linux-stable/include/net
Stewart Smith ecb741a17c tcp: Reduce chance of collisions in inet6_hashfn().
[ Upstream commit d11b0df7dd ]

For both IPv4 and IPv6 incoming TCP connections are tracked in a hash
table with a hash over the source & destination addresses and ports.
However, the IPv6 hash is insufficient and can lead to a high rate of
collisions.

The IPv6 hash used an XOR to fit everything into the 96 bits for the
fast jenkins hash, meaning it is possible for an external entity to
ensure the hash collides, thus falling back to a linear search in the
bucket, which is slow.

We take the approach of hash the full length of IPv6 address in
__ipv6_addr_jhash() so that all users can benefit from a more secure
version.

While this may look like it adds overhead, the reality of modern CPUs
means that this is unmeasurable in real world scenarios.

In simulating with llvm-mca, the increase in cycles for the hashing
code was ~16 cycles on Skylake (from a base of ~155), and an extra ~9
on Nehalem (base of ~173).

In commit dd6d2910c5 ("netfilter: conntrack: switch to siphash")
netfilter switched from a jenkins hash to a siphash, but even the faster
hsiphash is a more significant overhead (~20-30%) in some preliminary
testing.  So, in this patch, we keep to the more conservative approach to
ensure we don't add much overhead per SYN.

In testing, this results in a consistently even spread across the
connection buckets.  In both testing and real-world scenarios, we have
not found any measurable performance impact.

Fixes: 08dcdbf6a7 ("ipv6: use a stronger hash for tcp")
Signed-off-by: Stewart Smith <trawets@amazon.com>
Signed-off-by: Samuel Mendoza-Jonas <samjonas@amazon.com>
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20230721222410.17914-1-kuniyu@amazon.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
2023-08-03 10:22:34 +02:00
..
9p 9p: Add client parameter to p9_req_put() 2022-08-17 14:24:07 +02:00
bluetooth Bluetooth: L2CAP: Fix use-after-free caused by l2cap_chan_put 2022-08-03 12:03:40 +02:00
caif net: remove the caif_hsi driver 2021-07-01 13:19:48 -07:00
iucv net/af_iucv: don't track individual TX skbs for TRANS_HIPER sockets 2021-01-28 20:36:21 -08:00
netfilter netfilter: nf_tables: drop map element references from preparation phase 2023-07-23 13:46:44 +02:00
netns Remove DECnet support from kernel 2023-06-21 15:59:15 +02:00
nfc NFC: add NCI_UNREG flag to eliminate the race 2021-11-25 09:48:40 +01:00
phonet
sctp sctp: add a refcnt in sctp_stream_priorities to avoid a nested loop 2023-03-11 13:57:28 +01:00
tc_act net/sched: transition act_pedit to rcu and percpu stats 2023-03-11 13:57:29 +01:00
6lowpan.h 6lowpan: Replace zero-length array with flexible-array member 2020-02-28 14:51:30 +01:00
act_api.h net_sched: refactor TC action init API 2021-08-02 10:24:38 +01:00
addrconf.h ipv6/addrconf: fix a null-ptr-deref bug for ip6_ptr 2022-08-03 12:03:47 +02:00
af_ieee802154.h
af_rxrpc.h afs: Don't truncate iter during data fetch 2021-04-23 10:17:26 +01:00
