linux/net/netfilter
Florian Westphal 7d08487777 netfilter: connlimit: use rbtree for per-host conntrack obj storage
With current match design every invocation of the connlimit_match
function means we have to perform (number_of_conntracks % 256) lookups
in the conntrack table [ to perform GC/delete stale entries ].
This is also the reason why ____nf_conntrack_find() in perf top has
> 20% cpu time per core.

This patch changes the storage to rbtree which cuts down the number of
ct objects that need testing.

When looking up a new tuple, we only test the connections of the host
objects we visit while searching for the wanted host/network (or
the leaf we need to insert at).

The slot count is reduced to 32.  Increasing slot count doesn't
speed up things much because of rbtree nature.

before patch (50kpps rx, 10kpps tx):
+  20.95%  ksoftirqd/0  [nf_conntrack] [k] ____nf_conntrack_find
+  20.50%  ksoftirqd/1  [nf_conntrack] [k] ____nf_conntrack_find
+  20.27%  ksoftirqd/2  [nf_conntrack] [k] ____nf_conntrack_find
+   5.76%  ksoftirqd/1  [nf_conntrack] [k] hash_conntrack_raw
+   5.39%  ksoftirqd/2  [nf_conntrack] [k] hash_conntrack_raw
+   5.35%  ksoftirqd/0  [nf_conntrack] [k] hash_conntrack_raw

after (90kpps, 51kpps tx):
+  17.24%       swapper  [nf_conntrack]    [k] ____nf_conntrack_find
+   6.60%   ksoftirqd/2  [nf_conntrack]    [k] ____nf_conntrack_find
+   2.73%       swapper  [nf_conntrack]    [k] hash_conntrack_raw
+   2.36%       swapper  [xt_connlimit]    [k] count_tree

Obvious disadvantages to previous version are the increase in code
complexity and the increased memory cost.

Partially based on Eric Dumazets fq scheduler.

