddr;
- if (hinfo->cfg.mode & XT_HASHLIMIT_HASH_SIP)
- dst->src.ip = ip_hdr(skb)->saddr;
+ dst->dst.ip = maskl(ip_hdr(skb)->daddr,
+ hinfo->cfg.dstmask);
if (!(hinfo->cfg.mode &
(XT_HASHLIMIT_HASH_DPT | XT_HASHLIMIT_HASH_SPT)))
@@ -402,12 +522,16 @@ hashlimit_init_dst(const struct xt_hashlimit_htable *hinfo,
break;
#if defined(CONFIG_IP6_NF_IPTABLES) || defined(CONFIG_IP6_NF_IPTABLES_MODULE)
case AF_INET6:
- if (hinfo->cfg.mode & XT_HASHLIMIT_HASH_DIP)
- memcpy(&dst->dst.in6, &ipv6_hdr(skb)->daddr,
- sizeof(dst->dst.in6));
- if (hinfo->cfg.mode & XT_HASHLIMIT_HASH_SIP)
+ if (hinfo->cfg.mode & XT_HASHLIMIT_HASH_SIP) {
memcpy(&dst->src.in6, &ipv6_hdr(skb)->saddr,
sizeof(dst->src.in6));
+ hashlimit_ipv6_mask(dst->dst.ip6, hinfo->cfg.dstmask);
+ }
+ if (hinfo->cfg.mode & XT_HASHLIMIT_HASH_DIP) {
+ memcpy(&dst->dst.in6, &ipv6_hdr(skb)->daddr,
+ sizeof(dst->dst.in6));
+ hashlimit_ipv6_mask(dst->src.ip6, hinfo->cfg.srcmask);
+ }
if (!(hinfo->cfg.mode &
(XT_HASHLIMIT_HASH_DPT | XT_HASHLIMIT_HASH_SPT)))
@@ -447,10 +571,10 @@ hashlimit_init_dst(const struct xt_hashlimit_htable *hinfo,
}
static bool
-hashlimit_mt(const struct sk_buff *skb, const struct net_device *in,
- const struct net_device *out, const struct xt_match *match,
- const void *matchinfo, int offset, unsigned int protoff,
- bool *hotdrop)
+hashlimit_mt_v0(const struct sk_buff *skb, const struct net_device *in,
+ const struct net_device *out, const struct xt_match *match,
+ const void *matchinfo, int offset, unsigned int protoff,
+ bool *hotdrop)
{
const struct xt_hashlimit_info *r =
((const struct xt_hashlimit_info *)matchinfo)->u.master;
@@ -502,9 +626,62 @@ hotdrop:
}
static bool
-hashlimit_mt_check(const char *tablename, const void *inf,
- const struct xt_match *match, void *matchinfo,
- unsigned int hook_mask)
+hashlimit_mt(const struct sk_buff *skb, const struct net_device *in,
+ const struct net_device *out, const struct xt_match *match,
+ const void *matchinfo, int offset, unsigned int protoff,
+ bool *hotdrop)
+{
+ const struct xt_hashlimit_mtinfo1 *info = matchinfo;
+ struct xt_hashlimit_htable *hinfo = info->hinfo;
+ unsigned long now = jiffies;
+ struct dsthash_ent *dh;
+ struct dsthash_dst dst;
+
+ if (hashlimit_init_dst(hinfo, &dst, skb, protoff) < 0)
+ goto hotdrop;
+
+ spin_lock_bh(&hinfo->lock);
+ dh = dsthash_find(hinfo, &dst);
+ if (dh == NULL) {
+ dh = dsthash_alloc_init(hinfo, &dst);
+ if (dh == NULL) {
+ spin_unlock_bh(&hinfo->lock);
+ goto hotdrop;
+ }
+
+ dh->expires = jiffies + msecs_to_jiffies(hinfo->cfg.expire);
+ dh->rateinfo.prev = jiffies;
+ dh->rateinfo.credit = user2credits(hinfo->cfg.avg *
+ hinfo->cfg.burst);
+ dh->rateinfo.credit_cap = user2credits(hinfo->cfg.avg *
+ hinfo->cfg.burst);
+ dh->rateinfo.cost = user2credits(hinfo->cfg.avg);
+ } else {
+ /* update expiration timeout */
+ dh->expires = now + msecs_to_jiffies(hinfo->cfg.expire);
+ rateinfo_recalc(dh, now);
+ }
+
+ if (dh->rateinfo.credit >= dh->rateinfo.cost) {
+ /* below the limit */
+ dh->rateinfo.credit -= dh->rateinfo.cost;
+ spin_unlock_bh(&hinfo->lock);
+ return !(info->cfg.mode & XT_HASHLIMIT_INVERT);
+ }
+
+ spin_unlock_bh(&hinfo->lock);
+ /* default match is underlimit - so over the limit, we need to invert */
+ return info->cfg.mode & XT_HASHLIMIT_INVERT;
+
+ hotdrop:
+ *hotdrop = true;
+ return false;
+}
+
+static bool
+hashlimit_mt_check_v0(const char *tablename, const void *inf,
+ const struct xt_match *match, void *matchinfo,
+ unsigned int hook_mask)
{
struct xt_hashlimit_info *r = matchinfo;
@@ -536,7 +713,7 @@ hashlimit_mt_check(const char *tablename, const void *inf,
* create duplicate proc files. -HW */