af_unix.h af_unix: Add unix_stream_proto for sockmap 2021-08-16 18:43:39 -07:00
af_vsock.h vsock: each transport cycles only on its own sockets 2022-03-23 09:16:41 +01:00
ah.h
arp.h ipv4: Invalidate neighbour for broadcast address upon address addition 2022-04-13 20:59:05 +02:00
atmclip.h
ax25.h ax25: fix reference count leaks of ax25_dev 2022-04-20 09:34:22 +02:00
ax88796.h ax88796: export ax_NS8390_init() hook 2021-08-03 13:05:25 +01:00
bareudp.h bareudp: Reverted support to enable & disable rx metadata collection 2020-07-21 18:30:47 -07:00
bond_3ad.h net: bonding: Share lacpdu_mcast_addr definition 2022-09-28 11:11:48 +02:00
bond_alb.h bonding (gcc13): synchronize bond_{a,t}lb_xmit() types 2023-06-14 11:13:00 +02:00
bond_options.h Bonding: add arp_missed_max option 2023-06-05 09:21:19 +02:00
bonding.h bonding: fix send_peer_notif overflow 2023-06-05 09:21:19 +02:00
bpf_sk_storage.h bpf: struct sock is declared twice in bpf_sk_storage header 2021-03-26 17:43:55 +01:00
busy_poll.h net: Fix a data-race around sysctl_net_busy_poll. 2022-08-31 17:16:43 +02:00
calipso.h
cfg80211-wext.h
cfg80211.h cfg80211: fix management registrations locking 2021-10-25 15:20:22 +02:00
cfg802154.h cfg802154: Replace zero-length array with flexible-array member 2020-02-29 14:39:08 +01:00
checksum.h net: Force inlining of checksum functions in net/checksum.h 2022-03-02 11:47:58 +01:00
cipso_ipv4.h cipso: Remove unused inline functions 2020-07-15 07:45:24 -07:00
cls_cgroup.h bpf: Allow to retrieve cgroup v1 classid from v2 hooks 2020-03-27 19:40:38 -07:00
codel.h
codel_impl.h
codel_qdisc.h
compat.h net/ipv4/ipv6: Replace one-element arraya with flexible-array members 2021-08-05 11:46:42 +01:00
datalink.h
dcbevent.h
dcbnl.h
devlink.h devlink: Use xarray to store devlink instances 2021-08-14 13:59:10 +01:00
dsa.h net: dsa: introduce helpers for iterating through ports using dp 2023-06-05 09:21:17 +02:00
dsfield.h ipv6: Annotate bitwise IPv6 dsfield pointer cast 2019-12-16 16:09:44 -08:00
dst.h net: Remove unused inline function dst_hold_and_use() 2023-06-21 15:59:19 +02:00
dst_cache.h wireguard: device: reset peer src endpoint when netns exits 2021-12-08 09:04:46 +01:00
dst_metadata.h net: fix a memleak when uncloning an skb dst and its metadata 2022-02-16 12:56:30 +01:00
dst_ops.h net/dst: use a smaller percpu_counter batch for dst entries accounting 2020-05-08 21:33:33 -07:00
erspan.h erspan: Add type I version 0 support. 2020-05-05 13:23:29 -07:00
esp.h esp: limit skb_page_frag_refill use to a single page 2022-04-27 14:38:52 +02:00
espintcp.h xfrm: espintcp: save and call old ->sk_destruct 2020-04-20 07:34:16 +02:00
ethoc.h
failover.h
fib_notifier.h ipv6: Remove old route notifications and convert listeners 2019-12-24 22:37:30 -08:00
fib_rules.h ipv6: fix memory leak in fib6_rule_suppress 2021-12-08 09:04:43 +01:00
firewire.h
flow.h net: Remove DECnet leftovers from flow.h. 2023-06-21 15:59:19 +02:00
flow_dissector.h net/sched: flower: fix parsing of ethertype following VLAN header 2022-04-20 09:34:09 +02:00