Reviewed-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
2014-03-17 11:11:57 +01:00
..
ipset netfilter: Convert uses of __constant_<foo> to <foo> 2014-03-13 14:13:19 +01:00
ipvs ipvs: Reduce checkpatch noise in ip_vs_lblc.c 2014-03-07 12:36:21 +09:00
core.c netfilter: pass hook ops to hookfn 2013-10-14 11:29:31 +02:00
Kconfig netfilter: nf_tables: add reject module for NFPROTO_INET 2014-02-06 09:44:18 +01:00
Makefile netfilter: nf_tables: add reject module for NFPROTO_INET 2014-02-06 09:44:18 +01:00
nf_conntrack_acct.c netfilter: introduce nf_conn_acct structure 2013-11-03 21:48:49 +01:00
nf_conntrack_amanda.c netfilter: add my copyright statements 2013-04-18 20:27:55 +02:00
nf_conntrack_broadcast.c
nf_conntrack_core.c netfilter: conntrack: remove central spinlock nf_conntrack_lock 2014-03-07 11:41:13 +01:00
nf_conntrack_ecache.c netfilter: add my copyright statements 2013-04-18 20:27:55 +02:00
nf_conntrack_expect.c netfilter: conntrack: seperate expect locking from nf_conntrack_lock 2014-03-07 11:41:01 +01:00
nf_conntrack_extend.c netfilter: nf_ct_ext: support variable length extensions 2012-06-16 15:08:49 +02:00
nf_conntrack_ftp.c netfilter: Implement RFC 1123 for FTP conntrack 2013-05-27 13:32:43 +02:00
nf_conntrack_h323_asn1.c netfilter: h323: bug in parsing of ASN1 SEQOF field 2011-04-04 15:21:02 +02:00
nf_conntrack_h323_main.c netfilter: conntrack: seperate expect locking from nf_conntrack_lock 2014-03-07 11:41:01 +01:00
nf_conntrack_h323_types.c
nf_conntrack_helper.c netfilter: conntrack: remove central spinlock nf_conntrack_lock 2014-03-07 11:41:13 +01:00
nf_conntrack_irc.c netfilter: add my copyright statements 2013-04-18 20:27:55 +02:00
nf_conntrack_l3proto_generic.c
nf_conntrack_labels.c netfilter: connlabels: remove unneeded includes 2013-07-31 16:39:18 +02:00
nf_conntrack_netbios_ns.c
nf_conntrack_netlink.c netfilter: conntrack: remove central spinlock nf_conntrack_lock 2014-03-07 11:41:13 +01:00
nf_conntrack_pptp.c netfilter: add my copyright statements 2013-04-18 20:27:55 +02:00
nf_conntrack_proto_dccp.c netfilter: nf_conntrack_dccp: fix skb_header_pointer API usages 2014-01-06 17:40:02 +01:00
nf_conntrack_proto_generic.c netfilter: nf_conntrack: generalize nf_ct_l4proto_net 2012-07-04 19:37:22 +02:00
nf_conntrack_proto_gre.c netfilter: add my copyright statements 2013-04-18 20:27:55 +02:00
nf_conntrack_proto_sctp.c netfilter: add my copyright statements 2013-04-18 20:27:55 +02:00
nf_conntrack_proto_tcp.c netfilter: add SYNPROXY core/target 2013-08-28 00:27:54 +02:00
nf_conntrack_proto_udp.c netfilter: add my copyright statements 2013-04-18 20:27:55 +02:00
nf_conntrack_proto_udplite.c netfilter: nf_log: prepare net namespace support for loggers 2013-04-05 20:12:54 +02:00
nf_conntrack_proto.c netfilter: nf_conntrack: remove dead code 2014-01-03 23:41:37 +01:00
nf_conntrack_sane.c netfilter: nf_ct_helper: better logging for dropped packets 2013-02-19 02:48:05 +01:00
nf_conntrack_seqadj.c netfilter: only warn once on wrong seqadj usage 2014-01-06 14:23:17 +01:00
nf_conntrack_sip.c netfilter: conntrack: seperate expect locking from nf_conntrack_lock 2014-03-07 11:41:01 +01:00
nf_conntrack_snmp.c netfilter: nf_ct_snmp: add include file 2013-01-18 00:28:18 +01:00
nf_conntrack_standalone.c net: Convert uses of typedef ctl_table to struct ctl_table 2013-06-13 02:36:09 -07:00
nf_conntrack_tftp.c netfilter: add my copyright statements 2013-04-18 20:27:55 +02:00
nf_conntrack_timeout.c netfilter: nf_ct_timeout: move initialization out of pernet_operations 2013-01-23 12:56:02 +01:00
nf_conntrack_timestamp.c netfilter: nf_ct_timestamp: Fix BUG_ON after netns deletion 2013-12-20 14:58:29 +01:00
nf_internals.h net: misc: Remove extern from function prototypes 2013-10-19 19:12:11 -04:00
nf_log.c net: Convert uses of typedef ctl_table to struct ctl_table 2013-06-13 02:36:09 -07:00
nf_nat_amanda.c netfilter: add my copyright statements 2013-04-18 20:27:55 +02:00
nf_nat_core.c netfilter: nf_nat: add full port randomization support 2014-01-03 23:41:26 +01:00
nf_nat_ftp.c netfilter: nf_ct_helper: better logging for dropped packets 2013-02-19 02:48:05 +01:00
nf_nat_helper.c netfilter: nf_conntrack: make sequence number adjustments usuable without NAT 2013-08-28 00:26:48 +02:00