mutex_lock(&hlimit_mutex);
r->hinfo = htable_find_get(r->name, match->family);
- if (!r->hinfo && htable_create(r, match->family) != 0) {
+ if (!r->hinfo && htable_create_v0(r, match->family) != 0) {
mutex_unlock(&hlimit_mutex);
return false;
}
@@ -547,33 +724,111 @@ hashlimit_mt_check(const char *tablename, const void *inf,
return true;
}
+static bool
+hashlimit_mt_check(const char *tablename, const void *inf,
+ const struct xt_match *match, void *matchinfo,
+ unsigned int hook_mask)
+{
+ struct xt_hashlimit_mtinfo1 *info = matchinfo;
+
+ /* Check for overflow. */
+ if (info->cfg.burst == 0 ||
+ user2credits(info->cfg.avg * info->cfg.burst) <
+ user2credits(info->cfg.avg)) {
+ printk(KERN_ERR "xt_hashlimit: overflow, try lower: %u/%u\n",
+ info->cfg.avg, info->cfg.burst);
+ return false;
+ }
+ if (info->cfg.gc_interval == 0 || info->cfg.expire == 0)
+ return false;
+ if (info->name[sizeof(info->name)-1] != '\0')
+ return false;
+ if (match->family == AF_INET) {
+ if (info->cfg.srcmask > 32 || info->cfg.dstmask > 32)
+ return false;
+ } else {
+ if (info->cfg.srcmask > 128 || info->cfg.dstmask > 128)
+ return false;
+ }
+
+ /* This is the best we've got: We cannot release and re-grab lock,
+ * since checkentry() is called before x_tables.c grabs xt_mutex.
+ * We also cannot grab the hashtable spinlock, since htable_create will
+ * call vmalloc, and that can sleep. And we cannot just re-search
+ * the list of htable's in htable_create(), since then we would
+ * create duplicate proc files. -HW */
+ mutex_lock(&hlimit_mutex);
+ info->hinfo = htable_find_get(info->name, match->family);
+ if (!info->hinfo && htable_create(info, match->family) != 0) {
+ mutex_unlock(&hlimit_mutex);
+ return false;
+ }
+ mutex_unlock(&hlimit_mutex);
+
+ /* Ugly hack: For SMP, we only want to use one set */
+ info->master = info;
+ return true;
+}
+
static void
-hashlimit_mt_destroy(const struct xt_match *match, void *matchinfo)
+hashlimit_mt_destroy_v0(const struct xt_match *match, void *matchinfo)
{
const struct xt_hashlimit_info *r = matchinfo;
htable_put(r->hinfo);
}
+static void
+hashlimit_mt_destroy(const struct xt_match *match, void *matchinfo)
+{
+ const struct xt_hashlimit_mtinfo1 *info = matchinfo;
+
+ htable_put(info->hinfo);
+}
+
static struct xt_match hashlimit_mt_reg[] __read_mostly = {
{
.name = "hashlimit",
+ .revision = 0,
.family = AF_INET,
- .match = hashlimit_mt,
+ .match = hashlimit_mt_v0,
+ .matchsize = sizeof(struct xt_hashlimit_info),
+ .checkentry = hashlimit_mt_check_v0,
+ .destroy = hashlimit_mt_destroy_v0,
+ .me = THIS_MODULE
+ },
+#if defined(CONFIG_IP6_NF_IPTABLES) || defined(CONFIG_IP6_NF_IPTABLES_MODULE)
+ {
+ .name = "hashlimit",
+ .revision = 0,
+ .family = AF_INET6,
+ .match = hashlimit_mt_v0,
.matchsize = sizeof(struct xt_hashlimit_info),
+ .checkentry = hashlimit_mt_check_v0,
+ .destroy = hashlimit_mt_destroy_v0,
+ .me = THIS_MODULE
+ },
+#endif
+ {
+ .name = "hashlimit",
+ .revision = 1,
+ .family = AF_INET,
+ .match = hashlimit_mt,
+ .matchsize = sizeof(struct xt_hashlimit_mtinfo1),
.checkentry = hashlimit_mt_check,
.destroy = hashlimit_mt_destroy,
- .me = THIS_MODULE
+ .me = THIS_MODULE,
},
#if defined(CONFIG_IP6_NF_IPTABLES) || defined(CONFIG_IP6_NF_IPTABLES_MODULE)
{
.name = "hashlimit",
+ .revision = 1,
.family = AF_INET6,
.match = hashlimit_mt,
- .matchsize = sizeof(struct xt_hashlimit_info),
+ .matchsize = sizeof(struct xt_hashlimit_mtinfo1),
.checkentry = hashlimit_mt_check,
.destroy = hashlimit_mt_destroy,
- .me = THIS_MODULE
+ .me = THIS_MODULE,
},
#endif
};
-
To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
