linux-stable/net/ipv4
Kuniyuki Iwashima 325ac51bee udp: Improve load balancing for SO_REUSEPORT.
[ Upstream commit efc6b6f6c3 ]

Currently, SO_REUSEPORT does not work well if connected sockets are in a
UDP reuseport group.

Then reuseport_has_conns() returns true and the result of
reuseport_select_sock() is discarded. Also, unconnected sockets have the
same score, hence only does the first unconnected socket in udp_hslot
always receive all packets sent to unconnected sockets.

So, the result of reuseport_select_sock() should be used for load
balancing.

The noteworthy point is that the unconnected sockets placed after
connected sockets in sock_reuseport.socks will receive more packets than
others because of the algorithm in reuseport_select_sock().

    index | connected | reciprocal_scale | result
    ---------------------------------------------
    0     | no        | 20%              | 40%
    1     | no        | 20%              | 20%
    2     | yes       | 20%              | 0%
    3     | no        | 20%              | 40%
    4     | yes       | 20%              | 0%

If most of the sockets are connected, this can be a problem, but it still
works better than now.

Fixes: acdcecc612 ("udp: correct reuseport selection with connected sockets")
CC: Willem de Bruijn <willemb@google.com>
Reviewed-by: Benjamin Herrenschmidt <benh@amazon.com>
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2020-07-31 18:37:49 +02:00
..
bpfilter bpfilter: remove trailing newline 2018-07-24 14:10:42 -07:00
netfilter netfilter: nf_conntrack_pptp: prevent buffer overflows in debug code 2020-06-03 08:19:47 +02:00
af_inet.c net: don't clear sock->sk early to avoid trouble in strparser 2020-01-27 14:50:52 +01:00
ah4.c
arp.c proc: introduce proc_create_net{,_data} 2018-05-16 07:24:30 +02:00
cipso_ipv4.c netlabel: cope with NULL catmap 2020-05-20 08:18:35 +02:00
datagram.c inet: stop leaking jiffies on the wire 2019-11-10 11:27:37 +01:00
devinet.c devinet: fix memleak in inetdev_init() 2020-06-10 21:34:58 +02:00
esp4.c esp4: add length check for UDP encapsulation 2019-05-25 18:23:41 +02:00
esp4_offload.c Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec-next 2018-07-27 09:33:37 -07:00
fib_frontend.c ipv4: fix route update on metric change. 2019-11-10 11:27:50 +01:00
fib_lookup.h
fib_notifier.c
fib_rules.c net: fib_rules: add extack support 2018-04-23 10:21:24 -04:00
fib_semantics.c net: Fix the arp error in some cases 2020-06-30 23:17:06 -04:00
fib_trie.c ipv4: fix a RCU-list lock in fib_triestat_seq_show 2020-04-13 10:44:57 +02:00
fou.c net: fou: do not use guehdr after iptunnel_pull_offloads in gue_udp_recv 2019-04-27 09:36:31 +02:00
gre_demux.c gre: fix uninit-value in __iptunnel_pull_header 2020-03-18 07:14:11 +01:00
gre_offload.c Merge ra.kernel.org:/pub/scm/linux/kernel/git/davem/net 2018-07-03 10:29:26 +09:00
icmp.c net: icmp: fix data-race in cmp_global_allow() 2020-01-04 19:13:30 +01:00
igmp.c ipv4/igmp: fix v1/v2 switchback timeout based on rfc3376, 8.12 2019-12-01 09:17:05 +01:00
inet_connection_sock.c net: inet_csk: Fix so_reuseport bind-address cache in tb->fast* 2020-06-03 08:19:08 +02:00
inet_diag.c inet_diag: return classid for all socket types 2020-03-18 07:14:11 +01:00
inet_fragment.c net: IP defrag: encapsulate rbtree defrag code into callable functions 2019-04-27 09:36:33 +02:00
inet_hashtables.c tcp/dccp: fix possible race __inet_lookup_established() 2020-01-04 19:13:41 +01:00
inet_timewait_sock.c soreuseport: initialise timewait reuseport field 2018-04-07 22:32:32 -04:00
inetpeer.c inetpeer: fix data-race in inet_putpeer / inet_putpeer 2020-01-04 19:13:29 +01:00
ip_forward.c net: clear skb->tstamp in forwarding paths 2019-01-09 17:38:31 +01:00
ip_fragment.c net: IP defrag: encapsulate rbtree defrag code into callable functions 2019-04-27 09:36:33 +02:00
ip_gre.c net: ip_gre: Accept IFLA_INFO_DATA-less configuration 2020-04-02 15:28:13 +02:00
ip_input.c vrf: check accept_source_route on the original netdevice 2019-04-17 08:38:42 +02:00
ip_options.c vrf: check accept_source_route on the original netdevice 2019-04-17 08:38:42 +02:00
ip_output.c net: always initialize pagedlen 2020-01-27 14:50:03 +01:00
ip_sockglue.c net: bpfilter: fix iptables failure if bpfilter_umh is disabled 2019-12-01 09:17:18 +01:00
ip_tunnel.c ip_tunnel: fix use-after-free in ip_tunnel_lookup() 2020-06-30 23:17:06 -04:00
ip_tunnel_core.c ip_tunnel: allow not to count pkts on tstats by setting skb's dev to NULL 2019-08-04 09:30:57 +02:00
