Netfilter updates for net-next
Subsystems (6)
[email protected]
maintainercover Netfilter updates for net-next
From: Pablo Neira Ayuso <[email protected]>
Hi David,
The following patchset contains Netfilter/IPVS updates for your net-next tree, they are:
* Better performance in nfnetlink_queue by avoiding copy from the packet to netlink message, from Eric Dumazet.
* Remove unnecessary locking in the exit path of ebt_ulog, from Gao Feng.
* Use new function ipv6_iface_scope_id in nf_ct_ipv6, from Hannes Frederic Sowa.
* A couple of sparse fixes for IPVS, from Julian Anastasov.
* Use xor hashing in nfnetlink_queue, as suggested by Eric Dumazet, from myself.
* Allow to dump expectations per master conntrack via ctnetlink, from myself.
* A couple of cleanups to use PTR_RET in module init path, from Silviu-Mihai Popescu.
* Remove nf_conntrack module a bit faster if netns are in use, from Vladimir Davydov.
* Use checksum_partial in ip6t_NPT, from YOSHIFUJI Hideaki.
* Sparse fix for nf_conntrack, from Stephen Hemminger.
You can pull these changes from:
git://1984.lsi.us.es/nf-next master
Thanks!
Eric Dumazet (1):
netfilter: nfnetlink_queue: zero copy support
Gao feng (1):
netfilter: ebt_ulog: remove unnecessary spin lock protection
Hannes Frederic Sowa (1):
netfilter: nf_ct_ipv6: use ipv6_iface_scope_id in conntrack to return scope id
Julian Anastasov (2):
ipvs: fix hashing in ip_vs_svc_hashkey
ipvs: fix some sparse warnings
Pablo Neira Ayuso (2):
netfilter: nfnetlink_queue: use xor hash function to distribute instances
netfilter: ctnetlink: allow to dump expectation per master conntrack
Silviu-Mihai Popescu (2):
ipv4: netfilter: use PTR_RET instead of IS_ERR + PTR_ERR
bridge: netfilter: use PTR_RET instead of IS_ERR + PTR_ERR
Vladimir Davydov (1):
netfilter: nf_conntrack: speed up module removal path if netns in use
YOSHIFUJI Hideaki (1):
netfilter: ip6t_NPT: Use csum_partial()
stephen hemminger (1):
netfilter: nf_conntrack: add include to fix sparse warning
include/net/ip_vs.h | 2 +- include/net/netfilter/nf_conntrack_core.h | 1 + net/bridge/netfilter/ebt_ulog.c | 3 +- net/bridge/netfilter/ebtable_broute.c | 4 +- net/ipv4/netfilter/arptable_filter.c | 4 +- net/ipv6/netfilter/ip6t_NPT.c | 11 +-- net/ipv6/netfilter/nf_conntrack_l3proto_ipv6.c | 8 +- net/netfilter/ipvs/ip_vs_core.c | 8 +- net/netfilter/ipvs/ip_vs_ctl.c | 8 +- net/netfilter/ipvs/ip_vs_est.c | 2 +- net/netfilter/nf_conntrack_core.c | 47 +++++++---- net/netfilter/nf_conntrack_netlink.c | 100 ++++++++++++++++++++++-- net/netfilter/nf_conntrack_standalone.c | 16 ++-- net/netfilter/nfnetlink_queue_core.c | 96 +++++++++++++++++------ 14 files changed, 228 insertions(+), 82 deletions(-)
Signature · 1 line
1.7.10.4
[email protected]
maintainerPATCH 1/12 netfilter: ip6t_NPT: Use csum_partial()
From: YOSHIFUJI Hideaki <[email protected]>
[ Some fixes went into mainstream before this patch, so I needed
to rebase it upon the current tree, that's why it's different from the original one posted on the list --pablo ]