flow_offload.h netfilter: nf_tables: bail out early if hardware offload is not supported 2022-06-14 18:36:17 +02:00
fou.h
fq.h net/fq_impl: do not maintain a backlog-sorted list of flows 2021-01-21 13:33:45 +01:00
fq_impl.h net/fq_impl: do not maintain a backlog-sorted list of flows 2021-01-21 13:33:45 +01:00
garp.h treewide: Use sizeof_field() macro 2019-12-09 10:36:44 -08:00
gen_stats.h net_sched: extend packet counter to 64bit 2019-11-05 18:20:55 -08:00
genetlink.h mptcp: avoid lock_fast usage in accept path 2021-02-12 16:31:46 -08:00
geneve.h
gre.h ip_gre: add csum offload support for gre header 2021-01-29 20:39:14 -08:00
gro.h gro: add combined call_gro_receive() + INDIRECT_CALL_INET() helper 2021-03-18 19:51:12 -07:00
gro_cells.h
gtp.h
gue.h GUE: Fix a typo 2020-06-22 21:12:44 -07:00
hwbm.h net: hwbm: if CONFIG_NET_HWBM unset, make stub functions static 2019-10-25 16:24:32 -07:00
icmp.h ipv6: ICMPV6: add response to ICMPV6 RFC 8335 PROBE messages 2021-06-28 14:29:45 -07:00
ieee80211_radiotap.h mac80211: Use flex-array for radiotap header bitmap 2021-08-13 09:58:25 +02:00
ieee802154_netdev.h net: ieee802154: return -EINVAL for unknown addr type 2022-10-26 12:35:54 +02:00
if_inet6.h ipv6: fix locking issues with loops over idev->addr_list 2022-06-09 10:22:31 +02:00
ife.h
ila.h
inet6_connection_sock.h
inet6_hashtables.h net: allow unbound socket for packets in VRF when tcp_l3mdev_accept set 2022-08-17 14:23:36 +02:00
inet_common.h bpf: Allow rewriting to ports under ip_unprivileged_port_start 2021-01-27 18:18:15 -08:00
inet_connection_sock.h Revert "tcp: change pingpong threshold to 3" 2022-08-03 12:03:45 +02:00
inet_ecn.h inet_ecn: Use csum16_add() helper for IP_ECN_set_* helpers 2020-12-14 18:38:58 -08:00
inet_frag.h inet: frags: annotate races around fqdir->dead and fqdir->high_thresh 2022-01-27 11:05:35 +01:00
inet_hashtables.h net: allow unbound socket for packets in VRF when tcp_l3mdev_accept set 2022-08-17 14:23:36 +02:00
inet_sock.h net: allow unbound socket for packets in VRF when tcp_l3mdev_accept set 2022-08-17 14:23:36 +02:00
inet_timewait_sock.h
inetpeer.h
ioam6.h ipv6: ioam: Support for IOAM injection with lwtunnels 2021-07-21 08:14:33 -07:00
ip.h ipv{4,6}/raw: fix output xfrm lookup wrt protocol 2023-06-05 09:21:26 +02:00
ip6_checksum.h tcp: remove indirect calls for icsk->icsk_af_ops->send_check 2020-06-20 17:47:53 -07:00
ip6_fib.h ipv6: fix data-race in fib6_info_hw_flags_set / fib6_purge_rt 2022-02-23 12:03:10 +01:00
ip6_route.h Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2021-08-05 15:08:47 -07:00
ip6_tunnel.h ip_gre, ip6_gre: Fix race condition on o_seqno in collect_md mode 2022-05-09 09:14:36 +02:00
ip_fib.h ipv4: convert fib_num_tclassid_users to atomic_t 2021-12-08 09:04:49 +01:00
ip_tunnels.h ip_tunnels: allow VXLAN/GENEVE to inherit TOS/TTL from VLAN 2023-06-28 10:29:44 +02:00
ip_vs.h ipvs: Update width of source for ip_vs_sync_conn_options 2023-05-24 17:36:46 +01:00
ipcomp.h
ipconfig.h