nf_nat_irc.c netfilter: nf_nat: fix access to uninitialized buffer in IRC NAT helper 2014-01-06 14:17:17 +01:00
nf_nat_proto_common.c netfilter: nf_nat: add full port randomization support 2014-01-03 23:41:26 +01:00
nf_nat_proto_dccp.c netfilter: add protocol independent NAT core 2012-08-30 03:00:14 +02:00
nf_nat_proto_sctp.c net/sctp: Refactor SCTP skb checksum computation 2013-07-27 20:07:15 -07:00
nf_nat_proto_tcp.c netfilter: add protocol independent NAT core 2012-08-30 03:00:14 +02:00
nf_nat_proto_udp.c netfilter: add protocol independent NAT core 2012-08-30 03:00:14 +02:00
nf_nat_proto_udplite.c netfilter: add protocol independent NAT core 2012-08-30 03:00:14 +02:00
nf_nat_proto_unknown.c netfilter: add protocol independent NAT core 2012-08-30 03:00:14 +02:00
nf_nat_sip.c netfilter: nf_ct_sip: consolidate NAT hook functions 2013-10-01 12:47:09 +02:00
nf_nat_tftp.c netfilter: nf_ct_helper: better logging for dropped packets 2013-02-19 02:48:05 +01:00
nf_queue.c netfilter: move skb_gso_segment into nfnetlink_queue module 2013-04-29 20:09:05 +02:00
nf_sockopt.c
nf_synproxy_core.c netfilter: nf_conntrack: don't release a conntrack with non-zero refcnt 2014-02-05 17:46:06 +01:00
nf_tables_api.c netfilter: nf_tables: restore notifications for anonymous set destruction 2014-03-08 12:35:18 +01:00
nf_tables_core.c netfilter: nf_tables: unininline nft_trace_packet() 2014-02-07 17:50:27 +01:00
nf_tables_inet.c netfilter: nf_tables: fix error path in the init functions 2014-01-09 23:25:48 +01:00
nfnetlink_acct.c netfilter: nfnetlink_acct: fix incomplete dumping of objects 2013-06-05 12:36:36 +02:00
nfnetlink_cthelper.c netfilter: check return code from nla_parse_tested 2013-06-20 11:20:13 +02:00
nfnetlink_cttimeout.c netfilter: cttimeout: allow to set/get default protocol timeouts 2013-10-01 13:17:39 +02:00
nfnetlink_log.c netfilter: nfnetlink_log: remove unused code 2014-02-25 11:30:01 +01:00
nfnetlink_queue_core.c Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/jesse/openvswitch 2014-01-06 19:48:38 -05:00
nfnetlink_queue_ct.c netfilter: nf_conntrack: make sequence number adjustments usuable without NAT 2013-08-28 00:26:48 +02:00
nfnetlink.c netfilter: nfnetlink: add rcu_dereference_protected() helpers 2014-02-25 11:29:21 +01:00
nft_bitwise.c netfilter: nf_tables: expression ops overloading 2013-10-14 17:16:08 +02:00
nft_byteorder.c netfilter: nf_tables: expression ops overloading 2013-10-14 17:16:08 +02:00
nft_cmp.c netfilter: nf_tables: add compatibility layer for x_tables 2013-10-14 18:00:04 +02:00
nft_compat.c netfilter: nf_tables: restore context for expression destructors 2014-03-08 12:35:17 +01:00
nft_counter.c netfilter: nf_tables: expression ops overloading 2013-10-14 17:16:08 +02:00
nft_ct.c netfilter: nft_ct: remove family from struct nft_ct 2014-03-08 12:35:19 +01:00
nft_expr_template.c netfilter: nf_tables: expression ops overloading 2013-10-14 17:16:08 +02:00
nft_exthdr.c netfilter: nft_exthdr: call ipv6_find_hdr() with explicitly initialized offset 2013-12-20 11:25:10 +01:00
nft_hash.c netfilter: nft_hash: bug fixes and resizing 2014-03-07 11:42:07 +01:00
nft_immediate.c netfilter: nf_tables: restore context for expression destructors 2014-03-08 12:35:17 +01:00
nft_limit.c netfilter: nf_tables: expression ops overloading 2013-10-14 17:16:08 +02:00
nft_log.c netfilter: nf_tables: restore context for expression destructors 2014-03-08 12:35:17 +01:00
nft_lookup.c netfilter: nf_tables: restore notifications for anonymous set destruction 2014-03-08 12:35:18 +01:00
nft_meta.c netfilter: nft_meta: fix lack of validation of the input register 2014-01-09 20:04:16 +01:00
nft_nat.c netfilter: nft_nat: fix family validation 2014-03-08 12:35:19 +01:00
nft_payload.c netfilter: nf_tables: nft_payload: fix transport header base 2013-10-14 18:00:56 +02:00
nft_queue.c netfilter: nf_tables: fix log/queue expressions for NFPROTO_INET 2014-02-06 11:41:38 +01:00
nft_rbtree.c netfilter: nft_rbtree: fix data handling of end interval elements 2014-02-07 14:22:06 +01:00
nft_reject_inet.c netfilter: nf_tables: add reject module for NFPROTO_INET 2014-02-06 09:44:18 +01:00
nft_reject.c netfilter: nft_reject: split up reject module into IPv4 and IPv6 specifc parts 2014-02-06 09:44:10 +01:00