--- net/ipv6/netfilter/ip6t_NPT.c | 11 +++-------- 1 file changed, 3 insertions(+), 8 deletions(-)
net/ipv6/netfilter/ip6t_NPT.c +3 -8
@@ -18,9 +18,8 @@static int ip6t_npt_checkentry (const struct xt_tgchk_param *par) {struct ip6t_npt_tginfo *npt = par->targinfo; -__wsum src_sum =0 , dst_sum =0 ;struct in6_addr pfx; -unsigned int i; +__wsum src_sum, dst_sum;if (npt->src_pfx_len >64 || npt->dst_pfx_len >64 )return -EINVAL ;
@@ -33,12 +32,8 @@ static int ip6t_npt_checkentry(const struct xt_tgchk_param *par)if (!ipv6_addr_equal (&pfx, &npt->dst_pfx.in6))return -EINVAL ; -for (i =0 ; i <ARRAY_SIZE (npt->src_pfx.in6.s6_addr16); i++) { - src_sum =csum_add (src_sum, - (__force __wsum )npt->src_pfx.in6.s6_addr16[i]); - dst_sum =csum_add (dst_sum, - (__force __wsum )npt->dst_pfx.in6.s6_addr16[i]); - } + src_sum =csum_partial (&npt->src_pfx.in6,sizeof (npt->src_pfx.in6),0 ); + dst_sum =csum_partial (&npt->dst_pfx.in6,sizeof (npt->dst_pfx.in6),0 ); npt->adjustment = ~csum_fold (csum_sub (src_sum, dst_sum));return 0 ;
Signature · 1 line
1.7.10.4
[email protected]
maintainerPATCH 2/12 ipv4: netfilter: use PTR_RET instead of IS_ERR + PTR_ERR
In reply to [email protected]From: Silviu-Mihai Popescu <[email protected]>
This uses PTR_RET instead of IS_ERR and PTR_ERR in order to increase readability.
--- net/ipv4/netfilter/arptable_filter.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-)
net/ipv4/netfilter/arptable_filter.c +1 -3
@@ -48,9 +48,7 @@ static int __net_init arptable_filter_net_init(struct net *net) net->ipv4.arptable_filter =arpt_register_table (net, &packet_filter, repl);kfree (repl); -if (IS_ERR (net->ipv4.arptable_filter)) -return PTR_ERR (net->ipv4.arptable_filter); -return 0 ; +return PTR_RET (net->ipv4.arptable_filter); } static void __net_exit arptable_filter_net_exit(struct net *net)
Signature · 1 line
1.7.10.4
[email protected]
maintainerPATCH 3/12 bridge: netfilter: use PTR_RET instead of IS_ERR + PTR_ERR
In reply to [email protected]From: Silviu-Mihai Popescu <[email protected]>
This uses PTR_RET instead of IS_ERR and PTR_ERR in order to increase readability.
--- net/bridge/netfilter/ebtable_broute.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-)
net/bridge/netfilter/ebtable_broute.c +1 -3
@@ -64,9 +64,7 @@ static int ebt_broute(struct sk_buff *skb)static int __net_init broute_net_init (struct net *net) { net->xt.broute_table =ebt_register_table (net, &broute_table); -if (IS_ERR (net->xt.broute_table)) -return PTR_ERR (net->xt.broute_table); -return 0 ; +return PTR_RET (net->xt.broute_table); } static void __net_exit broute_net_exit(struct net *net)
Signature · 1 line
1.7.10.4
[email protected]
maintainerPATCH 4/12 netfilter: nf_ct_ipv6: use ipv6_iface_scope_id in conntrack to return scope id
In reply to [email protected]From: Hannes Frederic Sowa <[email protected]>
As in (842df07 ipv6: use newly introduced __ipv6_addr_needs_scope_id and ipv6_iface_scope_id).