ipv6.h tcp: Reduce chance of collisions in inet6_hashfn(). 2023-08-03 10:22:34 +02:00
ipv6_frag.h inet: frags: annotate races around fqdir->dead and fqdir->high_thresh 2022-01-27 11:05:35 +01:00
ipv6_stubs.h net: ipv6: add fib6_nh_release_dsts stub 2021-12-01 09:04:49 +01:00
iw_handler.h
kcm.h
l3mdev.h l3mdev: add infrastructure for table to VRF mapping 2020-06-20 17:22:22 -07:00
lag.h
lapb.h net: lapb: Make "lapb_t1timer_running" able to detect an already running timer 2021-03-23 14:14:50 -07:00
lib80211.h
llc.h llc: fix out-of-bound array index in llc_sk_dev_hash() 2021-11-18 19:17:10 +01:00
llc_c_ac.h
llc_c_ev.h
llc_c_st.h
llc_conn.h
llc_if.h
llc_pdu.h net: llc: fix skb_over_panic 2021-07-27 13:05:56 +01:00
llc_s_ac.h
llc_s_ev.h
llc_s_st.h
llc_sap.h
lwtunnel.h netfilter: add netfilter hooks to SRv6 data plane 2021-08-30 01:51:36 +02:00
mac80211.h mac80211: Fix Ptk0 rekey documentation 2021-09-27 12:02:54 +02:00
mac802154.h
macsec.h net: macsec: fix the length used to copy the key for offloading 2021-06-24 12:41:12 -07:00
mctp.h mctp: unify sockaddr_mctp types 2021-10-18 13:47:09 +01:00
mctpdevice.h mctp: Remove the repeated declaration 2021-08-25 11:23:14 +01:00
mip6.h net: mip6: Replace zero-length array with flexible-array member 2020-03-02 11:16:27 -08:00
mld.h mld: add new workqueues for process mld events 2021-03-26 15:14:56 -07:00
mpls.h net: Make mpls_entry_encode() available for generic users 2020-05-29 21:20:20 -07:00
mpls_iptunnel.h net: mpls: Replace zero-length array with flexible-array member 2020-02-28 12:08:37 -08:00
mptcp.h mptcp: remove MPTCP 'ifdef' in TCP SYN cookies 2023-01-12 11:58:52 +01:00
mrp.h mrp: introduce active flags to prevent UAF when applicant uninit 2022-12-31 13:14:42 +01:00
ncsi.h
ndisc.h ipv6: fix skb drops in igmp6_event_query() and igmp6_event_report() 2022-03-08 19:12:33 +01:00
neighbour.h neighbour: delete neigh_lookup_nodev as not used 2023-06-21 15:59:19 +02:00
net_failover.h
net_namespace.h net: initialize init_net earlier 2022-04-13 20:59:03 +02:00
net_ratelimit.h
netevent.h
netlabel.h
netlink.h net: netlink: add the case when nlh is NULL 2021-07-27 11:43:50 +01:00
netprio_cgroup.h netprio: use css ID instead of cgroup ID 2019-11-12 08:18:03 -08:00
netrom.h
nexthop.h net: ipv4: Fix rtnexthop len when RTA_FLOW is present 2021-09-24 14:07:10 +01:00
nl802154.h net: ieee802154: handle iftypes as u32 2021-12-01 09:04:46 +01:00
nsh.h
p8022.h
page_pool.h page_pool: fix inconsistency for page_pool_ring_[un]lock() 2023-06-05 09:21:22 +02:00
pie.h pie: realign comment 2020-03-04 13:25:55 -08:00
ping.h
pkt_cls.h sch_htb: Fix inconsistency when leaf qdisc creation fails 2021-08-30 16:33:59 -07:00
pkt_sched.h net/sched: make psched_mtu() RTNL-less safe 2023-07-23 13:47:45 +02:00
pptp.h
protocol.h tcp/udp: Make early_demux back namespacified. 2022-11-10 18:15:38 +01:00
psample.h psample: Add a fwd declaration for skbuff 2021-08-09 15:34:21 -07:00
psnap.h
raw.h raw: Fix a data-race around sysctl_raw_l3mdev_accept. 2022-07-21 21:24:27 +02:00
rawv6.h