ip_vti.c ip_vti: receive ipip packet by calling ip_tunnel_rcv 2020-06-03 08:19:46 +02:00
ipcomp.c
ipconfig.c Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next 2018-06-06 18:39:49 -07:00
ipip.c net: ipip: fix wrong address family in init error path 2020-06-03 08:19:10 +02:00
ipmr.c ipmr: Fix skb headroom in ipmr_get_route(). 2019-11-20 18:45:11 +01:00
ipmr_base.c net: ipmr: fix unresolved entry dumps 2018-10-17 22:35:42 -07:00
Kconfig vti[6]: fix packet tx through bpf_redirect() in XinY cases 2020-04-02 15:28:18 +02:00
Makefile net: remove blank lines at end of file 2018-07-24 14:10:43 -07:00
metrics.c net: metrics: add proper netlink validation 2018-06-05 12:29:43 -04:00
netfilter.c netfilter: utils: move nf_ip_checksum* from ipv4 to utils 2018-07-16 17:51:48 +02:00
netlink.c ipv4: Add ICMPv6 support when parse route ipproto 2019-03-10 07:17:17 +01:00
ping.c ipv4: fill fl4_icmp_{type,code} in ping_v4_sendmsg 2020-07-22 09:31:59 +02:00
proc.c tcp: tcp_fragment() should apply sane memory limits 2019-06-17 19:51:56 +02:00
protocol.c
raw.c ipv4: Use return value of inet_iif() for __raw_v4_lookup in the while loop 2019-07-03 13:14:46 +02:00
raw_diag.c inet_diag: return classid for all socket types 2020-03-18 07:14:11 +01:00
route.c net: revert "net: get rid of an signed integer overflow in ip_idents_reserve()" 2020-06-03 08:19:14 +02:00
syncookies.c tcp: handle inet_csk_reqsk_queue_add() failures 2019-03-19 13:12:39 +01:00
sysctl_net_ipv4.c tcp: add tcp_min_snd_mss sysctl 2019-06-17 19:51:56 +02:00
tcp.c tcp: md5: allow changing MD5 keys in all socket states 2020-07-22 09:32:00 +02:00
tcp_bbr.c tcp_bbr: improve arithmetic division in bbr_update_bw() 2020-01-29 16:43:17 +01:00
tcp_bic.c
tcp_cdg.c
tcp_cong.c tcp: make sure listeners don't initialize congestion-control state 2020-07-22 09:31:59 +02:00
tcp_cubic.c tcp_cubic: fix spurious HYSTART_DELAY exit upon drop in min RTT 2020-06-30 23:17:06 -04:00
tcp_dctcp.c tcp: Ensure DCTCP reacts to losses 2019-04-17 08:38:41 +02:00
tcp_diag.c tcp: annotate tp->rcv_nxt lockless reads 2020-01-09 10:19:08 +01:00
tcp_fastopen.c
tcp_highspeed.c
tcp_htcp.c
tcp_hybla.c
tcp_illinois.c net/tcp/illinois: replace broken algorithm reference link 2018-02-28 12:03:47 -05:00
tcp_input.c tcp: allow at most one TLP probe per flight 2020-07-31 18:37:48 +02:00
tcp_ipv4.c tcp: md5: refine tcp_md5_do_add()/tcp_md5_hash_key() barriers 2020-07-22 09:32:00 +02:00
tcp_lp.c
tcp_metrics.c net: Drop pernet_operations::async 2018-03-27 13:18:09 -04:00
tcp_minisocks.c tcp: annotate tp->rcv_nxt lockless reads 2020-01-09 10:19:08 +01:00
tcp_nv.c tcp_nv: fix potential integer overflow in tcpnv_acked 2018-01-31 10:26:30 -05:00
tcp_offload.c tcp: Don't coalesce decrypted and encrypted SKBs 2018-07-16 00:12:09 -07:00
tcp_output.c tcp: allow at most one TLP probe per flight 2020-07-31 18:37:48 +02:00
tcp_rate.c tcp: expose both send and receive intervals for rate sample 2018-07-11 23:01:56 -07:00
tcp_recovery.c tcp: add stat of data packet reordering events 2018-08-01 09:56:10 -07:00
tcp_scalable.c
tcp_timer.c tcp: fix SNMP TCP timeout under-estimation 2019-12-13 08:52:20 +01:00
tcp_ulp.c tcp, ulp: fix leftover icsk_ulp_ops preventing sock from reattach 2018-08-16 14:58:08 -07:00
tcp_vegas.c
tcp_vegas.h
tcp_veno.c
tcp_westwood.c
tcp_yeah.c
tunnel4.c inet: whitespace cleanup 2018-02-28 11:43:28 -05:00
udp.c udp: Improve load balancing for SO_REUSEPORT. 2020-07-31 18:37:49 +02:00
udp_diag.c inet_diag: return classid for all socket types 2020-03-18 07:14:11 +01:00
udp_impl.h
udp_offload.c net/udp_gso: Allow TX timestamp with UDP GSO 2020-01-27 14:50:56 +01:00
udp_tunnel.c
udplite.c proc: introduce proc_create_net{,_data} 2018-05-16 07:24:30 +02:00
xfrm4_input.c xfrm: reset transport header back to network header after all input transforms ahave been applied 2018-09-04 10:26:30 +02:00
xfrm4_mode_beet.c
xfrm4_mode_transport.c xfrm: reset transport header back to network header after all input transforms ahave been applied 2018-09-04 10:26:30 +02:00
xfrm4_mode_tunnel.c xfrm: Verify MAC header exists before overwriting eth_hdr(skb)->h_proto 2018-03-07 10:54:29 +01:00
xfrm4_output.c xfrm: Always set XFRM_TRANSFORMED in xfrm{4,6}_output_finish 2020-04-29 16:31:23 +02:00
xfrm4_policy.c net: add bool confirm_neigh parameter for dst_ops.update_pmtu 2020-01-04 19:13:37 +01:00
xfrm4_protocol.c
xfrm4_state.c
xfrm4_tunnel.c