--- net/ipv6/netfilter/nf_conntrack_l3proto_ipv6.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-)
net/ipv6/netfilter/nf_conntrack_l3proto_ipv6.c +2 -6
@@ -330,12 +330,8 @@ ipv6_getorigdst(struct sock *sk, int optval, void __user *user, int *len) sizeof(sin6.sin6_addr));nf_ct_put (ct); - -if (ipv6_addr_type (&sin6.sin6_addr) &IPV6_ADDR_LINKLOCAL ) - sin6.sin6_scope_id = sk->sk_bound_dev_if; -else - sin6.sin6_scope_id =0 ; - + sin6.sin6_scope_id =ipv6_iface_scope_id (&sin6.sin6_addr, + sk->sk_bound_dev_if);return copy_to_user (user, &sin6,sizeof (sin6)) ? -EFAULT :0 ; }
Signature · 1 line
1.7.10.4
[email protected]
maintainerPATCH 5/12 netfilter: ebt_ulog: remove unnecessary spin lock protection
In reply to [email protected]From: Gao feng <[email protected]>
No need for spinlock to protect the netlink skb in the ebt_ulog_fini path. We are sure there is noone using it at that stage.
--- net/bridge/netfilter/ebt_ulog.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)
net/bridge/netfilter/ebt_ulog.c +1 -2
@@ -319,12 +319,11 @@ static void __exit ebt_ulog_fini(void)for (i =0 ; i <EBT_ULOG_MAXNLGROUPS ; i++) { ub = &ulog_buffers[i];del_timer (&ub->timer); -spin_lock_bh (&ub->lock); +if (ub->skb) {kfree_skb (ub->skb); ub->skb =NULL ; } -spin_unlock_bh (&ub->lock); }netlink_kernel_release (ebtulognl); }
Signature · 1 line
1.7.10.4
[email protected]
maintainerPATCH 6/12 netfilter: nfnetlink_queue: use xor hash function to distribute instances
In reply to [email protected]From: Pablo Neira Ayuso <[email protected]>
Thanks to Eric Dumazet for suggesting this during the NFWS.
--- net/netfilter/nfnetlink_queue_core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
net/netfilter/nfnetlink_queue_core.c +1 -1
@@ -73,7 +73,7 @@ static struct hlist_head instance_table[INSTANCE_BUCKETS] __read_mostly;static inline u_int8_t instance_hashfn (u_int16_t queue_num) { -return ((queue_num >>8 ) | queue_num) %INSTANCE_BUCKETS ; +return ((queue_num >>8 ) ^ queue_num) %INSTANCE_BUCKETS ; } static struct nfqnl_instance *
Signature · 1 line
1.7.10.4
[email protected]
maintainerPATCH 7/12 ipvs: fix hashing in ip_vs_svc_hashkey
In reply to [email protected]From: Julian Anastasov <[email protected]>
net is a pointer in host order, mix it properly with other keys in network order. Fixes sparse warning.
--- net/netfilter/ipvs/ip_vs_ctl.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)
net/netfilter/ipvs/ip_vs_ctl.c +5 -3
@@ -271,16 +271,18 @@ ip_vs_svc_hashkey(struct net *net, int af, unsigned int proto, { registerunsigned int porth =ntohs (port);__be32 addr_fold = addr->ip; +__u32 ahash;#ifdef CONFIG_IP_VS_IPV6 if (af ==AF_INET6 ) addr_fold = addr->ip6[0 ]^addr->ip6[1 ]^ addr->ip6[2 ]^addr->ip6[3 ];#endif - addr_fold ^= ((size_t )net>>8 ); + ahash =ntohl (addr_fold); + ahash ^= ((size_t ) net >>8 ); -return (proto^ntohl (addr_fold)^(porth>>IP_VS_SVC_TAB_BITS )^porth) - &IP_VS_SVC_TAB_MASK ; +return (proto ^ ahash ^ (porth >>IP_VS_SVC_TAB_BITS ) ^ porth) & +IP_VS_SVC_TAB_MASK ; }/*
Signature · 1 line
1.7.10.4
[email protected]
maintainerPATCH 8/12 ipvs: fix some sparse warnings
In reply to [email protected]From: Julian Anastasov <[email protected]>
Add missing __percpu annotations and make ip_vs_net_id static.