red.h sch_red: fix off-by-one checks in red_check_params() 2021-03-25 17:40:43 -07:00
regulatory.h net/wireless: regulatory.h: drop duplicate word in comment 2020-07-31 09:24:23 +02:00
request_sock.h tcp: bpf: Optionally store mac header in TCP_SAVE_SYN 2020-08-24 14:35:00 -07:00
rose.h
route.h ip: Fix data-races around sysctl_ip_default_ttl. 2022-07-29 17:25:09 +02:00
rpl.h ipv6: rpl: Fix Route of Death. 2023-06-14 11:13:02 +02:00
rsi_91x.h
rtnetlink.h net: add extack arg for link ops 2021-08-04 10:01:26 +01:00
rtnh.h
sch_generic.h net/sched: sch_taprio: fix possible use-after-free 2023-02-01 08:27:09 +01:00
scm.h scm: fix MSG_CTRUNC setting condition for SO_PASSSEC 2023-05-11 23:00:26 +09:00
secure_seq.h secure_seq: use the 64 bits of the siphash for port offset calculation 2022-05-18 10:26:53 +02:00
seg6.h udp6: Use Segment Routing Header for dest address if present 2022-01-27 11:05:05 +01:00
seg6_hmac.h
seg6_local.h
selftests.h net: selftest: fix build issue if INET is disabled 2021-04-28 14:06:45 -07:00
slhc_vj.h
smc.h net/smc: introduce CHID callback for ISM devices 2020-09-28 15:19:03 -07:00
snmp.h
sock.h netlink: Add __sock_i_ino() for __netlink_diag_dump(). 2023-07-23 13:46:56 +02:00
sock_reuseport.h soreuseport: Fix socket selection for SO_INCOMING_CPU. 2022-12-31 13:14:07 +01:00
Space.h wan: remove sbni/granch driver 2021-08-03 13:05:26 +01:00
stp.h
strparser.h bpf, sockmap: sk_skb data_end access incorrect when src_reg = dst_reg 2021-11-18 19:17:11 +01:00
switchdev.h net: make switchdev_bridge_port_{,unoffload} loosely coupled with the bridge 2021-08-04 12:35:07 +01:00
tcp.h tcp: annotate data-races around tp->notsent_lowat 2023-07-27 08:47:04 +02:00
tcp_states.h
timewait_sock.h
tipc.h
tls.h net/tls: Check for errors in tls_device_init 2022-07-21 21:24:31 +02:00
tls_toe.h
transp_v6.h tcp: move ipv4_specific to tcp include file 2020-06-23 20:10:15 -07:00
tso.h net: tso: cache transport header length 2020-06-18 20:46:23 -07:00
tun_proto.h
udp.h tcp/udp: Call inet6_destroy_sock() in IPv6 sk->sk_destruct(). 2023-04-26 13:51:54 +02:00
udp_tunnel.h rxrpc: Fix ICMP/ICMP6 error handling 2022-09-15 11:30:05 +02:00
udplite.h tcp/udp: Call inet6_destroy_sock() in IPv6 sk->sk_destruct(). 2023-04-26 13:51:54 +02:00
vsock_addr.h vsock: remove include/linux/vm_sockets.h file 2019-11-14 18:12:17 -08:00
vxlan.h vxlan: calculate correct header length for GPE 2023-08-03 10:22:34 +02:00
wext.h
x25.h net/x25: add new state X25_STATE_5 2019-12-09 10:28:43 -08:00
x25device.h
xdp.h xdp: Allow registering memory model without rxq reference 2023-06-05 09:21:21 +02:00
xdp_priv.h page_pool: do not release pool until inflight == 0. 2019-11-16 12:39:10 -08:00
xdp_sock.h xdp: Add proper __rcu annotations to redirect map entries 2021-06-24 19:41:15 +02:00
xdp_sock_drv.h i40e: xsk: Move tmp desc array from driver to pool 2022-06-14 18:36:18 +02:00
xfrm.h xfrm: Treat already-verified secpath entries as optional 2023-06-28 10:29:45 +02:00
xsk_buff_pool.h xsk: Fix unaligned descriptor validation 2023-05-11 23:00:27 +09:00