x_tables.c netfilter: x_tables: fix ordering of jumpstack allocation and table update 2013-10-22 10:11:29 +02:00
xt_addrtype.c netfilter: xt_addrtype: fix trivial typo 2013-07-31 16:36:25 +02:00
xt_AUDIT.c netfilter: Convert uses of __constant_<foo> to <foo> 2014-03-13 14:13:19 +01:00
xt_bpf.c netfilter: x_tables: add xt_bpf match 2013-01-21 12:20:19 +01:00
xt_cgroup.c netfilter: x_tables: lightweight process control group matching 2014-01-03 23:41:44 +01:00
xt_CHECKSUM.c
xt_CLASSIFY.c
xt_cluster.c
xt_comment.c
xt_connbytes.c netfilter: introduce nf_conn_acct structure 2013-11-03 21:48:49 +01:00
xt_connlabel.c netfilter: add connlabel conntrack extension 2013-01-18 00:28:15 +01:00
xt_connlimit.c netfilter: connlimit: use rbtree for per-host conntrack obj storage 2014-03-17 11:11:57 +01:00
xt_connmark.c netfilter: Fix FSF address in file headers 2013-12-06 12:37:57 -05:00
xt_CONNSECMARK.c
xt_conntrack.c netfilter: add my copyright statements 2013-04-18 20:27:55 +02:00
xt_cpu.c
xt_CT.c netfilter: nf_conntrack: don't release a conntrack with non-zero refcnt 2014-02-05 17:46:06 +01:00
xt_dccp.c
xt_devgroup.c
xt_dscp.c
xt_DSCP.c netfilter: IPv6: fix DSCP mangle code 2011-05-10 10:00:21 +02:00
xt_ecn.c netfilter: xtables: collapse conditions in xt_ecn 2011-12-27 20:45:25 +01:00
xt_esp.c
xt_hashlimit.c Revert "netfilter: avoid get_random_bytes calls" 2014-01-06 14:00:55 +01:00
xt_helper.c
xt_hl.c netfilter: Reduce switch/case indent 2011-07-01 16:11:15 -07:00
xt_HL.c netfilter: Reduce switch/case indent 2011-07-01 16:11:15 -07:00
xt_HMARK.c ipv6: Move ipv6_find_hdr() out of Netfilter code. 2012-11-09 17:05:07 -08:00
xt_IDLETIMER.c netfilter: Remove unnecessary OOM logging messages 2011-11-01 09:19:49 +01:00
xt_ipcomp.c netfilter: xt_ipcomp: Use ntohs to ease sparse warning 2014-02-19 11:41:25 +01:00
xt_iprange.c
xt_ipvs.c ipvs: API change to avoid rescan of IPv6 exthdr 2012-09-28 11:34:33 +09:00
xt_l2tp.c netfilter: introduce l2tp match extension 2014-01-09 21:36:39 +01:00
xt_LED.c
xt_length.c
xt_limit.c netfilter: add my copyright statements 2013-04-18 20:27:55 +02:00
xt_LOG.c netfilter: xt_LOG: fix mark logging for IPv6 packets 2013-05-29 12:29:18 +02:00
xt_mac.c netfilter: Convert compare_ether_addr to ether_addr_equal 2012-05-09 20:49:18 -04:00
xt_mark.c
xt_multiport.c
xt_nat.c netfilter: xt_nat: fix incorrect hooks for SNAT and DNAT targets 2012-10-15 13:39:12 +02:00
xt_NETMAP.c netfilter: combine ipt_NETMAP and ip6t_NETMAP 2012-09-21 12:11:08 +02:00
xt_nfacct.c netfilter: xtables: add nfacct match to support extended accounting 2011-12-25 02:43:17 +01:00
xt_NFLOG.c netfilter: log: netns NULL ptr bug when calling from conntrack 2013-05-15 14:11:07 +02:00
xt_NFQUEUE.c netfilter: xt_NFQUEUE: separate reusable code 2013-12-07 23:20:45 +01:00
xt_osf.c netfilter: Fix FSF address in file headers 2013-12-06 12:37:57 -05:00
xt_owner.c userns: xt_owner: Add basic user namespace support. 2012-08-14 21:55:30 -07:00
xt_physdev.c
xt_pkttype.c
xt_policy.c
xt_quota.c net: Fix files explicitly needing to include module.h 2011-10-31 19:30:28 -04:00
xt_rateest.c net_sched: add 64bit rate estimators 2013-06-11 02:51:03 -07:00
xt_RATEEST.c Revert "netfilter: avoid get_random_bytes calls" 2014-01-06 14:00:55 +01:00
xt_realm.c
xt_recent.c Revert "netfilter: avoid get_random_bytes calls" 2014-01-06 14:00:55 +01:00
xt_REDIRECT.c netfilter: combine ipt_REDIRECT and ip6t_REDIRECT 2012-09-21 12:12:05 +02:00
xt_repldata.h
xt_sctp.c
xt_SECMARK.c
xt_set.c Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial 2013-11-15 16:47:22 -08:00
xt_socket.c netfilter: xt_socket: use sock_gen_put() 2013-10-17 10:27:25 +02:00
xt_state.c
xt_statistic.c net: replace macros net_random and net_srandom with direct calls to prandom 2014-01-14 15:15:25 -08:00
xt_string.c
xt_tcpmss.c
xt_TCPMSS.c netfilter: xt_TCPMSS: lookup route from proper net namespace 2013-09-27 16:18:23 +02:00
xt_TCPOPTSTRIP.c netfilter: xt_TCPOPTSTRIP: fix possible off by one access 2013-08-01 11:45:15 +02:00
xt_tcpudp.c
xt_TEE.c net: pass info struct via netdevice notifier 2013-05-28 13:11:01 -07:00
xt_time.c netfilter: xt_time: add support to ignore day transition 2012-09-24 14:29:01 +02:00
xt_TPROXY.c ipv6: make lookups simpler and faster 2013-10-09 00:01:25 -04:00
xt_TRACE.c
xt_u32.c