--- include/net/ip_vs.h | 2 +- net/netfilter/ipvs/ip_vs_core.c | 8 +------- net/netfilter/ipvs/ip_vs_est.c | 2 +- 3 files changed, 3 insertions(+), 9 deletions(-)
include/net/ip_vs.h +1 -1
@@ -459,7 +459,7 @@ struct ip_vs_estimator {struct ip_vs_stats {struct ip_vs_stats_user ustats;/* statistics */ struct ip_vs_estimator est;/* estimator */ -struct ip_vs_cpu_stats *cpustats;/* per cpu counters */ +struct ip_vs_cpu_stats __percpu *cpustats;/* per cpu counters */ spinlock_t lock;/* spin lock */ struct ip_vs_stats_user ustats0;/* reset values */ };
net/netfilter/ipvs/ip_vs_core.c +1 -7
@@ -69,10 +69,7 @@ EXPORT_SYMBOL(ip_vs_conn_put);EXPORT_SYMBOL (ip_vs_get_debug_level);#endif -int ip_vs_net_id__read_mostly ; -#ifdef IP_VS_GENERIC_NETNS -EXPORT_SYMBOL (ip_vs_net_id); -#endif +static int ip_vs_net_id__read_mostly ;/* netns cnt used for uniqueness */ static atomic_t ipvs_netns_cnt =ATOMIC_INIT (0 );
@@ -1181,9 +1178,6 @@ ip_vs_out(unsigned int hooknum, struct sk_buff *skb, int af) iph.len)))) {#ifdef CONFIG_IP_VS_IPV6 if (af ==AF_INET6 ) { -struct net *net = -dev_net (skb_dst (skb)->dev); -if (!skb->dev) skb->dev = net->loopback_dev; icmpv6_send(skb,
net/netfilter/ipvs/ip_vs_est.c +1 -1
@@ -56,7 +56,7 @@* Make a summary from each cpu */ static void ip_vs_read_cpu_stats (struct ip_vs_stats_user *sum, -struct ip_vs_cpu_stats *stats) +struct ip_vs_cpu_stats __percpu *stats) {int i;
Signature · 1 line
1.7.10.4
[email protected]
maintainerPATCH 9/12 netfilter: ctnetlink: allow to dump expectation per master conntrack
In reply to [email protected]From: Pablo Neira Ayuso <[email protected]>
This patch adds the ability to dump all existing expectations per master conntrack.
--- net/netfilter/nf_conntrack_netlink.c | 100 ++++++++++++++++++++++++++++++++-- 1 file changed, 95 insertions(+), 5 deletions(-)
net/netfilter/nf_conntrack_netlink.c +95 -5
@@ -2409,6 +2409,92 @@ out:return skb->len; } +static int +ctnetlink_exp_ct_dump_table (struct sk_buff *skb,struct netlink_callback *cb) +{ +struct nf_conntrack_expect *exp, *last; +struct nfgenmsg *nfmsg =nlmsg_data (cb->nlh); +struct nf_conn *ct = cb->data; +struct nf_conn_help *help =nfct_help (ct); +u_int8_t l3proto = nfmsg->nfgen_family; + +if (cb->args[0 ]) +return 0 ; + +rcu_read_lock (); + last = (struct nf_conntrack_expect *)cb->args[1 ]; +restart: +hlist_for_each_entry (exp, &help->expectations, lnode) { +if (l3proto && exp->tuple.src.l3num != l3proto) +continue ; +if (cb->args[1 ]) { +if (exp != last) +continue ; + cb->args[1 ] =0 ; + } +if (ctnetlink_exp_fill_info (skb,NETLINK_CB (cb->skb).portid, + cb->nlh->nlmsg_seq, +IPCTNL_MSG_EXP_NEW , + exp) <0 ) { +if (!atomic_inc_not_zero (&exp->use)) +continue ; + cb->args[1 ] = (unsigned long )exp; + goto out; + } + } +if (cb->args[1 ]) { + cb->args[1 ] =0 ; + goto restart; + } + cb->args[0 ] =1 ; +out: +rcu_read_unlock (); +if (last) +nf_ct_expect_put (last); + +return skb->len; +} + +static int ctnetlink_dump_exp_ct (struct sock *ctnl,struct sk_buff *skb, +const struct nlmsghdr *nlh, +const struct nlattr *const cda[]) +{ +int err; +struct net *net =sock_net (ctnl); +struct nfgenmsg *nfmsg =nlmsg_data (nlh); +u_int8_t u3 = nfmsg->nfgen_family; +struct nf_conntrack_tuple tuple; +struct nf_conntrack_tuple_hash *h; +struct nf_conn *ct; +u16 zone =0 ; +struct netlink_dump_control c = { + .dump = ctnetlink_exp_ct_dump_table, + .done = ctnetlink_exp_done, + }; + + err =ctnetlink_parse_tuple (cda, &tuple,CTA_EXPECT_MASTER , u3); +if (err <0 ) +return err; + +if (cda[CTA_EXPECT_ZONE ]) { + err =ctnetlink_parse_zone (cda[CTA_EXPECT_ZONE ], &zone); +if (err <0 ) +return err; + } + + h =nf_conntrack_find_get (net, zone, &tuple); +if (!h) +return -ENOENT ; + + ct =nf_ct_tuplehash_to_ctrack (h); + c.data = ct; + + err =netlink_dump_start (ctnl, skb, nlh, &c); +nf_ct_put (ct); + +return err; +} +static const struct nla_policy exp_nla_policy[CTA_EXPECT_MAX +1 ] = { [CTA_EXPECT_MASTER ] = { .type =NLA_NESTED }, [CTA_EXPECT_TUPLE ] = { .type =NLA_NESTED },
@@ -2439,11 +2525,15 @@ ctnetlink_get_expect(struct sock *ctnl, struct sk_buff *skb,int err;if (nlh->nlmsg_flags &NLM_F_DUMP ) { -struct netlink_dump_control c = { - .dump = ctnetlink_exp_dump_table, - .done = ctnetlink_exp_done, - }; -return netlink_dump_start (ctnl, skb, nlh, &c); +if (cda[CTA_EXPECT_MASTER ]) +return ctnetlink_dump_exp_ct (ctnl, skb, nlh, cda); +else { +struct netlink_dump_control c = { + .dump = ctnetlink_exp_dump_table, + .done = ctnetlink_exp_done, + }; +return netlink_dump_start (ctnl, skb, nlh, &c); + } } err =ctnetlink_parse_zone (cda[CTA_EXPECT_ZONE ], &zone);
Signature · 1 line
1.7.10.4
[email protected]
maintainerPATCH 10/12 netfilter: nfnetlink_queue: zero copy support
In reply to [email protected]From: Eric Dumazet <[email protected]>
nfqnl_build_packet_message() actually copy the packet inside the netlink message, while it can instead use zero copy.
Make sure the skb 'copy' is the last component of the cooked netlink message, as we cant add anything after it.
Patch cooked in Copenhagen at Netfilter Workshop ;)
Still to be addressed in separate patches :
-GRO/GSO packets are segmented in nf_queue() and checksummed in nfqnl_build_packet_message().
Proper support for GSO/GRO packets (no segmentation, and no checksumming) needs application cooperation, if we want no regressions.
--- net/netfilter/nfnetlink_queue_core.c | 94 ++++++++++++++++++++++++++-------- 1 file changed, 72 insertions(+), 22 deletions(-)
net/netfilter/nfnetlink_queue_core.c +72 -22
@@ -217,14 +217,59 @@ nfqnl_flush(struct nfqnl_instance *queue, nfqnl_cmpfn cmpfn, unsigned long data)spin_unlock_bh (&queue->lock); } +static void +nfqnl_zcopy (struct sk_buff *to,const struct sk_buff *from,int len,int hlen) +{ +int i, j =0 ; +int plen =0 ;/* length of skb->head fragment */ +struct page *page; +unsigned int offset; + +/* dont bother with small payloads */ +if (len <=skb_tailroom (to)) { +skb_copy_bits (from,0 ,skb_put (to, len), len); +return ; + } + +if (hlen) { +skb_copy_bits (from,0 ,skb_put (to, hlen), hlen); + len -= hlen; + }else { + plen =min_t (int,skb_headlen (from), len); +if (plen) { + page =virt_to_head_page (from->head); + offset = from->data - (unsigned char *)page_address (page); +__skb_fill_page_desc (to,0 , page, offset, plen); +get_page (page); + j =1 ; + len -= plen; + } + } + + to->truesize += len + plen; + to->len += len + plen; + to->data_len += len + plen; + +for (i =0 ; i <skb_shinfo (from)->nr_frags; i++) { +if (!len) +break ; +skb_shinfo (to)->frags[j] =skb_shinfo (from)->frags[i]; +skb_shinfo (to)->frags[j].size =min_t (int,skb_shinfo (to)->frags[j].size, len); + len -=skb_shinfo (to)->frags[j].size; +skb_frag_ref (to, j); + j++; + } +skb_shinfo (to)->nr_frags = j; +} +static struct sk_buff *nfqnl_build_packet_message (struct nfqnl_instance *queue,struct nf_queue_entry *entry,__be32 **packet_id_ptr) { -sk_buff_data_t old_tail;size_t size;size_t data_len =0 , cap_len =0 ; +int hlen =0 ;struct sk_buff *skb;struct nlattr *nla;struct nfqnl_msg_packet_hdr *pmsg;
@@ -246,8 +291,10 @@ nfqnl_build_packet_message(struct nfqnl_instance *queue,#endif +nla_total_size (sizeof (u_int32_t))/* mark */ +nla_total_size (sizeof (struct nfqnl_msg_packet_hw )) - +nla_total_size (sizeof (struct nfqnl_msg_packet_timestamp ) - +nla_total_size (sizeof (u_int32_t)));/* cap_len */ + +nla_total_size (sizeof (u_int32_t));/* cap_len */ + +if (entskb->tstamp.tv64) + size +=nla_total_size (sizeof (struct nfqnl_msg_packet_timestamp )); outdev = entry->outdev;
@@ -265,7 +312,16 @@ nfqnl_build_packet_message(struct nfqnl_instance *queue,if (data_len ==0 || data_len > entskb->len) data_len = entskb->len; - size +=nla_total_size (data_len); + +if (!entskb->head_frag || +skb_headlen (entskb) <L1_CACHE_BYTES || +skb_shinfo (entskb)->nr_frags >=MAX_SKB_FRAGS ) + hlen =skb_headlen (entskb); + +if (skb_has_frag_list (entskb)) + hlen = entskb->len; + hlen =min_t (int, data_len, hlen); + size +=sizeof (struct nlattr ) + hlen; cap_len = entskb->len;break ; }
@@ -277,7 +333,6 @@ nfqnl_build_packet_message(struct nfqnl_instance *queue,if (!skb)return NULL ; - old_tail = skb->tail; nlh =nlmsg_put (skb,0 ,0 ,NFNL_SUBSYS_QUEUE <<8 |NFQNL_MSG_PACKET ,sizeof (struct nfgenmsg ),0 );
@@ -382,31 +437,26 @@ nfqnl_build_packet_message(struct nfqnl_instance *queue, goto nla_put_failure; } +if (ct &&nfqnl_ct_put (skb, ct, ctinfo) <0 ) + goto nla_put_failure; + +if (cap_len >0 &&nla_put_be32 (skb,NFQA_CAP_LEN ,htonl (cap_len))) + goto nla_put_failure; +if (data_len) {struct nlattr *nla; -int sz =nla_attr_size (data_len); -if (skb_tailroom (skb) <nla_total_size (data_len)) { -printk (KERN_WARNING "nf_queue: no tailroom!\n" ); -kfree_skb (skb); -return NULL ; - } +if (skb_tailroom (skb) <sizeof (*nla) + hlen) + goto nla_put_failure; - nla = (struct nlattr *)skb_put (skb,nla_total_size (data_len)); + nla = (struct nlattr *)skb_put (skb,sizeof (*nla)); nla->nla_type =NFQA_PAYLOAD ; - nla->nla_len = sz; + nla->nla_len =nla_attr_size (data_len); -if (skb_copy_bits (entskb,0 ,nla_data (nla), data_len)) -BUG (); +nfqnl_zcopy (skb, entskb, data_len, hlen); } -if (ct &&nfqnl_ct_put (skb, ct, ctinfo) <0 ) - goto nla_put_failure; - -if (cap_len >0 &&nla_put_be32 (skb,NFQA_CAP_LEN ,htonl (cap_len))) - goto nla_put_failure; - - nlh->nlmsg_len = skb->tail - old_tail; + nlh->nlmsg_len = skb->len;return skb; nla_put_failure:
Signature · 1 line
1.7.10.4
[email protected]
maintainerPATCH 11/12 netfilter: nf_conntrack: add include to fix sparse warning
In reply to [email protected]From: stephen hemminger <[email protected]>
Include header file to pickup prototype of nf_nat_seq_adjust_hook
--- net/netfilter/nf_conntrack_core.c | 1 + 1 file changed, 1 insertion(+)
net/netfilter/nf_conntrack_core.c +1 -0
@@ -48,6 +48,7 @@#include <net/netfilter/nf_conntrack_labels.h> #include <net/netfilter/nf_nat.h> #include <net/netfilter/nf_nat_core.h> +#include <net/netfilter/nf_nat_helper.h> #define NF_CONNTRACK_VERSION "0.5.0"
Signature · 1 line
1.7.10.4
[email protected]
maintainerPATCH 12/12 netfilter: nf_conntrack: speed up module removal path if netns in use
In reply to [email protected]From: Vladimir Davydov <[email protected]>
The patch introduces nf_conntrack_cleanup_net_list(), which cleanups nf_conntrack for a list of netns and calls synchronize_net() only once for them all. This should reduce netns destruction time.
I've measured cleanup time for 1k dummy net ns. Here are the results:
<without the patch>
# modprobe nf_conntrack # time modprobe -r nf_conntrack
real 0m10.337s
user 0m0.000s
sys 0m0.376s
<with the patch>
# modprobe nf_conntrack # time modprobe -r nf_conntrack
real 0m5.661s user 0m0.000s sys 0m0.216s
--- include/net/netfilter/nf_conntrack_core.h | 1 + net/netfilter/nf_conntrack_core.c | 46 ++++++++++++++++++++--------- net/netfilter/nf_conntrack_standalone.c | 16 ++++++---- 3 files changed, 43 insertions(+), 20 deletions(-)
include/net/netfilter/nf_conntrack_core.h +1 -0
@@ -27,6 +27,7 @@ extern unsigned int nf_conntrack_in(struct net *net,extern int nf_conntrack_init_net (struct net *net);extern void nf_conntrack_cleanup_net (struct net *net); +extern void nf_conntrack_cleanup_net_list (struct list_head *net_exit_list);extern int nf_conntrack_proto_pernet_init (struct net *net);extern void nf_conntrack_proto_pernet_fini (struct net *net);
net/netfilter/nf_conntrack_core.c +32 -14
62 lines not on this page · raw message
net/netfilter/nf_conntrack_standalone.c +10 -6
@@ -545,16 +545,20 @@ out_init:return ret; } -static void nf_conntrack_pernet_exit (struct net *net) +static void nf_conntrack_pernet_exit (struct list_head *net_exit_list) { -nf_conntrack_standalone_fini_sysctl (net); -nf_conntrack_standalone_fini_proc (net); -nf_conntrack_cleanup_net (net); +struct net *net; + +list_for_each_entry (net, net_exit_list, exit_list) { +nf_conntrack_standalone_fini_sysctl (net); +nf_conntrack_standalone_fini_proc (net); + } +nf_conntrack_cleanup_net_list (net_exit_list); }static struct pernet_operations nf_conntrack_net_ops = { - .init = nf_conntrack_pernet_init, - .exit = nf_conntrack_pernet_exit, + .init = nf_conntrack_pernet_init, + .exit_batch = nf_conntrack_pernet_exit, };static int __init nf_conntrack_standalone_init (void )
Signature · 1 line
1.7.10.4
David Miller
maintainerFrom: [email protected]
Date: Mon, 25 Mar 2013 13:15:40 +0100
Quoted 2 lines
The following patchset contains Netfilter/IPVS updates for your net-next tree, they are:
...
Quoted 3 lines
You can pull these changes from:
git://1984.lsi.us.es/nf-nextmaster
Pulled, thanks a lot Pablo.